The Johnson Viking II was probably the best HF transmitter for AM and CW ever made. It covers nearly all frequencies from 1.8 MC to 30 MC continuously. There was even the coveted CDC model Viking II that had no gaps in coverage, due to its intended use for Civil Defense stations. It is ideal for modern time CW operation on the WARC bands, assuming you have a crystal or feed the VFO input with the proper frequency. You can use the original Viking 122 VFO or a modern synthesized digital VFO for wider frequency coverage. This is my vintage radio of choice, even over the Valiant. It employs WW2 surplus parts and is built like a brick outhouse. It does not use some cheesy pi network inter stage coupling network. It is all parallel tuned. This avoids subharmonic problems. It has been popular for use as a Pirate Radio Transmitter, but it sounds terrible with the stock audio. The audio can be corrected. I found an entirely different approach for audio improvement which does not require the invasive methods commonly employed.

The RF deck does not need anything other than neutralization and grid block keying to make it flawless. No weird compactron tubes or output tubes are used. While the 4D32 used in the Viking I is actually a good tube, once the surplus market drys up, you will not be able to get one for any price. All the 6146 family works fine in this rig, if you neutralize the final RF amplifier.

The grid block keying modification is fully documented here, with actual scope waveform photos of the keying characteristics. The Viking II is fully modernized by this EFJ factory mod and eliminates key clicks and chirp. There are a number of minor modifications that improve the reliability of the Viking II shown here.

This web page is to reorganize and condense the essential info (eliminating some extraneous tangential material) and update articles about the Johnson Viking II originally published in Electric Radio. Many photos are included that would not have fit in a print magazine. I have responded to a number of emails requesting help and explanation. Some have contributed their own unique solutions for this web page, with photos and links to obtain the parts. This page is intended to support Electric Radio readers and save me the labor of lengthy email exchanges with those wishing to work on their Viking II based on those articles. You can obtain the full article reprints from Electric Radio.
https://www.ermag.com/
Electric Radio is the one magazine I read cover to cover, including the ads. If you enjoy owning and operating vintage ham gear, SWL radios, WW2 Surplus, or want to learn more about repairs and home brewing, you should subscribe. Its an exceptional value.

REPRINTS: Request ER #367 12/2019 Neutralization, ER #368 1/2020 Repair & Reliability, ER #369 2/2020 Modulator & Power Stage Improvements, ER #370 3/2020 Modulator-Driver Performance Improvements
PART 5, ER #376 9/2020 Mods Wrap-up, EFJ Factory Mods, WARC Bands, CDC Model, Grid Block Keying SUPERCEDES ALL CHANGES IN THE PREVIOUS ARTICLES.

The series was a voyage of discovery that took me back to a more stock approach. The last article revisits some of the work in the earlier articles and comes up with a simplified audio system modification and supercedes all the previous work. I made some discoveries when working on a friend's CDC model, and revised my thinking. That work was organized chronologically. This web page reorganizes some of it as a step by step process based on that information. A lot of the ER material is not included here, because I am showing the basics you need. Also, some different photos are here, which would not have fit into the print media presentation. Some of the material here is brand new, based on correspondence with other Viking II aficianados.

If you want to have your Johnson transmitter professionally restored, there are resources to do that:
http://www.kiss-electronics.com/johnson.htm
http://www.johnsonradioresto.com/Home_Page.php

INDEX CLICK ON TOPIC TO NAVIGATE DONT START HERE! REPAIR & RELIABILITY EFJ FACTORY MODS SIMPLIFIED NEGATIVE CYCLE LOADING CIRCUIT NEUTRALIZATION & PARASITIC CHOKES WARC BAND OPERATION OF VIKING II AND CDC MODULATION LINEARITY AND HV SUPPLY MODS ADJUSTING THE BAND WIDTH VARIOUS AUDIO DRIVER MODS EFJ GRID BLOCK KEYING MOD REGULATING THE 807 SCREEN CIRCUIT WA1HLR, TIMTRON'S NOTES ON THE VIKING II VIKING II CDC REAR PANEL CONNECTIONS BY WA2LXB WA2LXB VIKING II TUNEUP PROCEDURE VIKING II HISTORY

The following is my personal opinion, based on years of electronic work. Many inexperienced people start at the wrong point in the restoration process with a lot of polish and paint for a piece which has not been proven to be restorable first. People remove transformers and chokes to do rust treatment and paint before verifying operation; this poorly prioritized approach can often damage the fragile cloth covered wires and may do irreparable damage to unobtainium components. If you are a commercial restorer with a lot of experience, and have lots of parts inventory, this may work for you, but the home based restorer would be well served by using a different method.
Many people begin with a shotgun replacement of all components, introducing wiring errors and new problems, before they have done the foundation work of just getting the equipment running.
Some even dive in with a bunch of questionable modifications they found on the internet, before the equipment is working right.
This sort of haphazard work is often done by rookies who lack the test equipment or basic electronic experience, and sadly it winds up in a dumpster or on someone else's bench as a major job of circuit tracing all wiring, ripping it all out to get back to stock, before genuine repair work can be done in a systematic fashion. Once the equipment is proven to work in stock form, any modifications you like can begin. Keep in mind, some gear is built by inexperienced people from a kit and contains wiring errors and bad solder joints. The Viking II was both factory assembled and kit form.

A restorer with a couple simpler projects like a receiver CAN do work on a Viking II successfully, if you focus on the real basic work. On this page, I do not get into any math or theory that you did not cover on your license exam; put it to practical application and learn something about electronics. That's real ham radio. You can write this off as grouchy crap from a curmudgeon if you want, but the truth is the truth, like it or not. More than a dozen Viking IIs have crossed my bench, and all of them benefited from the encounter. I am retired now, so please do not ask me to do repairs. I will offer what technical help I reasonably can. Now that I got that off my chest, let's get started.

I started in the wrong place. Like everyone else, I began with the audio driver and the modulator stage. This resulted in some of the info in the ER articles being out of sequence, and this caused some confusion. It was essential to reorganize this, eliminate anything but the necessary items, and make it step by step rather than a chronological process which might make it difficult for some to follow. Start with the basic repairs, then install the MODULATION LINEARITY AND HV SUPPLY MODS. This is the foundation work that must be done to fix the real problem with poor positive modulation peaks and less than 100% modulation. Then, and only then, tackle the audio driver in any way you wish. What I am offering here first is a way to have a "nearly stock" Viking II that works correctly with minimally invasive changes.

I have seen some radical approaches that make me wonder why they did not start out with a blank chassis, using the components they would rather have, with what is essentially a new design which bears little resemblance to the original product. If that's how you roll, you may still find something useful here.

Some of this topic first appeared in my article in ER #368. Request a reprint for the full article. I have received email inquiries and this is to supplement the article and expand on it with updated information. This avoids lengthy retyped emails explaining details. In the Part 5 final article, some changes were made which replace and supercede statements in the earlier articles. This revision is a condensed version which eliminates confusing elements. A lot of the information is brand new. Some of the information here is based on a web article by WA1HLR, the legendary Timtron (see topic navigator above). Some of it comes from other ER authors. Some of this page comes from other people who wrote me about the article or web page, and had their own solutions to contribute. Credit is given when I use other's work. If I missed you, send me an email and I will fix it.

This article will help you first get the transmitter operational in stock form. Some fixes cure a few of the known warts this otherwise handsome transmitter suffers from. The goal of this methodical procedure is to get it running without damage and correct known reliability problems.

If you have a Viking I or CDC, you will have to compare the schematics and figure out different parts numbering from this article, which is based on the Viking II schematic. The Viking I has big differences besides the 4D32 final. The Viking I did not include a clamp tube or many of the TVI measures in the Viking II. Early Viking IIs did not have the clamp tube either. There are factory mods on BAMA web site for updates. The CDC variant is different on parts numbering and some basic design features. I recommend you print out and study the schematics for ALL of the Viking I, II, and CDC variations. Many of these factory mods are ugly field installations, not properly done, or only partially done. In the end, you can own a fully updated final version Viking II style rig with all the problems solved by E. F. Johnson factory mods. Download the manuals from BAMA. Order Electric Radio reprints, especially the first two Viking II article references by AC0OB and K6AD. NEVER dive into mods before you get something running stock first. If your transmitter looks like a lot of mods or clumsy attempted repairs are present, get photocopies of the schematic and trace the circuits with a highlighter. Correct any questionable wiring. As with any modifications, a la carte partial mods are not recommended for best results.

This web page assumes you have adequate mechanical tools, good soldering equipment and skills (a 100 Watt Weller soldering pistol is a good idea), a good test meter, a scope, an audio generator, downloaded manuals from BAMA that match your radio, and lots of patience. Owning or having access to a tube tester is also recommended. Make friends wtih someone who has a tube tester or buy one (TV-7 or Hickok, even an Eico 667). This restoration is probably out of the reach of a person doing their first time project. The general methods here are applicable to any vintage ham gear.

Test the R25 25K 4 Watt drive pot with an ohm meter. If it feels scratchy, it may have loose windings ready to short out the LV supply. Look for a 4 or 5 watt replacement. You can still find them. These are failure prone because they need a dropping resistor to reduce the power dissipation. If you need to replace it, I recommend you put long teflon wires on the new pot and install it leaving a service loop near R25. Pull out the old wires from the bottom, using the to guide them through the grommet to the chassis bottom. If you ever have to replace it again, the teflon insulation won't melt like the original wire. If R25 is OK, we will install a series resistor to protect it.

Just in case the Drive control is bad, here is a work around if you cannot find the correct part, from the web, by Norm, WB2SYQ.
https://www.amwindow.org/tech/htm/drivepot.htm
Note that if you do this mod, you need to install a 25K 10 W bleeder resistor across the +300 LV supply to bring voltages back in spec, because the drive pot is missing.

Test the big HV divider resistor R13 in the center of the chassis for open circuit. This is still available at Mouser for $20. It is also failure prone because inexperienced people do not loosen the screw fully before adjusting it, destroying the fine wires. C8 usually is leaky or shorted, and that can wipe out the resistor as well. You can do a temporary work around with two separate resistors. The DX-100 does it that way. Temporarily make the 807 control grid bias adjustable to set the resting current with a 5K pot. Some people favor VR tubes for the 807 screens, which is not a trivial mod, and I do not recommend its use for a basic job. The wiring for this mod is tricky, so keep it stock til the rig is working.

NOTE: If R13 has two sliders and no clamp tube, you have an early version and need to perform the Johnson updates adding the clamp tube. This sounds scary, but many hams did this field update successfully, and you can too. If R13 is bad, salvage the sliders off the bad one, if you want to keep the two tap arrangement.

Order a replacement for the large wax capacitor C8 located underneath the oil filled HV capacitor C9. Loosen one clamp on C9, disconnect the wires, and swing C9 out of the way. C8 is always leaky, and causes failure of the divider resistor R13. Replace C8 with a 0.5 uF 600V or 1 KV Orange Drop on the 807 screens. This can be mounted on the tube socket instead of under the oil filled capacitor for easier access. I also substitute an electrolytic cap 5 - 10 uF @ 450V here to get things started.

Order a replacement for C56 0.1 uF 600 or 1000 VDC Orange Drop on clamp tube R30. Older schematics do not show this capacitor; it is essential for proper clamp tube operation. It is an official Johnson factory mod. The wax dripping C56, if it is installed, is always bad.

Test the driver transformer for continuity. If it is open, you need to get a replacement. Pick the proper one for your audio mods because single ended and push pull are not the same thing. Also, a push pull transformer has no gap in the core to allow unbalanced DC current for a single ended (single tube) driver, resulting in saturation and poor low frequency response.

For a stock or modified single ended driver, the P-T156 for $16 works. Others like the Hammond 124B. I found the original driver transformer fine for my use, once the actual problems with modulation were fixed. This means changing the 6146 screen supply and reducing the HV supply dynamic impedance, which is discussed later.

Order a replacement for R35, 56 Ohm 2 Watt non inductive resistor for the 4.7 uH coil that is commonly in the 6146 screen circuit. This was part of the official Johnson TVI mods for early Viking II. Save the RF choke as a spare if L21 fails. Be sure to inspect and install all steps 24 – 28 of the Johnson updates, which are discussed later in this web page:
https://bama.edebris.com/manuals/johnson/viking1/modification-b

Temporarily install a 1 amp fast fuse at F1. Do not install any tubes yet. Turn only the main power switch to ON. Slowly increase AC voltage via a variac to see if the LV transformer is shorted. Use a light bulb in series with the AC HOT line to limit current in case of sudden voltage breakdown. Check bias and LV and filament windings using the manual for the correct AC voltages. I like to do this at 30VAC line first, with the transformer outputs tested for roughly 1/3 the spec voltage in the manual. If it passes this test, gradually increase the variac to 110 VAC, and check the secondary AC output voltages against the values stated in the manual. ONLY AFTER the filament, bias, and LV tests OK, then test HV using the manual reference AC voltages too. All AC voltages should match those in the manual pages 21 & 22 using 115 V AC line input. At this point, you know you have good power transformers, and filament wiring is not shorted. These are dangerous voltages. Exercise safety precautions.

Another item that commonly needs replacement are the toggle switches on the front panel. Use an ohm meter (line cord disconnected from the AC line) to measure the contact resistance. Accept nothing more than 0.5 ohm contact resistance EVERY TIME you operate the switch, or you will have problems later. These switches can result in arcs and sparks in the power circuits, and scratchy erratic carrier on AM. Replace these switches with the best you can get, with a rating of 10 amps at 125 VAC. There is a trick to getting the old switch out. There is a nut between the front panel and the chassis. I remove the nut on the outside of the front panel. Disconnect all the wiring, marking it to get it back correctly. You may need to smash the bad switch with a pair of pliers to be able to unscrew the shaft of the switch backwards into the chassis area. Be careful of the ceramic rotary CW/Phone switch. The nut will fall free between the panel and chassis when the switch comes out. Try to retrieve it to prevent shorting by putting the Viking II upside down and letting it fall out. A Push to Talk relay comes later, after you get everything working right. The PTT relay contacts will be wired in parallel with the existing PLATE switch contacts, so the antenna relay works properly.

Replace all the filter capacitors in the Bias and Low Voltage supplies. Use 105 degrees C, 10,000 hour or better, for longer life. Read the Mouser catalog all the way to the right hand side of the page to see these specs. Before you install the new capacitors, test for shorts on the bias supply and LV supply using an analog VOM. This is essential for a "no surprises or smoke" power up. Good Bias supply reads about 5K due to R15, R16, R17. Good LV supply reads about 25K, due to R25. Good High Voltage supply reads 20K due to R13. Don't just leave the new capacitors dangling; I found terminal strips that fit the old screw locations, as shown in the photo below.

IMPORTANT: Do NOT install a tube shield on the bias rectifier, to prevent arcs to L7! Also dress L21, C32, and C46 away from L7 to prevent similar problems. Now install ONLY V11 6AL5. After you install new filter capacitors on bias and LV, test the Bias supply first, monitor the DC voltage for about 75 – 90 VDC NEGATIVE as you bring the line voltage up to 100 VAC. DON'T INSTALL THE CAPACITOR OR SOLID STATE DIODES BACKWARDS! Remember this is a NEGATIVE SUPPLY!

Disconnect the power, and then install the LV rectifier V10, 5V4. F1 now needs to be 2 amp for this test. Test the LV supply DC while bringing up the variac to 100 VAC slowly. Do not use full AC voltage because there is no current from the other circuitry, so the LV will exceed normal. This completes transformer test.

I installed a 3 wire power cord after this basic work was completed. I also installed a new fuse clip and a new 2 amp fuse in series with the primary of the LV transformer only. Almost none of the contemporary transmitters have a fuse to protect the LV and filament supply in standby or warm up conditions. (The 8 amp main fuse protects only the HV transformer.) I also installed a surge limiter thermistor to bring the bias, filament, and +300V supply voltages up slow on power up. I also added an MOV across the LV transformer primary to protect against AC line transients. I added a short cord and a standard AC female socket for the external antenna relay and receiver muting system while I had that disassembled. The crystal socket mating connector for the antenna relay is expensive, and I question its safety. Other people have suggested replacing the EFJ crystal socket style connector with banana jacks or pin jacks; this is a nice solution also, and much safer than the EFJ connector. I added an MOV across the HV transformer primary, which prevents inductive spikes when the HV is turned off.

Do NOT solid state the +300 LV supply without the thermistor. The solid state replacement MUST have a series resistor to drop the voltage back to EFJ manual specs. It allows the other tubes to warm up before LV comes to full voltage. Burnout of R25 may result. Also, a series resistor MUST be installed in the hot side of R25. I prefer for now to keep the 5V4 tube, for gradual LV B+ application during warm up.

For initial tests, you can use the existing bias resistor network, but the values will likely have drifted. This includes R24, R15, R16, R17. Use new resistors that don't drift. Use a new non inductive resistor for R24. R24 needs an old fashioned carbon composition or Ohmite OX or OY series available from Mouser. Instead of just replacing the bad resistors, I took this opportunity to install the Zener diodes to regulate the 807 grids. See parts list and schematic.

I used 1N4007 diodes and a 100 ohm 1 watt resistor to replace the 6AL5. Be sure to put them in correct polarity, this is a NEGATIVE supply. Due to the lower forward drop, it needed a resistor to bring the values back to nominal voltages in the Viking II manual.

The wires used for PLT and MOD meter shunts in Johnson gear are always inaccurate. Often they have bad solder joints because of the metal incompatibility with regular solder. If the shunt solder joints are bad, ALL the current flows through the special 5 mA meter, blowing it out! I put two 1N4007 diodes back to back directly in parallel across the meter to protect it. This is a 100 mV meter, and limiting it to 600 mV in either direction with the diodes will help protect it, without any inaccuracy.

Remove the PLATE and MOD shunts and throw them out. They can cause low power output or other problems, because you are adjusting tuning with false current readings. Replace them with commonly available modern resistors per the parts list. PLATE requires a 0.202 Ohm meter shunt. MOD requires the same for 500 mA full scale. I like to make MOD 250 mA full scale, which requires 0.404 Ohm meter shunt. See the parts list. If you use these resistors, it will be accurate. The other meter ranges are usually OK because they use real resistors. I test them, and trim them to 2% accuracy using series or parallel resistors. Here is how to check and adjust meter calibration:
MeterCalibration.html
http://www.w8ji.com/metering_amplifier.htm
Direct measurement of the shunt resistors is beyond the scope of the home workshop. A small power supply with a series resistor and a good DVM can confirm meter accuracy using the methods in the links above.

Have at least one crystal in a ham band for testing without the VFO. I like to treat the VFO rebuild separately, for ease of work. If the VFO cable has shorts, it can burn the insulation off the Viking II filament wiring. Look for a melted coil made of black wire near the VFO socket for warning signs. Install a 7.5K 5W resistor in the B+ end of Drive pot R25 to protect it from burn out. Install V6 6AU6, V5 6AQ5, and the 6146s. Do not switch HV on yet. Test the exciter section by tuning it up on all bands. You will get no grid current unless the 6146s are installed. DO NOT EXCEED 8 mA grid current! This is why the red line is on the meter. Despite what you might hear elsewhere, RCA specs of never exceeding 8 mA for a pair of 6146s give the longest tube life. Check R15, R19, C20, C25, and C14 if drive is low or unstable. Unplug the Viking II from the AC line. Lets fix the HV supply problems now before something gets damaged. The 807s are not installed yet, so secure the plate wires and prevent them from arcing to other components or ground.

Install the 5R4s. Test the HV supply to compare the manual DC voltages. If the plate current is not 10 - 20 mA with the key up, adjust the clamp tube pot per the manual. Turn power OFF before putting your hands or a screwdriver in the rig; make small incremental adjustments til the current is correct. Tune the rig up on 80 meters or 40 meters. You should get 100 W or so output, at 230 mA plate current. Note that it is normal for grid current to be higher with the HV OFF than during normal operation with HV ON. Do NOT exceed 8 mA grid current.

Retain the 5R4s, because it keeps the HV within design limits. Using solid state diodes, particularly on key up CW, will result in a much higher HV B+. I blew a filter choke this way from arcing. A series resistor and a high voltage zener diode is necessary to simulate the 5R4 forward drop. It may distort the CW keying wave form if you use solid state replacements.

Arcing commonly happens from the rivets on the 5R4 sockets to the chassis. Remove all the mounting hardware. Also remove and discard the two screws through the back panel near the HV and LV rectifier sockets, see the photos below. These screws are too long, and can even break the rectifier tube sockets. Be careful not to break the fragile cloth covered wires. Use the existing nuts as a spacer between the socket and the chassis. Provide a longer screw and another wide nut along with the original washers to mount the tube socket slightly below the chassis. Do not over tighten and break the ceramic. This increases the path the arc has to jump. Clean off any carbon tracks from old arcs. I also mount the 5V4 with spacers the same way, for cooling air flow from underneath the chassis past the tubes. Route all wires as original to prevent chafing and arcs. Retest all power supply resistances and AC cord leakage to ground before applying power again. If LV and Bias voltages are still OK, we are ready for the modulator test.

NOTE ON CDC MODELS: REMOVE DIODE RCT70 AT TOP OF T4 MOD TRANSFORMER AND REPLACE WITH A WIRE. This is late 1950s thinking which was later debunked, and may damage your mod transformer if you do not remove it. It even appears in radio handbooks of that era. The selenium rectifiers for the PTT circuit are always bad. When they fail, they emit poisonous gas. Replace them if you intend to use the old PTT relay. I recommend that you replace it all with a low voltage system to prevent mike damage.

Test R33 & R34, 22 Ohm non inductive resistors on the plate caps of the 807s. These may be missing because they were not installed during an update later, or damaged due to heat. Don't replace these parts if they are within +/- 20%. Install the 807s. Tune the RF stage up into a dummy load. NEVER OPERATE THE MODULATOR UNLESS THE RF AMP IS PROPERLY LOADED UP IN CW MODE. THIS PUTS A PROPER LOAD ACROSS THE MOD TRANSFORMER. IGNORING THIS IMPORTANT IDEA CAN IMMEDIATELY DESTROY THE MOD TRANSFORMER! With the Plate switch set to OFF, change the mode switch to PHONE. DO NOT MOVE THE CW/PHONE SWITCH WITH THE PLATE SWITCH TURNED ON! This can arc the switch or cause other damage. Test the modulator resting current on the MOD range. Don't touch any of this stuff with the power on! Adjust the clamp tube and 807 bias R13 per the manual for correct resting current with no audio. VERY IMPORTANT! Loosen the screw all the way on R13 before moving it. Do not slide the clamp across the fine wires, or it will cost you $20! The screen voltage on the 807s should be about 300 V in phone mode after all adjustments are finished. You will probably have to touch up the clamp tube adjustment on the RF PA because the adjustments interact.

Install the 6AU6s in the audio preamp/driver. No microphone, audio gain all the way down. If there is an inverse feedback circuit, and you have reversed the plate caps on the 807s, it may oscillate, showing higher MOD current. Immediately turn the HV OFF if it oscillates, to avoid damage. You may be able to fix it by swapping the 807 plate caps. Do not proceed further without correcting this problem. Plug in a microphone. See if you get any modulation. Use a monitor scope or look for kicks in the MOD current. Don't be disappointed if its distorted and does not make 100% modulation. Leave audio repairs and mods til later. The resistors and wax and electrolytic capacitors in this section probably all need replacement.

With AC power disconnected, disconnect 6146 plate caps. Unscrew the other end of the Parasitic Choke assemblies from the plate blocking capacitor C31. Remove the assemblies. Test and replace L11 and L12, 820 Ohm parasitic non inductive resistors, if needed. Just carefully heat and peel the coil off one end of the parasitic chokes. I was surprised the L11 & L12 parasitic resistors were 820 ohms and still good. This is not documented in the manual, I had to read it off the existing good resistors. Most other rigs use 47 ohms or 100 ohms. Failure to address this problem will result in instability or oscillation on 20 meters and above. Also test R35, 56 ohm non inductive resistor to the 6146 screens, which is part of the TVI mods. Early Viking IIs often do not have this fix. Use only older style carbon composition or the OX or OY series shown in the parts list attached. Don't replace these parts if they are good. The originals are better than modern parts. If it measures within +/- 20%, you can reassemble everything.

Add a 0.001 uF 500V ceramic capacitor from the junction of R24 and L6 to ground per AC0OB, to help keep RF and modulation audio out of the clamp tube circuit.

Install R28 on a terminal strip near the 6146s to better support it. If R28 is burned out, order a new 25K 25W. Do NOT reduce R28 in value as some say, or it will damage the 6146s from excessive screen dissipation. Increasing R28 slightly improves modulation linearity. But I later found a better method using two 10W resistors of 40K and 60K and a mix of modulated and unmodulated HV. I will describe that mod in detail later.

I modify the wiring to the CW key for my preferences. I do not like to wear the CW/Phone switch out, just to spot frequency. Cut the jumper on the KEY jack which grounds it when nothing is plugged in. Trace the V6 keying wire to the CW/Phone switch. I generally install a jumper wire, to directly connect the key to the V6 oscillator cathode and VFO keying line in all modes, independent of the setting of the CW/Phone switch. Now if you are in Phone listening, you simply close the CW KEY and adjust the VFO frequency to match the received station to zero beat. This works with a crystal also. See the schematic for details. Watch out for excessive grid current with HV off. See the drawing for details on switch wiring and the schematic.

Add a 2.5 mH RF choke in parallel with C30 loading for safety, and protection of the fixed loading capacitors from failure on modulation peaks driving high impedance antennas without a DC resistive connection. This prevents burn out of C33 – C38 and blows the fuse if HV blocking capacitor C31 fails, preventing shock hazard on the antenna. Later transmitters all include this fix. Use a classic non ferrite core choke with four winding groups to avoid unwanted self resonances.

At this point, you have a working CW transmitter. The stock audio and modulation transformer are proven good for AM operation. This completes the Phase One repairs. However, you may be running it on 125 VAC or possibly more. You will need to build a bucking transformer to get it back to 115 VAC design nominal. If you do not do this, you risk transformer damage and shortened tube life due to excessive filament voltage. Refer to Electric Radio #276, May 2012 for details on building a bucking transformer. See the schematic in this article for a simple setup that will handle your entire station, using modern parts. I found these transformers at ham fests for $5. I will provide and article on this site soon, which you can build using commonly available transformers.

CONGRATULATIONS! You have a working Viking II, with some useful reliability updates to eliminate problems which have shown up in some rigs.

Some other details:

Obtain some grease packets from Palstar to end noise and intermittent contact on the roller inductor, and lengthen its life. This will work on your antenna tuner too. (TNX WA2LXB) Apply sparingly only to the slider rail to lubricate the roller, and run the inductor to max and min positions to distribute a fine film. Do NOT allow any to get on the coil windings. Its conductive.
http://www.palstar.com/en/grease/

Replace the shiny shield with a black IERC shield (as used in R390) on V5 to reduce heat. See ER# 270, p 38, by N0DMS.
https://www.surplussales.com/Tubes-Sock-Acc/TubeShields-1.html

NOTE: Mechanical alignment of zero on the Plate Tuning dial is full capacitance of C29 and maximum inductance of L9 roller inductor. The Viking II maintains circuit Q across all its tuning range, unlike the fixed taps to a bandswitch of others. Do not make a bigger job of this than it is. Just loosen some setscrews and make the adjustment to the capacitor to sync it up with the roller inductor and dial.

I have seen some Viking IIs that have a solder blob shorting out a couple turns of the parasitic chokes near one end, reducing their inductance. It may have been an attempt to get the rig to load on 10 meters; if the above plate tuning alignment is correct, I have not found it necessary to do this on any of the dozens of Viking IIs I have reconditioned.

If the 10 meter tuning is out of range on the plate tuning, slightly stretch the turns of L8 to reduce its inductance and try again.

Neutralization is a simple inexpensive modification will enhance the stability of your Viking II, and make for quicker tune up and QSY. It is covered later on this web page.

TIP #1: The Viking II (with the plate switch OFF) will put out enough RF to the receiver to pin the S meter probably (no harm done by this). If the mike gain is up far enough, regardless of PHONE/CW switch, you may hear audio modulation. This is due to the LV B+ 350V being modulated by the driver stage. It is plate modulating all the RF driver stages. The effect is not noticeable in transmit with the HV+ Plate Switch ON. But if it worries you, turn the AUDIO control all the way down when on CW. I put a switch on the AUDIO control to shut off the filaments on the 807s during extended CW operation to save them from hours of unnecessary operation. NOTE: Do not increase the LV B+ filter capacitors to "cure" this non problem; huge values of filter capacitance may cause chirp, stick with stock or up to 2X stock. The specs on the old capacitors were Guaranteed Minimum Value (GMV) or -20, +80%.

IMPORTANT TIP #2: NEVER, EVER, OPERATE THE MODULATOR IN "PHONE" WITH THE "PLATE" SWITCH ON, UNLESS THE RF FINAL IS TUNED UP INTO A PROPER LOAD! If you make this mistake, the modulator has essentially an open circuit for a load and the mod transformer can be damaged.

TIP #3: BE VERY CAREFUL OF THE 7 PIN SOCKETS WHEN CLEARING OLD COMPONENTS FROM THEM. EXCESS STRESS ON THEM CAN CAUSE THE SOLDER TAB TO BREAK OFF.

TIP #4: Stick with the stock value of grid leak resistor for the 6146s (R24). If you increase this, it is supposed to improve linearity of the RF envelope. Instead, do the RF modulation linearity mod later on this page; that fixes the problem using widely documented ideas. Increasing R24 also will show increased negative grid current in key up, as the 6146s age. This is evident by the grid meter needle pinning left when not in transmit. Leave R24 alone and avoid premature replacement of 6146s. Stock R24 and never exceeding 8 mA operates the pair of 6146s according to the RCA Transmitting Tube Manual. Use the values of grid current specified in the EFJ Viking II manual for maximum life. I read specs and believe them. Pushing the specs for marginal performance changes does not seem to be good economical sense.

TIP #5: Use a bucking transformer as shown in the schematic below to obtain 110 - 115 VAC. The transformer in the parts list is enough to operate your whole vintage AM station in the 100 watt class. Its good for 15 amps.

TIP #6: In the event that you must remove the front panel (not recommended) you should be aware of a fixture which allows the chassis to be placed on your bench, bottom up, without damage to components on the top of the chassis. It consists of four wooden legs bolted to the sides of the chassis via existing holes placed there by EFJ. These assembly stands are used for assembly of factory built and kitted Viking IIs. This photo is courtesy of Bill Braun, K8ZCT.

VIKING II PHASE ONE REPAIRS & BASIC MODS. BILL OF MATERIALS

Source	Value	Qty	Where used
	33 uF 450V 105C 10,000 HOURS	1	C10 LV FILTER
	22 uF 160V 105C 10,000 hours	1	C13 Bias filter
	33 uF 160V 105C 10,000 hours	1	C12 Bias filter
	33 uF 450V 105C 10,000 hours ADD	2	HV filter added (series)
	470K 2 W ADD (see text)	2	HV Equalizing resistor
Mouser P/N 588-D50K20KE	20K 50W adjustable	1	R13
	0.5 uF 600 – 1000 VDC Orange Drop	1	C8, 807 screen bypass
	0.1 uF 600 – 1000 VDC Orange Drop ADD or replace old paper cap	1	C56 Clamp tube filter
	If burned out replace with 25K	1	R28 20K 20W screen
Mouser P/N 588-OY-56-E	56 Ohm 2 W non inductive resistor	1	R35, replaces 4.7 uH Johnson TVI mod B
Mouser P/N 588-OX-22-E	22 Ohm 1 W non inductive resistor	2	R33, 34 22 Ohm 1W 807 plate parasitic
	DIODE 1N4007 various locations	Many	6AL5, PTT, etc
Mouser P/N 588-OX-1.8K-E	1.8K 1W non inductive	1	R24 grid leak 6146s
	SPST TOGGLE SWITCH 10 AMP	1	SW1
	DPST TOGGLE SWITCH 10 AMP	1	SW2
	DIODE ZENER 34 V 5 WATT ADD	1	807 BIAS
	0.202 OHM 1 WATT SHUNT SH1 (0.22 Ohm paralleled by 2.7 Ohm)	1	PLATE METER 500 mA SH 1
	0.404 OHM 1 WATT SHUNT SH2 (0.47 Ohm paralleled by 2.7 Ohm)	1	MOD METER 250 mA SH 2
	0.001 uF 500 V ceramic disc ADD	1	Add Grid bypass 6146
Mouser P/N 576-V130LA20AP	GE V130LA20AP	2	HV, LV spike protection
Mouser P/N 527-CL-90A	2A, 120 OHM Amphenol	1	Thermistor Inrush current Protector, LV supply
	AC socket 2 wire, cable mounted ADD	1	ANT RELAY CONN
Mouser P/N 588-OY-820-E	820 Ohm 2 W Non-inductive	2	Parasitic Suppressor 6146, L11 & L12 if needed
	Eimac Heatsink Plate Cap	2	For 6146
	2000 pF 1.5 KV – 5 KV door knob cap	1	C31, if bad
	2.5 mH 500 mA RF choke air core ADD	1	Parallel C30 shock protect
	P-T156 driver transformer ($16) or Hammond 124B ($50)	1	T3 REPLACEMENT
RECOMMENDED BUCKING TRANSFORMER:	STANCOR P-6379 Transformer TRIAD F-109U OR EQUAL	1	6 or 12VAC 16A bucking

Select the correct wiring to adjust AC to 115V

Install jumper for key in PHONE mode

Install 6146 screen resistors here

Use wooden legs when front panel is off (K8ZCT)

22 Ohm oscillation stoppers (EFJ B MOD)

2.5 mH across loading capacitor

Remove this screw, stop rectifier shorts

New C8 Orange, C56 Green

Add Bypass Here to 6146 Grid Leak

Mount New Filter Caps Here

Install Meter Protect Diodes Here

Add Eimac Heatsink Plate Caps

Add Red MOV to Plate Transformer

Replace broken RF Connector

Bias Circuit Mods

Reliability Improvement Wiring

Study the Johnson factory "A", "B", and grid block keying modifications. I found them very effective compared to popular non-factory mods. Johnson really got the modifications right when they provided revisions for field updating early equipment they had sold previously. Before you do any of the mods on the web, be sure to check that YOUR Viking II has all the factory mods first.

If you have a Viking I or CDC, you will have to compare the schematics and figure out different parts numbering from this article, which is based on the Viking II schematic. I recommend you study the schematics for ALL of the Viking I, II, and CDC variations. Hi-res scan (TNX WA2LXB):
JohnsonVikingIICDCManual.pdf

In the end, you can own a fully updated final version Viking II with all the problems solved by E. F. Johnson factory mods. NEVER dive into third party mods before you get something running stock first.

The Viking I (built from 1949 to 1952) has big differences besides the 4D32 final. The 4D32 final was biased to cutoff.

The Viking I did not include a clamp tube or many of the TVI measures in the Viking II. Early Viking IIs did not have the clamp tube either. The Viking II was manufactured between 1952 and 1957. The 6146 RF stage screen voltage source of the early production Viking II was a second slider on the large power resistor R13. Unlike the 4D32, the RCA 6146 specs grid 2 at 250V MAXIMUM. I have found no documentation on how to adjust the second tap. I speculate that it should be ballparked at 1/3 of the way back from the front panel, and adjusted for almost zero PLT current on the RF stage in key up CW. This would compare to the proper adjustment with a clamp tube. I also speculate that the clamp tube mod was necessary to get some modulated voltage on the screen of the 6146s. Running them off the unmodulated HV B+ would not have provided 100% modulation, because the 6146 needs Plate AND Screen modulation to work right. WARNING: 6146 screen voltage is specified at +200V; do not exceed this.

Factory mod "A" is on the BAMA web site to replace the obsolete circuit with the clamp tube, which EFJ mounts underneath the chassis on the 160 meter coil screws. Just follow the step by step instructions. Check the final work against current EFJ Viking II schematics.
http://bama.edebris.com/manuals/johnson/viking2/screen%20clamp%20regulator

VIKING II that needs the B mod

If your Viking II has an R13 with two taps like this, it needs the A MOD installed. There will be no clamp tube, and you have to provide that. It is rare to encounter a Viking II lacking a clamp tube. Rather than mounting it below chassis like the mod sheet recommends, I make a flat plate with a hole for the 6AQ5 and the adjustment pot, and mount it to the long screws on the HV transformer, near the 807s. The 6AQ5 sticks out the back of the new clamp assembly chassis. I prefer the knob for the pot on the back of the chassis, to keep it away from the 807 plate caps, which are 700 VDC, live whenever the PLATE SWITCH is on (regardless of the position of the PHONE/CW switch.)

1953 – 1957 production has A and B mods done. Bulletin #3-1 is listed on BAMA under the Viking I directory, but it is essential to include very important updates for RF stability to install in your Viking II, regardless of what you decide to do with the audio.

It installs important RF stability components, like 22 ohm resistors on the 807 plates, a 56 ohm resistor in the 6146 screens, a resistor on the bandswitch, and adds C56, a 0.1 uF capacitor to the clamp tube circuit. The capacitor keeps audio out of the clamp tube circuit. The modification also replaces an RF choke of a different style on the VFO socket for better filament voltage. It adds an antenna relay connector (the infamous crystal socket on the back panel).
http://bama.edebris.com/download/johnson/viking1/modification-b/Johnson%20mods.pdf

Check to see that ALL of the mods are done. This modification also changes the audio system in the Viking I and early Viking II from a pentode output 6AU6 to a triode connected 6AU6, as in later Viking IIs. Frequency response is 250 to 3000 cycles with lower distortion. Get the rig running stock and updated now: do NOT get bogged down in audio mods til later.

The B mod link has a scheme using a 6AQ5 driver. This is NOT an EFJ factory recommended mod. I was unable to get this to work properly. But I did not try it again after I installed the 6146 screen resistor mod, which was the real cause of the distorted modulation. Read more on that later in this article.

DO NOT INSTALL THE JOHNSON PTT MOD. Instead, order the D-labs PTT kit with a PC board and simple installation instructions included. This provides low voltage PTT keying that is safe, unlike the EFJ mod which puts more than 100 V on the mike connector. I actually REMOVED the EFJ PTT relay from WA2LXB's CDC model, and replaced it with the DLabs unit. Write to n6tlu@comcast.net or call 269-962-8765. I highly recommend you enjoy other videos he has posted about the Viking II and other vintage gear.
https://www.youtube.com/watch?v=GJ2bpX45XgM

REGARDLESS OF WHICH PTT CIRCUIT YOU USE, I DO NOT RECOMMEND PLUGGING IN THE MIKE WITH THE POWER ON. I USE A STEREO 1/4 INCH TWO CIRCUIT JACK.

I also change the mike jack from a belly button connector to a simple 2 circuit stereo 1/4 inch jack. You will have to find a jack which has enough threads to go all the way through the front panel, the space between the chassis, and the chassis. This will be difficult to get tight, and if possible use a lockwasher on the inside of the chassis between the chassis and the jack body. Wire the AUDIO to the RING. Wire the PTT to the TIP. This connects the PTT LAST as you plug it in. It does not put the relay voltage on the mike as you plug it in.

There is a 2 pin female audio connector found on some Collins and Heathkit gear. I DO NOT RECOMMEND THIS CONNECTOR. I change it to a 4 pin connector in every piece of gear I encounter, for example the Johnson Valiant. The ground is made via the round ferrule that screws the body of the mating connector to the front panel mike connector. The ground connection is made LAST. If the power is on, the PTT current flows through the mike element or any processing board in the base of the D104 G stand. If the factory PTT is still in the rig, this puts hundreds of volts across the D104 crystal element or any audio PC board. It causes instant irreversible damage.

Some people like the 4 pin connector commonly found on CB transceivers, and a lot of ham equipment. I think it is a reliable connector that performs well; it overcomes some of the problems of the 2 pin female connector noted above. Be sure you are able to tighten it properly, with a lockwasher, to prevent it from spinning. The space between the chassis and front panel will not have a nut to hold the connector properly and space it from the front panel. Many people find it necessary to remove the front panel (not recommended) to mount the 4 pin connector.

Larger or commercial designed transmitters employ spark gaps to protect the modulation transformer or other components. The common 100 Watt class AM rigs of this era do not have that feature. Spark gap protection information is not often included in the popular handbooks. Adjustment can be tricky. Now that your Viking II will make 100% modulation, protect the mod transformer.

This simple circuit consists only of a diode string (with sufficient PIV rating) and a 3K 10W resistor. My circuit does not prevent overmodulation. This circuit is not a complicated 3 diode supermodulation circuit, and is not intended to increase asymmetrical upward modulation. There is some controversy about the efficacy of that circuit; see the W8JI comments noted in the references.
https://www.w8ji.com/Johnson%20audio%20mods.htm
"1.) Going to zero-carrier is actually not what causes splatter or excessive bandwidth. The slope of waveform abruptly changing, going in a new direction towards a straight line, causes the signal to get wide. It's really a "Fourier problem", where the rapid change in slope requires high-order harmonics to produce the waveform. 2.) A plate modulated tetrode tube, contrary to what we might assume from casual understanding of plate modulated stages, reaches zero carrier long before the modulated high voltage reaches zero volts. The diode limiter will limit too late, unless the modulated stage uses a hard class-C low-mu triode in the PA."

For the Viking II CDC model ONLY: The RCT70 diode in the CDC Viking II is a mistaken effort to prevent overmodulation, popular in the 50s, as noted above by W8JI. The CDC RCT70 diode not only does not prevent overmodulation, it disconnects the load from the transformer, inviting damage. Remove this CDC model RCT70 diode and replace it with a wire. (IF YOU WANT TO RETAIN THE ORIGINAL EFJ PTT CIRCUIT: Also, the other two Selenium diodes for the PTT relay should be replaced, since they emit a poisonous gas when they fail. Be sure to use a series resistor determined by experiment to produce the correct voltages in the manual. Selenium diodes have a high forward drop, and new silicon diodes only have 0.6V drop.)

The only function my circuit performs is to load the modulation transformer when the voltage on the 6146 plate hits zero or less during overmodulation. The mod transformer is designed to operate into a load approximating the resistance calculated by R = Ep / Ip. In the Viking II, this is approximately 600V / 0.2 A or 3000 ohms. A string of four 1N4007 diodes is used to keep the resistor out of the circuit except in the case of overmodulation. The wattage rating needs only be roughly 10% of the total modulation, for infrequent peaks occurring if the transmitter and audio is properly adjusted. Deliberate frequent overmodulation peaks will cause splatter and should be avoided.

NOTE: BE SURE TO ORIENT THE DIODES CORRECTLY.

Some may argue that the RF final turns off before the instantaneous plate voltage goes to zero, and they are right. It would require a bias voltage to correct this situation. However, that only happens for an instant in the audio waveform before the negative cycle loading circuit starts working. The ceramic disk capacitors across the modulation transformer primary and secondary offer some protection from these high speed transients. Extreme reductions in their values could invite trouble. Some people reduce the value of these capacitors to extend high frequency response, but I do not get too radical with changes. Anyway, in the interest of simplicity, the bias circuit is not included. Read the references and see if you want to skip this item in your mods. I think it can't hurt, its inexpensive, and its far better than doing nothing.

The audio source MUST include a hard limiter (clipper) to prevent overmodulation, but the simple protection shown here is a back up, for those who prefer just a microphone plugged in the front of the rig.

WHY I DID IT THIS WAY - WHAT IS NEGATIVE CYCLE LOADING, AND WHY YOU NEED IT

Once you improve modulator performance in the Viking II or any AM transmitter, you MUST protect the modulation transformer and prevent splatter. This is amazingly simple for a cheap insurance policy. Negative cycle loading keeps a resistive load across the modulation transformer when the plate voltage on the 6146 RF final goes below zero volts. The 6146 becomes a reverse biased diode, and no longer presents the nominal 2.5 K ohm load to the 807s and the modulator transformer. The load becomes infinity, and the voltage spikes can destroy the modulation transformer insulation. One Valiant I worked on had a terminal strip burned up from this problem. Larger well designed transmitters include an arc gap to limit the voltage spikes to prevent damage.

How can you fix this? You need another diode of the OPPOSITE POLARITY to the 6146 RF PA to connect a 2.5 K resistor when the voltage goes negative. Its that simple. This resistor only needs to be capable of absorbing the transient power peak. A 10 Watt resistor is sufficient for a 100 watt class transmitter, assuming occasional 10% overmodulation. The diodes only need to pass about 1 amp, but they need to be rated for PIV high enough not to break down. I used four 1N4007 series connected diodes. This is all you need. If you have a different transmitter, you can calculate the correct resistor from the plate voltage and current, using Ohm's Law.

There is another more complicated way to do this with a bunch of expensive HV zener diodes. It limits on both positive and negative cycles of modulation. I run 120% positive peaks on my class D rig, to give it more punch. The peaks you really need to worry about are the negative ones that hit the base line on the scope, which can cause splatter and possible damage. If your modulator and voice produce 120% positive peaks, that is a good thing, and there is absolutely no reason to limit them.

There was bad information in the late 1950s hand books that used a SERIES diode in the path from the modulation transformer to the RF power amplifier. This actually made the problem WORSE. It also was not necessary, since the RF PA tube was a diode already. Moreover, it did not prevent splatter. Some of these schemes also used an LC filter, often in a Pi network, to limit splatter. These sometimes resulted in resonances that increased peak voltages in the modulator transformer. One expert removes all these components in the Collins KW-1 to prevent mod transformer failure.
K4KYV Overmodulation Indicator and Transformer Protector article:
https://forums.qrz.com/index.php?threads/tech-talks-and-tips-by-k4kyv.525985/page-3

The Viking II CDC schematic shows this obsolete RCT-70, a 300 mA selenium diode at the top of the mod transformer. Remove this diode and replace it with a wire. Do not remove or excessively reduce the values of capacitors across the modulation transformer primary or secondary. They offer some protection from voltage spikes and limit splatter. Electric Radio #110 June 1998 page 34 has a table showing the frequency response for various values of C32 and C46.

When testing frequency response of the entire modulator, NEVER EXCEED 25% MODULATION for more than a few seconds. Use brief pulses of audio to test at 100% modulation. You may damage the 807s or the mod transformer. Be very cautious with frequencies well below or above 1 KHz, as transformer current can get excessive, especially if you are using inverse feedback. These systems are not made for 100% duty cycle. They are rated for intermittent speech, just like your linear amplifier. You can inject sine wave audio at reduced modulation and use a monitor scope to determine when the system frequency response drops off. The stock Viking II modulation transformer is capable of 150 to 200 cycles on the low end to 6 - 8 KHz on the high end. That is a 12 -16 KHz wide signal, which is too broad for night time work, or use on 40 meters near the band edge. Please be a good neighbor and limit it to about 4 KHz audio or 8 KHz channel width maximum.

ANY competent AM station MUST include a PEAK LIMITER in the audio section. The Valiant and Apache provided that, in the clipper/low pass filter section. A lot of people hack this circuit out. You CAN fix the Apache and Valiant to do the job of preventing mod transformer destruction and splatter, without ruining audio response. After the clipper, you must provide extreme low frequency response to prevent the "square waves" from the clipper from tilting. This requires huge coupling capacitors and a good driver transformer. Then use the clipper only as a peak modulation limiter, not a processor. Excessive clipping is distortion, and creates intermodulation distortion in the audio range, but small amounts will not be noticed, except in the protection they confer.

The Viking II does not have a peak limiter. I use the K7DYY D-104 mike processor for $80. It includes a high impedance load for the D-104 element, a phase rotator, a compressor, a peak limiter, and a gain adjustment, all on one circuit board the size of a Triscuit cracker that fits in the base of the G-10 stand for the microphone. It also reduces hum, because it has a low impedance output. I have some minor mods to the DYY board for my operating preferences.
http://www.k7dyy.com/

FIX THE ACTUAL PROBLEM. Messing with the wrong circuit to patch a problem elsewhere makes as much sense as replacing the tires because your car won't stop. Do the brake overhaul it really needs.

The problem is not the driver, its the RF final. The stock connection for V2 works, if the 6146 final screen supply is corrected for linearity. This circuit may be an artifact of the Viking I (if the 4D32 does not have the linearity quirks the 6146 has). Class AB1 for the 807s probably worked for the 4D32. When 4D32s became hard to obtain due to the military buying up all the stock, the switch to 6146s came with unanticipated problems. EFJ had painted itself into a corner by using two 7 pin tubes for the audio driver. They learned from that, and improved the modulator in the Valiant.

CHANGES TO THE 6146 SCREEN CIRCUIT

KC2RLQ suggested deriving the 6146 screen voltage proportionally mixed from the modulated and unmodulated HV to improve linearity of the RF envelope in response to the output of the 807 modulator. There is information on the web that confirms this. When I tried it, this nicely cleaned up distortion of the trapezoid pattern (audio vs RF) in the 0% output portion. You will need two 20W resistors in a 60K and 40K combination. These methods address the "kink" in the characteristic curves of tetrodes at lower screen and plate voltages during modulation. NOTE: This 60K/40K mix does not disturb CW operation, since the parallel value is the same as stock.

I mounted the new R28A and R28B 6146 screen resistors on a terminal strip between the 6146s and R13, to avoid hanging them off SW3, which does not provide very secure mechanical mounting.
If you want to simplify the wiring changes, replace the existing R28 on the PHONE/CW switch with a 40K 20W R28A resistor, and mount the other R28B 60K resistor by one leg to the rear end of the big adjustable resistor R13.

All the work on the audio section will not correct the modulation characteristics of the RF final. The Timtron DX100 and Valiant mods change the 6146 screen supply by raising the series dropping resistor. This partial solution probably would work even better with a split supply resistor scheme like I have done to the Viking II. I no longer have those rigs to try it on, but I would like to hear from anyone who has experimented with this.

This likely is due to the "kink" in the screen characteristics if you look at the curves in the tube manual. Historical note: This kink is the reason that audio tubes were developed, called "kinkless tetrodes" as in the KT-## series of audio output tubes. The mods keep the 6146 OUT of this nonlinear kink region.

I am not alone in asserting this effective remedy for Viking II & DX-100 & 6146 RF final screen circuits for 100% modulation:

• http://billthompson.us/kc2rlq/
• http://www.amwindow.org/tech/htm/dx100.htm
• http://www.qsl.net/ve7khz/VikingII.html
• https://en.wikipedia.org/wiki/Beam_tetrode
• https://www.nostalgickitscentral.com/heath/mods/dx100.html
• http://www.amwindow.org/tech/htm/scrnmod.htm
• https://www.amwindow.org/tech/htm/scrnmodd.htm

Here are the schematic and design formulas from Dan McGorill as posted in various places on the web, for those who want to understand "how" the changes work, and the justification for the values selected. These extra components are needed because the 4D32 and the 6146 are very different animals. The 6146 was marketed as a "high perveance" beam power tetrode. https://www.john-a-harper.com/tubes201/
Also see: https://en.wikipedia.org/wiki/Perveance

NOTE: Be certain the C2 capacitor shown in the drawing is rated for at least 3 KV. This is for the 700V plate voltage PLUS the 700V or more of modulating voltage, PLUS a safety factor. Otherwise, if the capacitor shorts, excessive voltage at high current will be applied to the 6146 screens, damaging the tubes and possibly power supply and other components.

6146 Sceen Resistors Calculator B+ Voltage: V Created by Don, AC2RS Screen Voltage: V Total Screen Current: mA C1 Value: pF R1 Value: _____Ω Calculate R2 Value: _____Ω C2 Value: _____pF Screen Voltage MUST be 200V. Screen Current for 2 tubes should total 22mA. Control Grid current should be set to 6mA for 2 tubes. These values are appropriate for a plate B+ voltage of about 650V. They come direct from the RCA Transmitting Tube Manual. NOTE: Under no circumstances should the Control Grid Current exceed 3mA per tube or 6mA total. The RCA Tube Manual specifies absolute maximum Control Grid Current of 4mA per tube or 8mA total. Tube Life will be dramatically shortened if you exceed these parameters. It is not necessary for good linearity to challenge these grid current ratings. Ignore misguided information on the web which suggests you do so. This technique should work on any transmitter which employs 6146 finals. It improves AM positive modulation peaks and has no adverse effect on CW operation. For example, it works on the DX-100 (which Dan McGorill originally developed it for), the B&W 5100, and the Johnson Valiant (don't forget to compensate for use of three tubes). The Heathkit Apache TX-1 will require more adaptation, due to its unique tune up circuit in the screen of the finals.

The second key change to the Viking II to get 100% modulation is to the dynamic impedance of the HV supply.

In 1949, when the Viking I was created, designers did not use large amounts of filter capacitance in HV supplies because they were expensive, and electrolytic capacitors are unreliable. The DX-100, which copied many parts of Johnson's design used about 50 uF filter caps compared to the Viking II's 8 uF. This was not for ripple reduction. If you analyze the circuit of the secondary winding of the modulator transformer, you find this:

The top of the modulation transformer goes where you would expect it to, to the plates of the 6146s.
The problem is revealed when you look at where the bottom end of the secondary of the modulation transformer goes. It connects to the + side of the HV supply. Any AC current from the modulation transformer secondary must go through the HV supply output filter capacitor to ground to complete the circuit. If a high impedance exists there, it limits the current. It gets worse for lower audio frequencies, as the reactance of an 8 uF capacitor rises to a level that audio is attenuated. This is a major reason a stock DX-100 sounds better and gets fuller modulation than a stock Viking II. The easy fix is to add more capacitance to the output filter of the HV supply. Johnson figured this out, and provided 40 uF for HV filtering on the later Valiant.

The power supply is shared between the RF stage and the audio stage, and the HV sags on modulation peaks, which is why you get 100% modulation on the negative peaks near the base line on the scope when the positive peaks are not quite 100%. Beefing up the HV filter capacitors (within reason) helps with that too. The RF PLATE meter reading shows "downward modulation" in AM. Increasing the HV filters reduces that downward deflection on AM peaks.

I therefore added two series connected 33 uF 450V capacitors, for a net capacitance of 16 uF to my Viking II. There are equalizing resistors (value not critical, but watch the wattage) across each capacitor. The total capacitance is 24 uF.

I do not recommend increasing the value of C9 to extremes above this value. If you read the 5R4 tube specs for surge current and recommended value for filter capacitor, this is near the limit. If you go to extremes, the 5R4s may flash over, which could cause arcing and damage to the plate transformer.
24 uF (8 + 16 uF) is sufficient capacitance to furnish a low reactance compared to the 3000 ohm load of the RF finals matched by the secondary of the mod transformer.
Furthermore, excess capacitance may distort the keying wave form in CW.
If you solid state the 5R4s, AND increase the filter capacitance to insane values, the surge current in the HV choke and transformer can become impressive. It becomes necessary to limit surge currents by placing a thermistor or perhaps a Harbach soft start board in the primary circuit, to step-start the voltage at the beginning of transmit.
Solid stating the 5R4s dramatically increases the plate voltage, requiring a downward adjustment in plate current to keep the RF stage wattage input correct. It further stresses vintage components beyond their design limits. I arced a HV choke in this fashion when I was using solid state 5R4s when that happened. As long as 5R4s are available, I have decided to continue to use them rather than the solid state alternative. It may be possible to simulate a 5R4 with a series resistance and a series Zener, but high voltage high wattage Zeners are expensive and getting hard to find. Maybe you could design a high voltage switching regulator using solid state components. As the available high voltage power supply iron shrinks, that may be a solution that needs study.
This type of extreme mods increase complexity, can reduce reliability, and may provide no worthwhile increase in performance in the exchange. There are some schools of thought that advocate that in the interest of HV supply regulation. KC2RLQ successfully used two 68 uF capacitors in series (34+8=42 uF). The DX-100 uses two 100 uF caps (50+8=58 uF). These are YOUR design decisions to experiment with.

Combined with the 6146 screen resistors mod, this was all that was needed to get 100% modulation.

TIP: Swap the 807 plate caps for best positive peaks for your voice and microphone.

ONLY SIMPLE CHANGES TO THE AUDIO DRIVER ARE NEEDED AT THIS POINT, TO MAKE IT PERFECT

I made only one necessary change to the audio driver circuit. I changed the V1 preamp 6AU6 screen to feed from a voltage divider rather than a simple dropping resistor from the plate supply. This prevents the screen voltage from going higher than the plate supply, with a high value cathode resistor. The high value cathode resistor on V1 is required to prevent clipping with a high output mike like the D-104.

I learned this simple audio trick from work on WA2LXB's Viking II CDC, which had STOCK AUDIO. If you compare the Viking II CDC circuit, a low value cathode resistor is fed from B+ with a 50 K 5 watt resistor. This is required due to the compressor circuit varying the screen voltage on V1 to control the gain. In changing that gain, cathode current varied wildly; this is not the case in a standard Viking II, so the cathode bias stabilization circuit is not required. In a standard Viking II, this circuit can be simplified to the circuit I show, which to eliminate the compression components. It reduces distortion. Check out the Viking II CDC schematic to verify this conclusion.

Choice of two capacitors in the V1 audio stage provides either a "presence rise" for more punch, or a "flat" more mellow sound. Play with it and pick one. Or try a value in between like 1 uF, 4.7 uF. With a D104, it will sound good. With a dynamic mike, you may need the presence rise.

Any attempt to force frequencies lower than 250 cycles through the stock audio transformer will result in harmonics of that low frequency audio appearing, along with intermodulation distortion with mid band audio. This covers everything up with a film of grundge which get worse the more bass boost you add.

Furthermore, I have done tests on the mod transformer that show similar results. It is possible to reduce this by a fancier, much more complicated circuit, which includes the mod transformer and driver transformer within a single inverse feedback loop. Unless you are willing to go all the way with a W1CKI style driver, anything in between is futile.

At this point, with the mods described to the 6146 screens and the HV filtering, you have a decent sounding vintage AM rig. There are some additional tweaks you can do to treat the higher frequencies in the audio.

WARNING: Audio amps have power ratings, based on transformer current and plate dissipation of the output tubes, etc. DO NOT USE A CONTINUOUS TONE TO TEST YOUR MODULATOR AT FULL POWER. The plates of the 807s may glow red or the modulation transformer may overheat or fail, if you use a single tone test for too long. One of the "amplifier tuneup pulsers" could be employed in the audio chain to use full output but reduced duty cycle for testing. Otherwise, do frequency response tests at low levels, 20% of full modulation or less.

NOTE: BE VERY CAREFUL OF THE 7 PIN SOCKETS WHEN CLEARING OLD COMPONENTS FROM THEM. EXCESS STRESS ON THEM CAN CAUSE THE SOLDER TAB TO BREAK OFF.

Viking II RF Linearity, Bias, Audio Mods

AM TRAPEZOID
X AXIS MODULATED B+
Y AXIS RF OUT

The sides of the trapezoid must be a straight line. If there is any bending, it means that the audio modulation waveform does not produce a corresponding linear change in the RF output. This is due to the RF stage, in this case, the screen circuit modifications installed fixed the problem.

There are ways to adjust the high frequency response of the Viking II and other transmitters. This technique is known a "building out" the modulation transformer. This consists of installing small value capacitors across the primary and secondary of the modulation transformer. This filtering reduces harmonic distortion and higher frequency audio components, thereby reducing splatter. (Push pull modulator stages have less even order harmonic distortion than single ended amplifiers.) There are mods on the web that produce excessive bandwidth, splatter, and put the modulation trsnsformer at risk for damage. This information provides a moderate increase in high frequency response without introducing splatter. Playing with these capacitor values requires that you sweep the response well past the normal 10 KHz. Excessively low values of parallel capacitance on a modulation transformer resonate with the windings and can produce a peak out near 100 KHz from the carrier, even with a proper 100% modulation. This can be baffling for the inexperienced ham. I recommend you do not deviate much from the stock values.

There are also circuits in that appeared in many of the handbooks around the mid to late 50s that placed an L/C network (often a variation of a Pi network) in the path between the mod transformer secondary and the RF finals. One rebuilder of the Collins KW-1s removes that network from all he works on, to prevent mod transformer failure. The input impedance of some Pi networks at frequencies outside the normal range can be very high, causing no load to be presented to the mod transformer secondary. Similarly, series diodes like the Viking II CDC has (RCT70) should be removed. They are not necessary, since when the 6146 plate voltage reaches zero, it is already turned off and any diode is only a duplicate. It is important to prevent the plate voltage on the 6146s from approaching zero by peak limiting in the modulator stage. A negative cycle loading circuit as provided on this page provides a load if that condition is not met, protecting the modulation transformer.
K4KYV Overmodulation Indicator and Transformer Protector article:
https://forums.qrz.com/index.php?threads/tech-talks-and-tips-by-k4kyv.525985/page-3

There is a chart in the Electric Radio #110 and #370 page 7 that displays values for C32, C46, and C55 that result in various RF band widths from 5.4KHz (stock), 9 KHz, to 14.8 KHz. (Audio frequency band width is half those figures.)

I recommend you change C55 (across the modulation transformer) from 0.01 uF to 0.005 uF 3 KV CERAMIC. I recommend you leave C32 and C46 alone. This will widen the response enough for intellibiity, while staying safe when working near that band edges on 7295 and 3705.

If you want 9 KHz band width, change C32 and C46 from 0.01 uF to 0.0068 uF 3 KV CERAMIC also. Do NOT use anything but ceramic capacitors for this location, since this is an RF BYPASS. Sprague Orange Drops won't do the job right. Also, very important, dress all capacitors and the small RF choke away from the windings of the large main RF choke. Do NOT install a shield on the 6AL5! Overlooking this simple mechanical precaution will cause arcing and failure of these components. You might reduce the capacitor across the secondary of the stock driver transformer slightly too. Depending on the driver transformer you choose, you should sweep the audio response to test it, and listen to your signal on a remote SDR like the K3FEF web site to verify your work. K3FEF has a feature which displays band width, and you can zoom in to nearly full screen for a standard AM signal.
http://k3fef.com:8901/
A PT156 or 124B will give a wider signal (and require different termination networks across their secondaries.) For the PT156, use a 390 pf in series with 390 ohms across the driver secondary. A series resistor is needed to damp out ringing when the driver tube saturates. Those values may be in the ballpark for the 124B; you will have to experiment.

DO NOT, UNDER ANY CIRCUMSTANCES, REMOVE C55! This is an important component that reduces transient voltages in the mod transformer which might cause failure. If the 807s saturate, ringing can occur, which causes high voltages that may arc over the insulation in the transformer. Where are you going to get a replacement? I changed it from 0.01 uF to 0.005 uF, and recommend you do not go further.

Reducing C32 & C46 to 0.0033 uF results in 15 KHz RF band width. Overdoing changes to these components can also result in unanticipated peaks in audio response out near 100 KHz away from the carrier, due to resonances within the mod transformer. This will be seen as splatter out that far, and get people really annoyed, and may get a Volunteer Monitor's attention.

The following section will cause some to question, and accuse me of heresy in the long standing discussion of Viking II audio. I will be the first to admit the stock audio is an excruciating assault on the auditory nerves. However, I did not want to adopt any solution that included enlarging the 7 pin tube socket holes to fit 9 pin tubes. This limits choices in treatment of the problem.

There are a lot of audio driver mods out there. I found some did not work properly. Not only that, they incorrectly treated the audio distortion and lack of positive peaks as a driver problem. I found the real problem was non linearity as the 6146 RF finals responded to modulating voltage. Correcting the 6146 RF screen source to provide linearity fixed this problem, without major modifications to the audio.

NOTE: The 6AQ5 driver mod shown in the B mod link is not from the EFJ factory. It recommends a 12AX7 requiring a 9 pin socket, making a major amount of work punching a hole in the tight space near the meter. Also, the 6AQ5 did not develop proper undistorted voltage swing on the plate because it shunt fed the driver transformer via a coupling capacitor. A lot of the potential power was wasted in the big resistor to B+. There are suggestions to derive the 6AQ5 plate voltage from the 600V supply, but a triode connected 6AQ5 should not have more than 275 absolute max, according to the RC-24 RCA tube manual. The peak positive plate voltage in a TV vertical deflection driver is 1100 VDC, but at a specific duty cycle which might not be met in audio service. The coupling capacitor would have to be conservatively rated for 1KV, to prevent driver transformer burn out in the event the 6AQ5 filament opens (as it would with normal aging failure). If a standard electrolytic capacitor is used for this purpose, it would be subject to normal aging failure (shorting), which would also ruin the driver transformer. Stuffing a large oil filled coupling cap under the chassis seemed like a challenge. Because of these factors, I did not try the 600V plate supply scheme. You would have been better off punching TWO 9 pin socket holes and installing a 12AU7 driver like the Valiant uses. An even better choice would be the 12B4 (with 1K plate impedance), but that needs a 9 pin socket. With both the parallel connected 12AU7 and 12B4, plate current could get too high for the driver transformer primary and burn it out. You are back to a coupling capacitor scheme like the 6AQ5 mod.

On the surface, you would expect a power tube like a 6AQ5, and a triode, to provide lower impedance drive for the 807 grids. The triode connected 6AQ5 transconductance is 4800 micromhos. It turns out the plate impedance of a triode connected 6AQ5 is 1970 ohms (Sylvania Technical Manual 16th Edition 1979). It might not be low enough to produce undistorted class AB2 operation with the 807s. I evaluated this circuit installed in the Viking II, and abandoned it.

It would be great if a driver transformer that was rated for 60 mA plate current was available. The 6AQ5 driver would not need a coupling capacitor or feed resistor in that case. Heathkit and others used a 12BY7 for a driver, and their transformers might be OK for that, if you have a junker chassis for parts.

The 6C4 triode has a plate impedance of 7.7K, and provides better voltage swing and linearity than the 6AQ5 in tests I performed. The 6C4 plate impedance is not low enough to produce undistorted class AB2 operation from the 807s. However, its transconductance is only 2200 micromhos, and lacked enough gain to work with a variety of D-104 heads. I evaluated this circuit installed in the Viking II, and abandoned it. The 6AU6 transconductance in triode connection is 4800 micromhos, enough difference in gain to make the audio driver work right, if the RF linearity is improved by the 6146 screen grid circuit mod.

The Viking II's triode connected 6AU6 has no rating for plate impedance in the handbooks, but its cousin the triode connected 6AH6 is rated for 3.6K in the GE Essential Characteristic Tube Manual (14th edition). Similarly, the 6BH6 is used in the Viking Mobile as a triode connected driver for 807 modulators. Class AB1 operation of the 807s is sufficient to produce 100% modulation, if the 6146 RF linearity mod is done. Various EFJ advertisements and literature make conflicting statements as to whether the 807s run class AB1 or AB2 in the Viking II. If you want a low impedance driver for the 807 grids, try the 6AH6. NO rewiring is required, but you may need to adjust the biasing. Please let me know how you make out.

A final note: The tube manuals instruct you to tie screen grid, suppressor grid, and plate together to make the triode connection for a 6AU6, with a maximum plate supply voltage of 250 VDC. The Viking II leaves the suppressor grid grounded and supplies B+ to the screen through a 22K screen resistor. I speculate that this nonstandard connection is used to keep screen dissipation within acceptable limits when operating off a +350 VDC supply rather than +250 VDC. I agree that this unique EFJ choice sets them apart from everybody else, but they did a good job on the rest of the Viking II. It also appears that they knew what they were doing here.

Use the simple divider network shown with V1 to stabilize its screen voltage. Use the stock V2 6AU6 driver, with the stock driver transformer. Choose the cathode bypass capacitor on V1 to suit your preferences. The lower value provides a lot of presence rise; the larger value is flat 300 - 3000 cycles. The stock capacitor load across the driver is correct for a stock transformer.

If your driver transformer is burned out, use a series R/C network and a fixed resistor as a load to flatten response and prevent ringing.

Absolutely nothing more is needed, since you have already fixed the real problem in the 6146 screen supply and the HV dynamic impedance.

The AM Trapezoid pattern shows the linearity of an AM plate modulated RF stage.

Viking II RF Linearity, Bias, Audio Mods

AM TRAPEZOID
X AXIS MODULATED B+
Y AXIS RF OUT

Anyway, here are the other audio mods, if you want to give them a whirl. But be sure to do the RF stage linearity and HV impedance mods FIRST.

I am not throwing rocks at other people's work here, just offering reasons why I took a different path. I tried most of these mods myself. I posted these links, with my cautions based on my work, to point you to other resources. You may find them better for your purposes than mine.

One helpful trick working in these close quarters is to use Vector Tube Sockets, which have a center stalk with tie points for components. You can still find these at swap meets and evil-bay. I recommend using tube shields, since the modulator plates nearby can couple signals into the low level circuitry. If you retain the original terminal strips, you may run out of lugs. Drilling more holes in those cramped quarters to mount more terminal strips is not a good idea. Punching out the 7 pin socket holes to 9 pin size can be done without generating metal chips from drilling. And removing the front panel to make the socket holes is a hassle. NOTE: BE VERY CAREFUL OF THE 7 PIN SOCKETS WHEN CLEARING OLD COMPONENTS FROM THEM. EXCESS STRESS ON THEM CAN CAUSE THE SOLDER TAB TO BREAK OFF.

This one is pretty well thought out, but requires the 9 pin 12AX7, followed by a 6C4.
https://www.amwindow.org/tech/htm/viking2.htm

Here is a simple mod with a 6C4 driver.
https://www.amwindow.org/tech/htm/vikinga.htm
I tried this one, and its not too bad either. The RFB and CFB are fed back to a large electrolytic C54, which shorts out all the inverse feedback; that part does not work. Also, the biasing of the 6AH6 can be improved by changing R32 to a voltage divider circuit. This prevents the 6AH6 screen voltage from going above the plate voltage. The cathode resistor in the first stage may be too low, not enough DC bias for the high output D104 mike. The 10 ohm 2 W resistor in the cathodes of the 807 reduces distortion in class AB1. In class AB2, cathode bias is not permitted. The Zener stabilized 807 bias is a good idea, but needs a separate supply resistor, not the stock bias string. C8 screen bypass is a good idea, does not overdo it, and eliminates the need for VR tubes. The mike RF filter C52 - C26 is done correctly; EFJ got that wrong.

You will occasionally see something with a "floating paraphrase" phase inverter. Keep in mind this is a high impedance source that will not work in class AB2. The design has some inherent imbalance problems, and the frequency response of the outputs that drive the grids of the 807s are not equal. As a design idea, it seemed too cute for my taste.
http://www.vias.org/crowhurstba/crowhurst_basic_audio_vol2_063.htm
http://www.r-type.org/articles/art-097.htm
http://www.tubecad.com/2020/10/blog0517.htm
These discussions expose the complicated design problems and limitations of this circuit.
I had a Valiant with similar mods, and removed it. Simple straightforward circuits seemed to work better. This valiant did not have the extreme separate HV supply and transformer lamination changes, so the partial mods were not able to produce distortion free 100% modulation in class AB1 mode. LESSON: Half a mod doesn't work.
However, this clever phase inverter overcomes some of the deficiencies of its cousins, and is worth your study, if you are curious.
http://www.amwindow.org/tech/htm/valiant/audio.htm
http://www.amwindow.org/tech/htm/valiant/valiantmods.htm
BTW, changing the modulator tubes from 807s to 6146s (as used in the B and W 5100) could require a different turns ratio in the mod transformer. Its not just a drop in replacement.

If you are going to modify the audio, and punch 9 pin socket holes, do it right. This one is the best thought it out of any I have seen. Because both the driver and modulation transformer are contained in the feedback loop, very careful attention must be paid to phase relationships in the amplifier:
http://www.w1cki.net/A%20Speech%20Amplifier%20For%20Everyone.htm
I once had a Viking II that had this mod, and its audio was fantastic. I also had a Valiant with something similar, push pull driver. If you are a novice restorer, read the article and note that you will have do some math and understand what you are doing. Don't get in over your head. This time, older, I do not have the energy to punch socket holes, and wanted to stick with a nearly stock approach.

Another recent offering is this one, found in two places:
I wish the poster of this mod had included the info in ER #110, 111, 112 which had a very good push pull driver, using a FET for the phase inverter, so you did not have to punch a third socket hole. You should obtain the ER reprints if you want a really good push pull driver configured exactly for the Viking II.
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwi1ufy2-fXtAhVOrFkKHYtmC50QFjAAegQIARAC&url=https%3A%2F%2Fdocplayer.net%2F61992050-Update-for-viking-ii-cdc-upgrades.html&usg=AOvVaw0LvWWh5znnPRuFfuIOf-Z7
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwi1ufy2-fXtAhVOrFkKHYtmC50QFjACegQIAxAC&url=http%3A%2F%2Famfone.net%2FAmforum%2Findex.php%3Faction%3Ddlattach%3Btopic%3D43061.0%3Battach%3D56226&usg=AOvVaw2TODcYZcqg2Bm7VSOgD0RJ
There is some very useful information there, for instance this:
"Note: One cannot simply replace the first stage amp 6AU6 with a 6AH6 without re-biasing that stage, contrary to some misinformation in the public domain."
"The Door Knob dc blocking and plate coupling capacitor C31 usually requires replacement because of heating and value changes so I usually replace this cap with a HEC 2000 pF 7.5kV unit."
I followed his schematic of the high voltage added capacitance, the RC networks for driver transformer load, and other ideas in my Viking II.
I did not use his single ended driver with the 124E transformer. Neither did this YouTuber (DLabs) during a Viking II repair. I highly recommend you enjoy other videos he has posted about the Viking II and other vintage gear. He ripped it all out and put it back stock, with the correct transformer:
https://www.youtube.com/watch?v=W6h_5dnyzL4
I took issue with the use of the Hammond 124E driver transformer in the single ended circuit, because of this:
The Hammond spec sheet says of the 124 D, E, and F: "Should not be used for single ended applications. They have no gap for DC (plate) current present in Single Ended mode." A further step down wired transformer from a 6AU6 gives away already scarce driver circuit gain, without lowering driving impedance sufficiently to achieve distortion free AB2.
The stock single ended driver transformer is OK for single ended use, since it can pass 10 mA plate current in the primary. If it is burned out, the Hammond 124B or cheaper P-T156 ($17) are nearly identical replacements. They are rated for 10 mA plate current. They are all 1:3 step up (with center tapped secondary). A different termination network is required across the secondary of each of these replacement transformers. The manufacturer does not recommend use of the 124D, E or F in single ended drivers which put DC current in the primary; neither do I.
I did not use the 807 screen regulator, since it complicated the clamp tube operation and was not necessary in a minimal mod. I at one point used VR tubes for that, and later removed them.
NOTE: The RCA Transmitting Tube Manual TT-5 states the 6146 grid #2 voltage must not exceed 250, with 200 V typical for class C.
WA1HLR recommends INCREASING the 6146 screen resistor (correctly reducing the screen voltage) in his DX-100 article (which partially solves the problem). If the 807 grids are regulated, it contributes little to the overall distortion improvement, when the 6146 screen mod is installed, and HV filter caps are increased.

Another simple very effective approach is this approach, shown for the Valiant, but equally applicable for the Viking II:
http://www.motormouthmaul.com/linked/high%20fidelity%20transformer%20modification%20for%20johnson%20viking.pdf
He uses a reverse connected hi fi audio amp output transformer. You could use an external 10 watt audio amp and drive a scrapped push pull audio output transformer connected as a step up to drive the 807 grids. Only 0.2 watt is required for AB2.
A Hammond 117K4 25 volt PA speaker line transformer is a better choice because it is much less step up compared to a hi fi output transformer. A 25 volt to 8 ohm transformer is the closest ratio available.
Use the full primary winding for the 807 grids (COM and 1/4W) and 1W for center tap for the bias connection. Use two 750 ohm, 5 watt swamping resistors from each grid to center tap (bias) for low impedance and flat response, and to use up excess drive power.
Drive the secondary speaker winding with the amplifier. Note that you are using this transformer backwards as a slight step up. Limit the drive to 4 watts and avoid overdriving the 807 grids. USE A PEAK LIMITER IN THE AUDIO CHAIN. A Symmettrix 528E is a nice piece of gear, but it MUST be followed by a peak limiter. Here is a transformer which will provide very low impedance drive for the 807 grids for class AB2 operation:
https://www.hammfg.com/electronics/transformers/audio/117
Of course, the Viking II is not a stand alone rig any more. It is of necessity connected to an audio chain and a small power amplifier. Any small hi fi amp or PA amp will do fine for the driver.
DO NOT OVERDRIVE THE 807 GRIDS OR YOU WILL WRECK THE TUBES.

Another "outboard mod" that is very aggressive is this approach by KC2RLQ. He uses a large solid state amplifier and toroids to step up the voltage to modulate a Viking II used solely as the RF deck:
http://billthompson.us/kc2rlq/

There is a need and a business opportunity lurking here, which would solve this problem. If someone created a universal solid state class AB2 driver board with low impedance output and sufficient voltage swing, they could probably replace the audio driver circuits in the Viking II, DX-100, Johnson Valiant and 500, Collins 32V, or any other vintage AM transmitter. It could even be transformerless, solving yet another problem. Any takers to this challenge?

It is possible to use even a standard Viking II on all the currently available WARC bands (60, 30, 17, and 12 meters). This information shows how to use a slightly modified 122 VFO or a synthesized VFO to have fun on the WARC bands with this vintage classic. Be sure to include the carbon resistor on the grid of the VFO 6AU6 to prevent oscillation.

This is the only correct schematic for the Johnson 250-122 VFO. It includes R55 (22 ohms) directly at the grid of the 6AU6 oscillator tube to stop VHF oscillation above 80 meters. I only use NONINDUCTIVE carbon or OX or OY type 150 ohms for R55. It also includes the complete text and schematics for the grid block keying modification. Don't forget to replace R51 (Chernobyl resistor) with a 5 watt unit. I even replaced the AUDIO pot with a pot including a switch to turn off the 807 filaments when on CW for extended periods.
http://bama.edebris.com/manuals/johnson/vfo122/
http://bama.edebris.com/download/johnson/keyermod/keyermod.pdf

I was surprised to find the standard Viking II will cover all the WARC bands. The basic ranges of the 250-122 VFO are 1.75 – 2 MHz and 7.0 – 7.425. 11 meters can be modified to cover either 5.05 – 5.075 and/or 6.02 – 6.056 MHz with the addition of a mica padder capacitor in parallel with C56. I mounted a miniature toggle switch on the rear panel of the VFO to select 30 meters (100pF) or 17 meters (33 pF). 60 meters (5.3 – 5.405) works using the third harmonics of the existing 1.75 – 1.8 MHz. Be sure to use a frequency counter to get exactly on "center frequency".

I mounted a counter output BNC on the Viking II rear, with a piece of RG-174/U. Remove the shield from a very short bit on the free end and place about 1/4" near the buffer plate tuning capacitor wire. WARNING: This circuit has +350V DC, so insulate it well and use care to avoid shock or damage to equipment.

Changing the VFO coax to RG-62/U allows you to make the cables 5 feet long instead of 3 feet, due to the lower capacitance per foot. This maintains the resonance of L51 to the Johnson design.

If you examine the output waveform on the 1.75 – 2.0 range, you will find that L51 resonates to 80 meters. The resulting scope trace shows the 80 meter harmonic output stronger than the 160 meter fundamental. This gives more drive for 80 and 40 meters. The 40 meter VFO output is tuned to 40 meters, as expected. When the VFO is modified to cover 5 or 6 MHz, reduced output occurs, but L51 tunes very broadly and tripling works. Allow warm up time for frequencies above 40 meters. Quadrupling to 12 meters may not provide sufficient grid current and requires more circuitry.

I did not change the original VR tube or 6AU6 types, as some have recommended, since the grid block keying mod may depend on those components. Furthermore, I found drift acceptable with the original parts. The time sequence element of this modification starts the oscillator immediately (delaying the transmitter output), and holds the VFO "on" long after the key is released, giving similar stability as keeping the VFO running all the time. R202 is set to keep the VFO just below the verge of oscillation in key-up. Adjust the keyer control R202 and the clamp tube R30 per the EFJ manual.

Keying voltage is now about 4 mA at 67 Volts Negative. Use the HamGadgets UKA-3 keyer to protect your solid state keyer:
http://www.hamgadgets.com/UKA-3

Viking II - CDC WARC Tuning Band Band Switch Xtal/VFO Freq. (MHz) 160 1.7 – 2.3 1.8 – 2.0 N/A 2.3 – 3.3 N/A 80 3.3 – 4.5 1.75 – 2.0 60 4.5 – 5.8 1.75 – 1.8 40 5.8 – 8.4 7 – 7.3 30 8.4 – 11.7 5.05 – 5.075 20 11.7 – 16.6 7 – 7.17 17 16.6 – 23 6.0267 – 6.056 15 16.6 – 23 7 – 7.15 12 23 – 30 6.22 – 6.25 10 23 – 30 7 – 7.425

Viking II WARC Tuning FREQ BAND VFO OSC BUF DRIVE FINAL LOAD 1.89 160  1.89 10 40 2 17 3 3.71 80 3.71 60 75 2 17 2 *5.33* 40 1.78 49 8 6 42 3 7.16 40 7.16 87 80 3 56 4 *10.1* 20 5.05 38 14 9 69 5 14.33 20 14.33 82 94 5 80 6 *18.07* 15 6.03 60 57 4 86 6 21.43 15 21.43 82 90 4 91 6 29 10 29  88 80 5 100  7 International SW BC Bands 49 MTR 40 6 MHz 72 48 4 48 4 25 MTR 20 6 MHz 60 60 6 72 5 31 MTR 20 5 MHz 38 14 9 69 5

Comparison of Vintage VFOs FEATURE LAF. HA-90 VF-1 EFJ 122 HA-5 HG-10 DRIFT* Hz 900 1170 240 10 250 1 DIAL TURN 40 meter 70 KHz 270 KHz 25 KHz 50 KHz 30 KHz 1 DIAL TURN 80 meter 140 KHz 300 KHz 25 KHz 50 KHz 35 KHz SPOT? Y Y N Y Y CHIRP? Y Y N (with grid block) ZERO (heterodyne) Y

Here is a table comparing some of the VFOs available. All tests are performed on the 40 meter VFO range, since that is the range most prone to drift. The capacitors are smaller, and the inductor construction is more critical on that range in most designs. Also, the 80 meter range is used on fundamentals. The 40 meter range is multiplied up to 4 times; the drift is also multiplied by 4 on 10 meters. That is the most important consideration in the choice of 40 meters for my test.

NOTE: The VF-1 could not be set to exactly to 7.0 MHz, due to the tuning rate. Due to backlash, the HG-10 could not be set to exactly to 7.0 MHz. DRIFT is defined as the change from 20 minutes warm up to final time of 2 hours. Raw data is provided so you can interpret it differently if you wish. In any event, the stock Johnson 122 and HG-10 are tied as the best of the simple VFOs. The HA-5 heterodyne VFO is better, but is not in the same class due to its complexity.

You can also use a solid state synthesized VFO.
You can also use a solid state analog VFO, with an appropriate driver interface to step up the signal level for tube gear:
http://www.f6hoy.com/johson-vicking/

If you only plan on using your Viking II on 160, 80, and 40 meters you could skip this mod. Above that, the variations in grid current as the Plate Tuning is brought to resonance become increasingly problematic. The rig does not oscillate, but getting it tuned up is a chore. I plan on using mine extensively on the WARC bands and 10 meters AM once the sunspots return.

Some of this information first appeared in ER #367. Request a reprint for the full article. I have received email inquiries and this is to supplement the article and expand on it with updated information.

BACKGROUND INFORMATION: The DX-100 and Apache had to add neutralization as a fix in the field. The Valiant and other better rigs included it in the original design. Sadly, many commercial power amps have no neutralization these days. The best article ever general article explaining the theory behind neutralizaton is from Electric Radio Magazine Jan 2014 ER#296. Get reprints at ermag.com or write ray@ermag.com.

Here is a link to W8JI's advice about why neutralization is essential: http://www.w8ji.com/neutralizing__amplifier.htm

Here are tutorials on RF amplifier tubes:

• www.cpii.com/docs/related/22/C&F5Web.pdf
• http://www.vias.org/basicradio/basic_radio_28_04.html
• http://www.rfampguy.com/linear101.html

Care and Feeding of Power Tetrodes full book download:
http://www.tubebooks.org/Books/Atwood/Eimac%201951%20Care%20and%20Feeding%20of%20Power%20Tetrodes.pdf

Full description of home brew 6146 amp with neutralizing: http://faculty.frostburg.edu/phys/latta/ee/6146amp/6146amp.html

Navy report on SSB and RF amps including neutralization: www.navy-radio.com/manuals/ssb_report.pdf

The 6146 is a well screened tube which actually works without neutralization in most rigs, such as the Viking II, Ranger, DX-40, B&W 5100, DX-100 or the early Apache. Why bother with the extra circuit complexity? Under most conditions with a low impedance antenna like a dipole, the final will not self oscillate near the intended frequency. With highly reactive loads such as when adjusting an antenna tuner into open wire feeders or an end fed wire, it may not be so happy.

Neutralization stabilizes the grid current as the Plate Tuning is varied. This becomes very troublesome on 20 meters and up, though in my Viking II, it was clearly evident even on 40 meters. Grid tuning and drive adjustment required frequent touch up. My Viking II grid current swung above and below nominal by 2 mA in both directions. The 6146s do not tolerate above 8 mA drive, per the RCA specs, or the lifetime will be reduced. It is easy to exceed this during normal tuning and use, without neutralization. After neutralization, the grid current shows a slight 0.5 mA peak when the Plate Tuning is exactly at resonance. This makes QSY much easier, on a rig with a lot of knobs.

THE MODIFICATION: Proper neutralization can be achieved with the addition of only 3 components, with minimal expense. Refer to the schematic for the circuit. The only other components necessary are a common 750 pF 500 Volt mica capacitor and a standard 2.5 mH RF choke. The C23 bypass capacitors in the 6AQ5 buffer stage would be very difficult to access to change entirely. I simply lifted the ground end and inserted the new 750 pF 500 V mica capacitor, C*, in series. See photos for installation detail. You may have to disconnect one end of certain components and swing some of the wiring out of the way to get access. The new RFC* choke is inserted in series with the existing VHF choke L20 on the B+ end.

If you can find a 5 KV rated wide spaced capacitor of about 10 - 15 pF maximum for a reasonable price, you can do a nicer job than the one shown here using highly insulated wires. The "GIMMICK" neutralizing capacitor, CN*, can be fabricated from 5 KV test probe wire, some insulated #14 solid house wiring, and a piece of shrink sleeve. It mounts to C29, the Plate Tuning capacitor. Two solder lugs and nuts mount the 5 KV wire "gimmick" capacitor to C29 at both ends. The shrink tube holds the #14 building wire parallel to the 5KV wire. There is no direct connection on this end of the #14 building wire. The free end of the #14 wire solders to the topmost terminal on L5, as shown in Figure 2. Increase the CN* capacitance by moving the #14 wire closer to C29 stator rail. Reduce CN* capacitance by rotating it away from C29 stator, or shortening the stiff #14 wire. Temporarily sliding the #14 wire out of the shrink sleeve will get you in the ball park for a lower value. Be sure to keep the exposed #14 wire end away from C29 to avoid arcing on modulation peaks. The final B+ is lethal. The buffer is also at +350 VDC. DO NOT MAKE ANY OF THESE ADJUSTMENTS WITH THE POWER ON. DISCHARGE ALL FILTER CAPACITORS.This neutralization modification has been successfully implemented on several Viking IIs, by myself and others I have corresponded with to help them do it.

Proper neutralization adjustment can be done by using the following rules:
Turn all power off when adjusting CN*!
1. If grid current rises, as Plate Tuning goes HIGHER in frequency, INCREASE CN*.
2. If grid current rises as Plate Tuning goes LOWER in frequency, REDUCE CN*.
3. Ideal setting is a slight peak in grid current at Plate Tuning resonance (maximum output, Plate current dip).
4. The goal is to reduce the excursion in grid current as the Plate Tuning is tuned thru resonance.

What makes it work? The purpose of C* and RFC* is to raise the end of the grid tuning network above ground slightly to allow injection of inverse feed back of the proper phase around the RF final amplifier. The neutralizing capacitor will not work properly without this measure. Compare the schematic of the Johnson Valiant to verify this is the proper method, since this is exactly the same circuit.

One Other Detail for Stability: When working on vintage transmitters, I often find the resistor in the parasitic chokes on the finals are well out of tolerance. If they are too high, they can actually promote instability, especially on the higher bands. Disconnect the 6146 plate caps. Unscrew the lug from C31 and remove the parasitic choke assemblies from the transmitter. Carefully remove one end of the parasitic coil L11 or L12 to isolate the resistor for testing. The resistor value used in the Viking II is 820 ohms. This is not documented in the manual, I had to read it off the existing good resistors. This is a wide departure from the common values of 47 or 100 ohms in other rigs. If it measures within +/- 20%, you can reassemble everything. To replace the resistor, use only non-inductive resistors like the original carbon composition ones, if you can find them. A modern replacement is the OX or OY style non inductive resistors available from Mouser.

NOTE: Mechanical alignment of zero on the dial is full capacitance of C29 and maximum inductance of L9 roller inductor. The Viking II maintains circuit Q across all its tuning range, unlike the fixed taps to a bandswitch of others.

I have seen some Viking IIs that have a solder blob shorting out a couple turns of the parasitic chokes near one end, reducing their inductance. It may have been an attempt to get the rig to load on 10 meters; if the above plate tuning alignment is correct, I have not found it necessary to do this on any of the dozens of Viking IIs I have reconditioned.

If the 10 meter tuning is out of range on the plate tuning, slightly stretch the turns of L8 to reduce its inductance and try again.

Neutralization is a simple inexpensive modification will enhance the stability of your Viking II, and make for quicker tune up and QSY.

A beautiful solution using a real capacitor rather than my economical "gimmick" capacitor has been sent from Bill Braun, K8ZCT. He has agreed for the information to be posted here, if you want to use that method. He has found a commercial source where you can order the exact capacitor needed, still in production. You cannot use a receiving grade capacitor for this job; you need at least the spacing of the final Plate Tuning capacitor. This capacitor plate spacing provides sufficient air insulation for the 650V B+, plus the RF+Modulation AC signal that appears on one side of the neutralizing capacitor; a safety factor of at least 1.5 should be used. Note that a shield extension is necessary to prevent final RF from coupling to the oscillator circuit.

K8ZCT Neutralizng Cap

K8ZCT Neutralizng Cap

AB2RA Gimmick Neut Cap (see text)

Added C*

Added RFC*

Neutralizing Schematic

From Bill Braun, K8ZCT: Contact: Steven Elliott at Oren Elliott Products http://www.orenelliottproducts.com/
The 80APL10DE is a 3.5 - 9.7 pf Capacitor.
This capacitor was in stock, and I believe is a good replica. The APL air gap is 0.80 and the Valiant was measured and calculated to be almost 0.10. close replica of the Johnson Valiant neutralizing capacitor.
The red insulator is
GPO-3, 1/8" thick fiberglass reinforced insulating board, and cut a piece 6" High x 2- 1/2 " Wide

If you only plan to use your Viking II on phone, you do not need to do this mod. I was surprised on how many Viking IIs I have worked on that have obviously had this mod, and it was later removed. Here is the reason: when the factory grid block keying mod is installed, the OSC and BUF meter positions may not be helpful in tuning up the exciter. There are some bands which will not show resonance. You will have to use your receiver S meter to adjust the oscillator and driver tuning. I personally like the improvement enough to employ it, even with this obvious deficiency. IF YOU DON'T LIKE USING THE RECEIVER S METER TO TUNE THE EXCITER STAGES, DON'T DO THIS MOD.

Keep in mind that this only works with a 122 VFO that has also been modified for grid block keying. Also, the tuneup of the grid drive is a little quirky (but manageable) with grid block keying as compared to standard cathode keying. The stock cathode keyed Viking II with the matching VFO does not click as bad as a DX-100, and I have had zero complaints when I was living within walking distance downtown from 3 other hams. Modern standards and Volunteer Monitors are more critical of these matters, and I want to be a good neighbor. The keying wave forms shown here are not available anywhere else. Complete photos of the original EFJ kit clone I built are shown here. The step by step Grid Block keying instructions are available on the BAMA web site.
http://bama.edebris.com/download/johnson/viking1/mods/vik1mods.pdf

METER SWITCH	+ Wire Color	- Wire Color	Shunt Number
MOD	VIOLET	YELLOW	2
PLT	BLACK (GND)	YELLOW	1
GRID	WHITE	ORANGE	4
BUF	GRAY	BLACK (GND)	5
OSC	BLUE	BROWN	3

This is the only correct schematic for the Johnson 250-122 VFO. It includes R55 (22 ohms) directly at the grid of the 6AU6 oscillator tube to stop VHF oscillation above 80 meters. I only use NONINDUCTIVE carbon or OX or OY type 150 ohms for R55. It also includes the complete text and schematics for the grid block keying modification. Don't forget to replace R51 (Chernobyl resistor) with a 5 watt unit. I even replaced the AUDIO pot with a pot including a switch to turn off the 807 filaments when on CW for extended periods.
http://bama.edebris.com/manuals/johnson/vfo122/
http://bama.edebris.com/download/johnson/keyermod/keyermod.pdf

For a very complete explanation of how the grid block keying circuit works, see:
https://www.frostburg.edu/personal/latta/ee/ranger/schematic/diffkeying/rangerkeying.html

Changing the VFO coax to RG-62/U allows you to make the cables 5 feet long instead of 3 feet, due to the lower capacitance per foot. This maintains the resonance of L51 to the Johnson design.

If you examine the output waveform on the 1.75 – 2.0 range, you will find that L51 resonates to 80 meters. The resulting scope trace shows the 80 meter harmonic output stronger than the 160 meter fundamental. This gives more drive for 80 and 40 meters. The 40 meter VFO output is tuned to 40 meters, as expected. When the VFO is modified to cover 5 or 6 MHz, reduced output occurs, but L51 tunes very broadly and tripling works. Allow warm up time for frequencies above 40 meters. Quadrupling to 12 meters may not provide sufficient grid current and requires more circuitry.

I did not change the original VR tube or 6AU6 types, as some have recommended, since the grid block keying mod may depend on those components. Furthermore, I found drift acceptable with the original parts. The time sequence element of this modification starts the oscillator immediately (delaying the transmitter output), and holds the VFO "on" long after the key is released, giving similar stability as keeping the VFO running all the time. R202 is set to keep the VFO just below the verge of oscillation in key-up. Adjust the keyer control R202 and the clamp tube R30 per the EFJ manual.

Keying voltage is now about 4 mA at 67 Volts Negative. Use the HamGadgets UKA-3 keyer to protect your solid state keyer:
http://www.hamgadgets.com/UKA-3

KEYING WAVE FORMS WITH GRID BLOCK KEYING MODIFICATION (CLICK TO ENLARGE)

RF waveform of about 5 mS rise and fall times.

Horizontal time base: 5 mS/div.
Upper trace: 6146 grids RF.
Lower trace: Key closure.
Increase C205 for longer rise and fall times.

Time Sequence Keying Operation
with RF output and VFO operation.

Timebase: 10 mS/div, Ext. triggered on Key closure.
Upper trace: 6146 grids RF.
Lower trace: VFO.
Ignore level shifts (inadvertently left on DC coupling)
Note: VFO starts before RF output, and stays on after output stops,
   to eliminate chirp.
Increase C205 for longer rise and fall times.

Output keying waveform —
shows good rise and fall times.

Keying waveform on transmitter output 100 watts
   (uncalibrated time base).
Note droop from HV power supply, which does not affect key clicks.
Increase C205 for longer rise and fall times.

CONSTRUCTION OF KEYER CHASSIS AND INSTALLATION (CLICK TO ENLARGE)

Grid Block Keyer Chassis (closeup)

Grid Block Keyer Chassis (w/wires)

Keyer Chassis Mounting

The first scope photo shows that the VFO signal at the 6AU6 grid comes on before, and persists long after, the signal at the 6146 grids. This means the VFO starts and stabilizes (no chirp) during times the 6146 is cut off from transmitting.
The second scope photo shows that the key closure is slightly before the RF at the 6146 grids, giving time for the oscillator to start. The RF waveform is the same length as the key closure. Edges on the 6146 grid are rounded to avoid key clicks.
The monitor scope wave form shows the actual output to the antenna has keying rise and fall times which meet modern standards to eliminate key clicks. The power supply has some droop due to DC regulation. This is not discernable to the listening station. Excessive changes in the values of the Viking II filter capacitors will have adverse effects if pushed radically beyond the values I suggest in these modifications.
Keying speed is approximately 15 WPM.

Next↓Top⇑

REGULATING THE 807 SCREEN CIRCUIT

PLEASE NOTE: This type of modification shows up elsewhere. Often serious errors are included, such as driving the VR tubes off the R13 tap or returning the VR tubes directly to ground (which causes errors in plate current readings). I tried it and abandoned it. With the other measures shown on this web page, this mod is not necessary for good performance. Nevertheless, I am offering it here if you want to pursue it. Please do not write me about this mod, as I do not plan to support it. You are on your own, and maybe you can find something you can use for your project.

The 807 screen circuit changes are tricky, as observed by previous ER authors. The 807 screens must only be powered in the following condition: 1. AM mode ONLY 2. PLATE SWITCH ON. Much of the information available does not make that clear. It often does not specifically describe the switching necessary to accomplish that goal. This article and its schematic will make that wiring change explicit.

If the screens have +300 V on them without plate voltage, screen current is excessive and it destroys the 807s. Also, you do not want to activate the modulator in CW mode. Some authors suggest using the +LV supply (which exceeds the 300V rating for the 807 screen); if you do this, you MUST provide a third set of contacts on the PTT relay to keep the voltage OFF when in CW, or AM and not transmitting. The LV supply can be converted to choke input to lower the voltage, but then the grid drive on higher bands may be inadequate, and it can also affect the speech amp available voltage swing. A separate screen regulator with proper switching as shown here avoids all those consequences.

Another problem is often overlooked in modulator improvement articles. Removing the 807 screen current from R13 upsets the clamp tube circuit, and can place excessive voltage on the 6146 screens, often around +300V. The maximum rating for 6146 screens is +200 V per the RCA transmitting tube manual TT4. Fortunately, this can be corrected by adjustment of R13 to provide the correct voltage. Using the switching I describe, the clamp tube adjustment and the 807 resting current do not interact any more.

Some propose using R13 for a dropping resistor for the 807 screen regulators. There is no apparent switching in this case to allow CW operation. Another problem with this approach is that the clamp tube requires R13 to operate properly. In addition, CW requires R13 as a bleeder resistor for the HV during key up conditions to prevent the HV from soaring; it also would put a huge leading edge spike on the keying waveform from the charge in C9. Furthermore, it introduces a safety problem by eliminating the bleeder current function for C9. I avoided this by leaving R13 as originally purposed for the bleeder and clamp tube circuits, and add a dedicated (inexpensive) 22K 10 watt to operate the separate 807 screen regulator.

These changes install two OA2 VR tubes to obtain +300 V for the 807 screens. That is the recommended voltage for the 807 in the tube manuals. This voltage was originally provided by R13 on the slider tap. NOTE: The lower OA2 cathode goes to the -HV return of SH1 located at the front panel end of R13, as shown in the schematic, NOT to ground, to avoid metering problems. The PLT current will not read correctly if the VR tube current flows through the SH1 meter shunt, complicating clamp tube adjustment and tuning. This annoying problem is ignored in some modification source material.

I chose not to use a solid state regulator. The high voltage zener diodes can be hard to find, expensive, and subject to drift with heat of operation. OA2s are commonly available and inexpensive at swap meets or on line. A simple flat mounting bracket can hold them in a location above chassis where they are easy to see or change. I mounted them horizontally to the top side on the protruding T4 screws on a flat piece of metal, pointing to the rear of the Viking II. This avoids extra holes in the chassis and metal chips getting in the rig.

Note when tracing wires that SW3D is only a tie point and does not switch anything. The new switching circuitry can be provided by the original SW2 PLT and SW3 mode switch in the Viking II without the use of any added relay contacts. A small terminal strip can be added to one end of R13 to hold the dropping resistor for the OA2s and wiring to the 807 screens. NOTE: NEVER operate the transmitter without the OA2s installed, or the 807 screen voltage will be excessive.

The original C8 and an additional 4.7 uF 450 V capacitor on the 807 screens is required to filter any minor audio variations. Some have reported VR tubes acting as a "relaxation oscillator" with the VR tubes blinking on and off and the 807 current bouncing with no audio. I did not observe this problem, but you can add a 330 ohm 1 watt resistor in series between the two VR tubes if this happens. My VR tubes light immediately, and stay lit regardless of modulation level, varying in intensity with modulation. A 150K resistor is used in parallel with the top OA2 in the string to ensure the lower VR tube fires. A new 22K 10W dropping resistor provides the current for the VR tubes and 807 screens. The 807 screen current no longer comes from R13 tap.

In this scheme, the 807 resting current is not adjustable (as in the DX-100) but anything from 30 to 65 mA seems to work fine. The original 807 negative bias string is abandoned and replaced with an "R17" replacement 4.7K 1 watt resistor and a zener diode. NTE 147A (33 V 1 W) seems like a possible choice. The nominal handbook 807 grid bias is 32 V for 300 V screen voltage and 600 V on the plate. Any adjustment would consist of changing the value of the new grid bias zener to a lower voltage and adding series zener diode(s), or series connected 1N4007s to move it in 0.6V steps. I wound up having to improvise the zener diode from a group of series connected 10 and 12 volt units in my junk box. A 2uF 160V (observe polarity!) to ground at the terminal strip near the 807s and the driver transformer provides additional stability to the 807 grid bias. This makes the 807 negative bias independent of any other drains on the bias string, including 6146 grid current. The original bias resistors are probably all out of spec, so why not replace them anyway? The SH4 meter shunt for 6146 grid bias is wire wound, so if it tests good, leave it alone. R24 grid leak resistor for the 6146s MUST be non inductive old style carbon composition or OX or OY style modern non inductive ONLY; if it tests good, leave it alone. See the bias modification schematic for details.

NOTE: Do NOT remove the 100 ohm control grid resistors on the 807s; these suppress the tendency to oscillate. They serve a similar function to the 22 ohm resistors on the 6146 plate caps. Use only non inductive resistors for all these locations.

This 807 circuit, when properly driven, 100% modulates the 6146s running at 230 mA PLT current for full 100 watt carrier output. It does so without exceeding the 300 V screen voltage spec for the 807s. (Some advocate running the 807 screens at +400 V and adjusting the negative bias to get 55 mA MOD current to obtain 100% modulation with inferior speech amp drivers.) I prefer to observe the tube specs and properly fix the real problems with the 6146 screen voltage, the HV dynamic impedance, and minor changes to the driver circuit.

I also provided a 27K ½ watt resistor to keep the current draw from the bias supply the same as before. It provides the original return impedance for the 6146 grid current (which tries to drive current INTO the bias supply).

807 Regulator Mods

PHONE/CW SWITCH WIRING

Next↓Top⇑

WA1HLR, TIMTRON'S NOTES ON THE VIKING II

WA1HLR, TIMTRON'S NOTES ON THE VIKING II, UNEDITED: Get ready for some heavy duty drilling and blasting!

Many of the ideas here have germinated into my present Viking II. Some of them are pretty agressive mods. Keep in mind that a lifetime in broadcasting and a mulititude of projects are the basis for this information. WA1HLR has been a lifetime advocate for the AM mode and vintage hollow state radio, and has offered free help to many seeking to enjoy amateur radio in that way. I found this apparent transcript of one of his legendary old buzzard transmissions and thought it appropriate to include it here.
http://amfone.net/Amforum/index.php?topic=34647.25

OOOOOla, After following the threads about the Viking II i felt I must interject. The HV plate transformer in the Viking1&II is a piece of crap. It has porr voltage regulation and runs hot. I have a number of Viking II plate transformers that I will never use. If you are interested I could part with one. There are better pieces.

-----------------Overview--Viking I & II -----------------

The Viking I and II is the best thought out in the RF department. It is almost general coverage as it is. Yep! 17 and 12 meters on AM . The rest of the transmitter is in great need of fixing:

Power supplies. Solid stating of ALL supplies is a must. Why screw with lossy heat producing vacuuuuuuum tube rectifiers. Biass supply: Get rid of the 6AL5. Use two 1kv PIV diodes . Go choke inpoot filter. ( Eliminate the first cap at rectifier anodes. Use existing filter reactor. Use at least a 100Mfd 150 volt cap for filtering the Biass. The outpoot voltage will be in the order of -80 -90 volts or so.This is fine. Next, the LV supply; Solid state this one as well. Use 3Kv worth of diode per leg. Set this one up for choke inpoot as well. If left cap inpoot ,the B+ will end up being about 370 volts,A bit much! Going choke inpoot will drop the voltage to the order of 250 volts. This will be a lot easier on various components and circuit considerations. A low drive issue may be on 10 and 15 meters if operation there is contemplated. I used a 6BF5 instead of a 6AQ5. The pin out is the same. It is a drop in. The filter cap should be at least 200Mfd@450 volts used in the low voltage supply. The HV supply must be solid stated as well a pair of K2AW type diodes ,at least 6-8Kv PIV rating to be used or two strings of 1Kv diodes mounted on a terminal board works well too. With the stock plate transformer the outpoot voltage will be about 700VDC. The 8MFD oil cap is barely sufficient for the job . Jim WD5JKOs suggestion to use a pair of 330 Mfd caps in series is the best way to go. In terms of plate voltage anything from 500-800 volts will work. Obviously the 800 volt level will make about 125 watts outpoot. Whereas 500 volts will make about 75-80 watts out. In the grand scheme of things the difference in outpoot power will be hard to notice at the other end. A machine tool transformer as suggested will work well . A 480 volt secondary with a bridge rectifier and cap inpoot filter will deliver about 625-650 volts under load. I noticed a discussion of a transformer that was rated @ 575 volts. This would deliver 500 volts under load choke inpoot bridge rectified. You have many options here.

---------------AUDIO--------------------

Probably the weakest part of the Viking I & II is the modulator. The speech amp has barely enough gain when using a good D-10-4 microphonium. The audio quality is like that of two dixie cups and a string. There are many modification articles to fix the problems more or less. Problem is they all involve drilling and blasting and replacement of at least the driver transformer.

The mod transformer can be a weak link as well. The existing circuit can be improved by doing the following : K.i.s.s. mod.( Keep It Simple Stupid) If you are using a preamp /EQ and good microphonium . Low gain of the system is not an issue. Replace the carbon resistors used on first stage of audio,6AU6, After 60 years or so the values can change radically. Use at least a .2Mfd screen bypass cap. Use at least a .01 cap for the coupling cap to microphonium gain pot. The cathode bypass cap should be at least 25 Mfd@ 10 volts or so . Don't forget the plate de-coupling cap. A.1 mfd cap is used. This should be changed to at least a 10 MFD 450 volt cap. Change the second 6AU6 to a 6AQ5. The pinout is very similar. The cathode pin of the 6AQ5 is what was the suppressor grid pin on the 6AU6. The cathode pin of the 6AU6 is one of the two pins used for the grid of the 6AQ5. This makes it simple. The cathode resistor can be of the same value but a two watt rating for safety. Use at least a 50Mfd 25 volt bypass cap. The 6AQ5 is triode connected by connecting the screen to the plate through a 4.7K resistor,value very non critical. This will at least open up the audio response of the speech amp without getting too carried away.

Not so K.i.s.s. mod. If more gain is needed to use a D-10-4 microphonium do the following: Follow instructions for the upgrade of the first 6AU6 but change the 1Meg grid resistor to at least a 10 meg resistor. This change should be done to ANY transmitter when a D-10-4 or other peizo-electric microphonium is used. Relocate the goose tube(6AQ5) from where it is located to the tube socket vacated by the 6AL5. Place another 6AQ5 in the socket and wire it according to the K.I.S.S mod. The second 6AU6 becomes a resistance coupled stage. Simply use a 47K 1watt resistor from the LV B+ to the plate of the 6AU6. The 22K screen resistor is connected to the plate. A .01-.05 cap connects the plate of the second 6AU6 to the grid of the 6AQ5 audio driver. A 100K resistor is connected from the plate of the 6AQ5 to the plate of the second 6AU6. This is a loop of negative feedback around the audio driver. The cathode resistor of the second 6AU6 should be 1K unbypassed. This can be a point to wrap a loop of negative feedback around the modulator. This mod with negative feedback around the modulator still has plenty of gain with a D10-4 microphonium.

--- Modulator addendum---

It is important to place a 10-15K 1 watt resistor in parallel with the secondary of the driver transformer. This provides a much steadier load impedance presented to the audio driver. When the grid voltage swing does not drive the mod tubes into grid current the driver transformer is looking into an infinite impedance. Some audio clipping distortion will be noticeable as one leans into to audio. There are also phase shifts that take place near the ultrasonic audio range that will cause the modulator go into parasitic oscillation when negative feedback is placed around the modulator. I introduced negative feedback with two 470K 2 watt resistors directly from the secondary of the mod transformer to the cathode of the second 6AU6. Adding negative feedback is the very last step to be considered before calling it done. If the modulator takes off reverse the modulator tube plate caps. This is what happens with positive feedback. The biass and screen voltages are very poorly regulated in the Viking I & II. This must be fixed. First of all deal with the screen voltage. In one Viking II I did up 20 years ago I used the choke inpooted low voltage supply(250 volts) through a set of relay contacts as part of the PTT circuit I installed. This voltage also feeds the screen of the goose tube. This required a biass voltage of about -24 volts or so. I used a string of 6 Volt radio shack zener diodes. I also placed a 100 Mfd or so 50 volt cap across the zener string. The only disconcerting thing to someone that does this mod it will be noticed that the PA plate current indication will take a negative swing. THIS IS NOT REALLY HAPPENING. Being that the modulator cathode return is tied to the HV B- rail and the meter shunt for the PA stage is from chassis ground to the B- rail. The screen current drawn from the modulator tubes is referenced to chassis ground . This varying current is in opposition to the current flow through the PA plate shunt and there fore causes the meter to swing in a negative direction. To anally retentive people this is a problem. For those that understand it is a non-issue. The usual 20K 50 watt bleeder with the slider tap has always been a failure point in the Viking I & II transmitters. A good stiff voltage source for the screens of the modulators is just as important as a stiff biass supply. There is an alternative. In my collection of gear I have a practically mint Viking II CDC . I was very hesitant to do any rip tear mods. The 20K bleeder was in perfect shape. I simply placed a pair of VR150s in the sockets vacated by the 5R4s I did not even have to reset the tap, It worked perfectly. One thing though. A 10 MFD 450 volt cap must be place across the VR tubes.A 500 ohm 2 watt resistor is placed in series with the VR tubes to prevent the possibility of any relaxation oscillation effect that can occur with gaseous regulator tubes. Zener diodes are free from this effect. See AM Window- Valiant mods I authored.

In the case of your Viking II ,you are free to make some bigger changes. Move the modulator tubes to the socket holes used by the 5R4s This opens up space for a much larger and better modulation transformer, IE Stancor A 3894 etc.

-------------------------PA Stage---------------------------

The 6146s are not neccessarily operating in class C. The biass voltage from the biass divider string goes through a low value resistor to the grid RF choke The value should be 4.7-5.6K to provide more grid leak biass. This improves modulation linearity. The screen dropping resistance is 20K .This should be 30-35K depending on plate voltage. If the replacement plate transformer delivers 500 volts,20 K is OK. If a DX 100 plate transformer is used.(800+volts) 35-40K is needed. The disc ceramic bypass caps used at the base of the plate FR choke are .01. This really impacts high audio frequency response. They should be pared down to 1000-3000Pf 3Kv units.

---------------------------Addendum-------------PTT----------------

A PTT circuit is very useful. There are a number of different schemes. 4 poles of relay should be contemplated. One pole to key the exciter, another for the plate transformer and Keyed AC for antenna changeover relay. Another pole for modulator screen voltage and the last pole for auxillary equipment control. RX muting etc. With the solid stating of the power supplies this leave two 5 v philament windings. Seriesing up of these windings and use of a bride rectifier with 4700MFD filter cap will give you 12-14 volts DC for 12V relays. If a higher voltage keying system is contemplated. The biass supply will work well. A 240 AC coil relay should pull in @ 70-90 volts DC . A pair of 120 DC coil relays with the coils in series should work. If somewhat more voltage is needed simply hang two more 1Kv diodes off of the Biass taps on the LV xfmr and 100 or so 200 volt cap will provide 130-150 volts or so.

In most locations the AC line voltage is around120--125 volts. This results in higer than 6.3V philamnet voltage. This causes the transformer to run hotter than usual along with reduced life expectancy of the tubes. The LV transformer is wound for 115 VAC. I wired the two unused seriesed 5 volt philament windings in series with the primary of the LV transformer. This dropped the philament voltage to 6.1 or so. In the end ,you will have a reliable transmitter that will have respectable audio with full modulation and minimal distortion. The Viking II will give many hours of pleasurable operation in it's modified state.

I realize I made a real old buzzard transmission here but when I see someone in a sand trap I thought I would help out. To others who have a Viking I or II who are contemplating modification please consider what is said here. Hmmmmm I smell another Henrynellar modification article Unless Steve HX edits it and put it up on AMFONE

-----------------------------------FINI- Tim WA1HnyLR

Next↓Top⇑

VIKING II CDC REAR PANEL CONNECTIONS BY WA2LXB

Viking II rear panel connections (Click to enlarge)

The rear panel connections to the Viking II are not shown anywhere in the Johnson Viking II manuals or documentation on line. Hugh, WA2LXB has provided photos of the CDC model we restored. The labels and photo are his work, shared here for Viking II aficionados. The Johnson TVI filter is the exact unit specified in the CDC documentation, with mounting straps fabricated using actual factory made straps as a model. You always wondered what the four holes in the back of the cabinet were for. Now you know!

All of the information here is equally applicable to the standard Johnson Viking II.

Next↓Top⇑

VIKING II CDC TUNEUP PROCEDURE BY WA2LXB

The Viking II has a LOT of knobs! There are multiple band switches, one on the VFO, one for the exciter, one to select 160/normal 3 - 30 MHz (more for the CDC model), and the big crank knob for the final plate tuning (continuous 3 - 30 MHz). To a newcomer to AM and vintage transmitters, this can be daunting. WA2LXB wrote this detailed step by step tuning procedure he uses. This is provided for those who are just getting acquainted with the Viking II. It provides valuable supplemental information to what is in the EFJ manual.

WARNING: NEVER CHANGE THE PHONE/CW SWITCH OR THE COARSE LOADING SWITCH OR BAND SWITCH WITH THE PLATE HIGH VOLTAGE SWITCH ON! It can damage the switches and other components.

CW OPERATION

1. Plug in a crystal in position 1 of the crystal selector socket
2. Set CRYSTAL dial on "0"
3. Set Frequency to correct range
4. Set DRIVE on "0"
5. FINAL (large spinner knob) on "0" (full tank in)
6. CW-PHONE on CW
7. PLATE (high voltage) toggle switch to off (down)
8. COUPLING on "1"
9. FINE COUPLING on "0"
10. AUXILLARY switch on "3.5 – 30"
11. Approximate settings of the oscillator and buffer may be obtained from tuning curves on sheet following page 10
12. Throw FIL toggle switch to SW-1 (up). This applies all voltages EXCEPT plate and screen voltages for the 6146 final and 807 modulators
13. Turn CRYSTAL switch to position 1. The "osc" current should DROP perceptibly.
14. Turn METER to "bfr" and advance DRIVE to about position 3 in the CLOCKWISE direction
15. Tune OSCILLATOR SLOWLY for a rise in buffer current. Adjust tuning MAXIMUM buffer current. In SOME cases a SLIGHT dip of buffer current may occur between two maximum current points. This is true with very active crystals or a strong 6AU6 oscillator tube. Correct tuning, when this occurs, is usually ON THE DIP between maximum points.
16. Turn METER to "grid" and note 6146 grid CURRENT. If it should be OVER 8MA, REDUCE IT AT ONCE by turning DRIVE COUNTER-CLOCKWISE. Usually there would be no grid current at this point but the buffer tuning could happen to be nearly correct.
17. Tune BUFFER for MAXIMUM "grid" current keeping the grid reading BELOW 8MA by means of the DRIVE control. After BUFFER has been properly adjusted, turn DRIVE to its "0" position.
18. Turn METER to "plt" position and PLATE ON. Turn DRIVE CLOCKWISE until PLATE current rises to 100MA.
19. Tune FINAL toward 100 until plate current DIPS SHARPLY. Be CERTAIN that the FINAL is tuned to the FIRST dip in plate current thus assuring that the amplifier is tuned to the fundamental and not to a harmonic.
20. Turn METER to "grid" and add touch-up BUFFER tuning for MAXIMUM grid current. Advance DRIVE until 6MA of grid current is obtained. Turn METER to "plt"
21. Proceed to load the antenna or dummy load by advancing the COUPLING CONTROL toward higher numbers, returning to the FINAL for a DIP after each adjustment of the COUPLING CONTROL. Continue this procedure, adjusting COUPLING CONTROL and re-tuning FINAL, until a METER reading of approximately 200MA is obtained under resonant conditions (FINAL dipped).
22. A minimum coupling capacity equivalent to that of position "6" of COUPLING should be maintained in any case on the 5.8-8-4 MC range. The antenna is usually of the range of the Viking II CDC if less than a total of 150 mmfd of output capacity is required on this range. If COUPLING is turned to "7", FINE COUPLING should be advanced no further than 60 in the following steps.
23. Turn FINE COUPLING toward higher numbers (no higher than 60) until the METER "plt" reading rises to about 250MA. DIP the plate current with the FINAL adjustment again and follow this procedure until a desired plate current between 250MA and 300MA has been reached. If the desired coupling cannot be obtained with the COUPLING setting of step 22, turn the FINE COUPLING to "0" and advance COUPLING another step and re-adjust FINAL and FINE COUPLING. PLATE current on CW should not exceed 300MA.
24. The preceding steps complete the normal tuning procedure. However, when initially setting up the equipment the 6AQ5 clamper tube must be adjusted and once adjusted will remain stable and need not be changed over long periods of time.
25. 3.3 to 4.5 MC Tuning: a) AUXILLARY switch on "3.5-30MC" PLATE OFF while changing b) Use 1.65 to 2.25MC or 3.3 to 4.5 MC crystals c) DO NOT attempt to use COUPLING beyond step 6 or FINE COUPLING beyond 50 in cases of high antenna impedance as in the case of 2.3-3.3MC tuning.

PHONE OPERATION

1. Tune up Viking II CDC in CW position
2. Turn the PLATE switch off
3. Turn the PHONE-CW switch to PHONE and turn the PLATE toggle switch ON and adjust for 6MA "grid" current and a "plt" current of 230MA. DO NOT EXCEED 230MA plate current for PHONE operation. Turn PLATE switch OFF.
4. Turn AUDIO to "0" and connect a high impedance crystal or high output dynamic microphone to the connector marked "mic". 5. Turn METER switch to "mod". Turn PLATE toggle switch ON. The modulator no-signal current should be between 60MA and 80MA.
6. Place a DC volt meter between jack J70 and ground to read positive voltage and adjust R76 for a plus 4 volt reading. This completes the modulation limiter adjustment.
7. Talking into the microphone, advance AUDIO until the METER "mod" current rises to 120MA to 130MA on audio peaks. 120MA to 130MA of modulator current corresponds to 100% modulation. A small downward reading on the "plt" METER setting is normal during modulation.
8. Check 6AQ5 clamper tube adjustment by switching crystal out of the oscillator circuit momentarily. "Plt" current should drop to less than 50MA.
9. If the "Plt" current is above 50MA, refer to clamper tube adjustment section.

Top⇑

VIKING II A BRIEF DESIGN HISTORY

This article was intended for a presentation at the Antique Wireless Association yearly conference. We were unable to arrange a suitable schedule time for it. You should consider joining the AWA.
https://antiquewireless.org/homepage/

Why an engineer made a product the way he did is often the key to get the best performance out of it. Sometimes a unique product has reason for its distinct design. The reason for this brief archaeology dig is curiosity about the reasoning in the evolution of the Johnson Viking series and its contemporaries. Its a distinctive and handsome rig with an interesting past.

The Johnson Viking I, Viking II, and Viking II CDC were some of the earliest 100 watt class transmitters available beginning in 1949. It was Johnson's first amateur transmitter. This explains why they had some design quirks which have caused problems. Johnson actually issued later factory modification instructions which did a pretty good job of correcting these issues. But hams have a way of introducing their own fixes; some of those were effective, and some made problems worse. Given that many Vikings were kits (before Heathkit set the Gold standard for clear assembly manuals), some problems arose from that as well. Taken in context of the time, it is understandable how those shortcomings could arise. When the Viking I came out, the Collins 32V1 was already using the 4D32, a sturdy WW2 surplus part. It was logical to use that tube or its cousin the 829, which was an acceptable substitute. The 32V1 also used 807s (another popular WW2 surplus item) for the modulator. Specified 32V audio distortion is less than 8% at 90% modulation with a 1000 cps input frequency, with frequency Response of 2 db from 200-3000 cps. You thought it was better than that, didn't you? The legendary Collins 32V was truly the Harley Davidson of transmitters. The stability of the internal VFO was way ahead of its time.
Collins32V.html

The 32V1 exciter section was a very complex system, with a sophisticated PTO oscillator and tracking tuning for the exciter. E. F. Johnson made simplifications that made a 100 watt class Viking I transmitter more affordable in kit form, but had lots of exciter knobs to tune. WW2 surplus transmitters often were general coverage over a frequency range of 1.7 to 20 MHz, so it was logical to build the Viking I that way. The special general coverage Viking II CDC covers 1.7 – 30 MHz with no gaps. There is a rare Canadian version the C-C or Continuous Coverage. The stock Viking II will cover CW for 30 meters as well as CW/AM for 17 and 12 meters. This is one reason why it is still my favorite. One of the better early linear amplifiers was the Gonset GSB-101 and 201, which featured three band switch positions for 80 meters, to keep circuit Q of the plate tank optimum. The Viking series and the Hallicrafters HT-20 are the ONLY transmitters with continuous HF coverage I know of which did the same, with gear driven plate tuning capacitor and roller inductor.

Popular WW2 surplus rigs with carbon microphones also had a unique sound on phone and generally covered a lot of HF spectrum outside the amateur bands. The reason for their popularity, besides price, was that the actual frequencies for amateur use after WW2 were in a state of change. When EFJ designed the Viking I, 15 meters was brand new, and a lot of rigs (like the AT-1 Novice Rig) did not cover it. 15 meters was a very popular band back in the late 50s with its high sunspot count. By using continuous coverage, the rig would not become obsolete with FCC allocation changes.

At the time, phone operation was not the mode of choice. Class B licenses could only operate phone on 160, 11 and 10 meters. Class B was "CW only" in the rest of the HF spectrum. Only class A licenses could operate phone on 75 and 20 meters. That is why Mike, W2OY, sent his classic CQ requesting only class A operators, no lids, no kids. Until 1960, even F3 narrow band FM phone was allowed on 75 meters and other HF bands. Around 1951, many changes to the licensing structure created the Novice class and later 15 meters. The unique crank knob and geared roller inductor for the plate tuning system of the Viking I and its descendants made a lot of sense in that time of chaotic spectrum management. This also solidly positioned it for the cold war environment to make the special edition CDC civil defense model, and a rare Canadian continuous coverage model. The Viking family was conceived as versatile transmitters that would not quickly become obsolete. Before you do any modifications or repairs, the Johnson factory updates are required reading to get the best results.

Most operation back then was CW, and a lot of it was still crystal controlled, explaining the 10 position crystal switch. For civil defense operation outside the ham bands, just plug in a crystal and tune it up. The stable Viking 122 VFO was also available for ham band operation in 1949 as part of the Johnson Viking system. Elimination of key clicks and chirp were common subjects for articles in QST and CQ magazine. The new Novices who home brewed were learning how to tackle that project. It was also required for a new Novice operator to design a T/R switching system that did not have to toggle multiple controls, to smoothly integrate the components of his station. How many hams these days would need help to get that done?

The switch from the 4D32 to an 829B or newer 6146s arose from a shortage of the 4D32 due to the Korean war. In 1952, RCA introduced the 6146. A kit with a metal plate to mount a pair of 6146s into a Viking I appeared around that time, but the instructions are unavailable. I have an early Viking II for restoration that has 6146s, but no clamp tube from that era. The HV bleeder resistor had two slider taps, one for the 300V 807 screens, and one for about 150 – 200 V for the 6146 screens in CW mode. It was logical for new production of the Viking II to use 6146s. There was a series of important official Johnson mods that addressed TVI (for Viking I) and mods for the Viking II, Mod A, Mod B, and the grid block keying mod. Most every 100 watt class rig that came after that used 6146s as well.

When a Johnny Novice back in the early 1960s graduated to General, they had a plethora of (often poor) choices to get on phone. These included less than effective cathode or screen modulators kludged onto their CW rig, often from kits or home brewed. Many did not own an oscilloscope, something necessary for any AM design work. Consequently, many AM signals of that era were just plain awful. WW2 vets had the training and experience to rework a surplus transmitter for the ham bands. To a high school kid with a limited budget and technical skills, a used Viking II or DX-100 looked like a solution straight from heaven. All of these radios were frog princes with a lot of warts, but we learned to love them over time. In 1960, with savings from a part time job cleaning glassware for the high school chemistry instructor, I walked into an electronics store in Cortland, NY and fell in love with a used Viking II. This article is a reminiscence of that life long love affair.

So when you speak badly of the Viking II, keep in mind the time of its creation and the rapid progress in technology in the Sputnik era. The DX-100 had the benefit of waiting just a few years after the Viking before production, and its trendy but unstable built in VFO was not something you bragged about. If you managed to get within 1 KHz, and stay within 2 KHz of the other station, that was good enough. When the SB-10 SSB adapter came out, Heath had to create the Apache to get a stable enough VFO. The DX-100 key clicks were notorious. The "Scratchy Apache" AM audio was as bad as the Viking. The driver transformers in the Viking, DX-100 and Apache reflected the intent to restrict audio band width. At the time, the ARRL responded to the 75 meter night time bedlam by recommending narrow audio, along with peak clipping to reduce splatter, to fight QRM. There were not enough notches in your QF-1 Q multiplier or Selectoject to eliminate all the heterodynes in 1960. The new Valiant appeared magnificent with its internal VFO and band switching. But the limited Valiant Plate Tuning doesn't quite reach the edges of 80 meters or operate in the WARC bands. The 32V limited matching range needed a tuner for even a standard dipole and the output loading network was fragile. The Viking series was set up to directly match anything: an end fed wire, a coax fed dipole, or my trusty Gotham vertical.

WA1HLR says: "The Viking I and II is the best thought out in the RF department. It is almost general coverage as it is. Yep! 17 and 12 meters on AM. The rest of the transmitter is in great need of fixing. The audio quality is like that of two Dixie cups and a string. There are many modification articles to fix the problems more or less. Problem is they all involve drilling and blasting and replacement of at least the driver transformer."

I plan to delve into the repairs and reliability work first. The Viking 122 VFO holds some secrets, and it is possible to make it work on the WARC bands, 30, 17, and 12 meters. A new digital kit VFO will also work. The popular audio modifications are plentiful and sometimes not so good. I plan to get 100% modulation at full rated AM output and keep both the 7 pin sockets but use different tubes, for less extreme mechanical changes. A new current production good quality driver transformer costing about $16 is now available, which I will compare to the alternatives.

There are cures for the Viking II's warts, if you have the patience. The technical series of articles covering the Viking II in Electric Radio offers solutions to those problems.

POPULAR COMMERCIAL AND KITTED AM/CW TRANSMITTERS OF THE ERA:

• Harvey Wells 1933-1968 TBS-50 (1947)
• WW2 surplus, BC-610 (Hallicrafters HT-4) began 1942, 25,000 produced
• ART-13 (1940)
• Collins 32V1 (1947) $475 See "Zen and the Art of 32V Maintenance" website
• E.F. Johnson Viking 1 (1949-52) 4D32 final 2 807s modulators
• Hallicrafters HT-9 (1940-1948)

• Heathkit AT-1 (1951), VF-1 VFO (1951), DX-20 (1956), DX-35 (1956), DX-40 (1957),
• DX-60 (1961) with model A and B , DX-100 (1956) DX-100B (1959)
• Apache TX-1 (1959) SB-10 SSB adapter
• Hallicrafters HT-20 (1952-1954) (4D32, 807 modulators) HT-40 (1960-1964)
• E.F. Johnson Viking 2 (1952-57), CD variant, Adventurer (1954-63),
• Navigator CW (1957), Challenger (1959-65), Ranger I (1954-61), Ranger 2 (1961-65),
• Valiant 1 (1956-62) and 2 (1962-65), Viking 500 (1956-63),
• Johnson Desk KW (1955-1964), Pacemaker (1956-59) a Phasing SSB/AM/CW rig.
• Barker & Williamson 5100 (1955 – 1961) plus 51SB SSB adapter

• Collins 32V: Internal VFO very stable. 4D32 final still $25. 2x807 modulator good audio, 100% modulation. Weak loading caps in Pi-L network. Modular construction. Exciter very complex and hard to work on. Expensive. NO 160 meters.
• B&W 5100, Internal VFO. 2 6146s, mod by 2 6146s. Good modulation. Modular, could be hard to work on without extender cables. Expensive. NO 160 meters.
• Johnson Valiant: 3x6146 mod by 2x6146. Internal VFO. Speech clipping. 866 Mercury vapor rectifiers, replace with 3B28 or solid state. Arcing problems. Barely makes 100% modulation stock.
• Heathkit Apache TX-1: 2 6146s mod by 2 6CA7s. Claims 150 W input AM, exceeds tube ratings. Arcing under mod sockets, can redesign for 6146 modulators. Internal VFO more stable than DX-100. Later versions neutralized. SSB adapter SB-10. Speech Clipper. "Scratchy Apache" due to poor audio driver transformer, capacitors. NO 160 meters. Can do mod for 160 meters.
• Heathkit DX-100: 2 6146s mod by 2 807s. Internal VFO unstable, drifty, jumps, due to band switch and 500 pF capacitors. Barely stable enough for AM. Modulation fairly good stock form. Key clicks. Grid block keying, factory mod, cures chirp; causes "backwave". One QST keying mod causes low drive and 40 meter and up drift.
• Johnson Viking II: 2 6146s mod by 2 807s. External VFO 122 very stable. Click complaints cured by Factory grid block keying mod, which cured chirp and click. Covers all WARC. Worst stock audio of any of its contemporaries. Problem can be fixed, lots of options.

• Heathkit offered the SB-10 phasing adapter for the Apache, NOT recommended for DX100. I tried an SB-10 on a Viking II, and switching modes was very inconvenient, only worked on 80/40.
• Johnson offered an expensive add on SSB adapter for Valiant. Others could be adapted.
• Johnson offered the Pacemaker (NOTE: spurious signals on 15 and 10) for the Johnson Desk KW, to replace the Ranger exciter and allow SSB operation.
• B & W offered an expensive SSB adapter.
• Central Electronics model 10 and 20 exciters were phasing style, 10 and 20 watts output to drive a larger amplifier.
• Collins was nice but expensive, it made crystal filter based SSB and transceivers the standard.
• Swan made crystal filter SSB transceivers affordable. Due to mixing scheme, drifty.
• Inexpensive filter transceivers from Heathkit included SB-100 (1965) and HW-100 (1969).
• Drake made high performance transceivers in mid price range; it still ranks well on the Sherwood site.
• Later, Ten Tec made high performance solid state CW and SSB available. Still my favorite.
• Finally, Yaesu ended a lot of American amateur manufacturing with the wildly popular FT-101 (1971)

These rigs transformed SSB from a largely home brew crowd hanging out on narrow slots on 3999 KHz or 20 meters to the mainstream pursuit of amateur radio. The end of AM, coupled with incentive licensing, caused many to leave the hobby rather than upgrade and buy all new equipment. Thankfully, we have rediscovered AM, and restored boat anchor classic rigs as well as created all new solid state incarnations. Long live AM!

This article is intended to give background and try to explain some of the quirky design choices Johnson made in creation of the Viking I and II, in the historical context of its contemporaries.

REFERENCE MATERIAL FOR FURTHER READING:

1. AC0OB VIKING II- CDC, PARTS 1 – 3, ELECTRIC RADIO #323, 324, 327 April – August 2016

2. K6AD VIKING II Modifications, PARTS 1 – 3, ELECTRIC RADIO #110, 111, 112 June – August 1998

3. KT2L VIKING II Improvements, ELECTRIC RADIO #9, January 1990

4. https://bama.edebris.com/manuals/johnson/viking1/modification-b
Johnson Modification B, adds 22 ohm resistors to 807 plates, corrects VFO filament choke voltage drop, fixes stability with 56 ohm resistor on 6146 screen, adds 10K 2W resistor to driver coil, adds C56 0.1 uF to clamp tube circuit, adds a PTT circuit. Very important modification sheet to implement. Latest version. Also shows an audio quality K8TV mod with 6AQ5 driver circuit

5. https://bama.edebris.com/download/johnson/viking2cdc/viking2CDC.pdf
Excellent copy of CDC version. NOTE: DIODE RCT70 AT TOP OF T4 MOD TRANSFORMER. REMOVE THIS DIODE AND REPLACE WITH A WIRE. Also note 6J6 compressor circuit is different. Has C56 and R33, 34, and R35 updates. Has a PTT circuit. NOTE: parts numbering is totally different from a standard Viking II.

6. Radiotron Designer's Handbook 4th edition Pages 496-501, Effect of Incomplete by-passing on Gain

7. Electronic and Radio Engineering 4th edition (F. E. Terman) pages 257-275

8. KD0HG ER #11 March 1990 "Audio Response, How Much Do You Really Need?