This legendary rig is probably one of the most well known and best loved transmitters of the 50s and 60s. It affordably got a lot of new hams on the air. The number of modifications and articles out there is a testament to the longevity of this product. Some DX100s passed through my hands. I got a lot of enjoyment out of getting them running again, and upgrading the CW keying and audio.

I will not reinvent the wheel here. I draw on others proven work when doing my own restorations. I do have a few additional recommendations you may find useful. I have since sold the rig and will have to fall back on the work of others for the pictures. Here are some of the best I have seen on the web.

The official Heathkit modifications, including the loading capacitors and grid block keying are essential reading. The loading capacitor mod replaced the coarse and fine loading (small red) knob with a single three gang loading condenser. The grid block keying mod typically installed a push button on the right hand side of the front panel and made some internal changes.

I DO NOT RECOMMEND THE SB10 SSB MODIFICATION. As noted later in my article, the badly needed neutralization will not work with the SB10 mod. The DX100 VFO is barely adequate in stability for AM and CW, not good enough for SSB operation. If you have to use a SB10 sideband adapter, get an Apache and leave the DX100 alone! The drive on 15 and 10 meters will never be right with the SB10 modification to a DX100 either due to cable capacitance. Switching modes from AM/CW to SSB involves cable swapping and is mondo inconvenient. The Apache has a knob to switch modes.

The following resource has the essential Heathkit factory modifications and also includes some valuable trouble shooting tips: http://k9sth.com/uploads/Heathkit_DX-100_mods-1.pdf

I found that the original Heathkit block grid keying mod produces backwave. What is backwave? Modern hams new to the hobby may have not heard the term. Basically, it means that during CW key up conditions, some small percentage of signal is still produced. It is such a low level it is unlikely that you will detect it on your scope or power meter. If you have a S9+ signal, the receiving station may see an S4 signal that he reports to you as an oddity. It is no cause for concern. It is not a spurious signal that will produce a signal out of band. It is only on your frequency. It happens because to reduce chirp, the Heathkit modifications run the VFO constantly when the PLATE switch is ON, even in CW. The driver stage is fully cut off, but the stray capacitance couples enough energy to the final RF amplifier and antenna that signal leakage occurs. The factory grid block keying modifications can improve stability on CW and PHONE in addition to removing completely any key clicks on CW, the primary reason to perform the mods. The Heathkit keying modifications provide a better ground for the cathode of the VFO tube, helping even on phone. Switch contacts on the HV toggle switch and jack and key resistance can give a warbly drift of the VFO. This mod partially corrects this problem and is badly needed. See comments at the end of this article for my fix to the rest of the VFO frequency jump problem.

There is another grid block keying mod in a QST around 1956. I did this to one of my DX100s. It uses a complicated time sequence keying system similar to the Johnson Valiant. It is a well written how to article, complete with good pictorials. It would be good for full break in CW operation, but it is not at all necessary for PHONE. It is more complex than needed to get adequate CW performance. But it also contributes to MORE DRIFT on bands using the 40 meter VFO output. This is probably due to increasing the value of the grid leak resistors in the VFO. This makes the dependence on 6AU6 tube selection more critical. I do not recommend any modification that increases the value of these resistors. Worse yet, it results in LESS DRIVE on 10 meters. I SPECIFICALLY DO NOT RECOMMEND THIS MODIFICATION. That is why no reference is included in this article on the DX100. April 1959 QST had a keying mod, which I cannot recommend because I did not personally try it. WB2EJG claims he had good results with the 1959 mod, so give it a try.

I worked on a DX100 with the following mod, and it worked well. It brought the DX100 up to Johnson Valiant power levels. The modulator worked hard, but came close enough to getting the job done after some improvements. With some work, you could get more peak watts out of a pair of 1625s or 807s. But 80 to 90 watts should modulate 160 watts OK, and the three 6146s would do 180 watts (three times 60 watts apiece) fully loaded in ICAS specs. Back it off a bit and get 100% modulation if you have to. One concern is the increased filament current for the extra 6146. Another concern is the extra drain on the HV power supply. I do not recommend the modification for you to do at this point. But if you encounter a DX100 which is done up this way, there is no great cause for concern. It remains a mildly popular commercially available kit mod which used three 6146 finals: http://www.ami-west.com/DX100.W6OM.html

No discussion of power output is complete without taking a look at the ICAS ratings of the 6146. The power input in class C is rated at 60 - 67 watts per tube or 120 - 134 watts for a pair. Some manufacturers rate their standard 6146s up to 67 watts per tube. With 3 you get egg roll. No, you get 180 to 202 watts input on AM or 90 X 3 = 210 watts on CW. Curiously, the AM tune up specs in the stock DX100 manual call for 240 mA X 750 volts = 180 watts FOR ONLY TWO TUBES! The three 6146 mA modification simply adds the necessary plate dissipation to make the DX100 make its claimed input work within the ICAS specifications. What if you simply plugged in 6146Bs for the RF final and operated them within their ICAS ratings? The ratings for a 6146B are ICAS plate modulated AM are 84 watts per tube or 168 for a pair. This is with the same filament current drain. So there is an advantage of using the pair of 6146Bs that is nearly as effective as the three standard 6146 modification! Since the power supply has demonstrated its ability to withstand the extra power from operating the rig outside ICAS ratings with two 6146s or within ICAS ratings for three 6146s, the choice of two 6146Bs seems to work. The referenced three 6146 modification sheet does not give instructions for tuning up at all. It only explains how to install the modification. So I do not know if they intended to further hot rod the DX100 past ICAS ratings or use the stock DX100 tune up instructions.

There is one cautionary tale in the use of 6146Bs: http://k9sth.com/uploads/The_6146_Family_of_Tubes_1.pdf

I do take some exception to his comments. The VHF and UHF operation of the tube realistically can make demands above HF operation, in particular, interelectrode capacitance and lead inductance for various elements, in particular the cathode and screen leads. Also, the Collins HF SSB rigs used inverse feedback around the RF final and driver tube. Off resonance operation during tuning can cause oscillation of a system like that, which is stable at resonance. With a HF rig like the DX100, Apache, Johnson Valiant, or Viking 2, which can be neutralized, and has no inverse feedback complications, I have had no problems. But your mileage may vary. Another consideration is that modern line voltages often produce closer to 800 V or even more on the HV. Reduce the plate current accordingly to get the proper RF plate input power. Especially if you raise the HV even more by solid stating the HV power supply. Its basic math. Desired plate current = desired power input watts divided by plate voltage. Allow 15 to 20 mA for screen current, since the meter reads cathode current, not plate current.

Can you do the three 6146s modification, put in 6146Bs, and run them at 242 watts? Well, will the 807s modulate that RF power and the HV power supply take it? Will the secondary of the modulation transformer do the extra current? I ain't gonna go there, you are on your own. Keep in mind that you have to make a 6 dB increase in order to get an S unit on the other end. All of these modifications are subject to that limitation versus the life of the equipment being shortened from abuse. I am all for improving the functionality and quality of sound of a rig, but do not favor blowing it up from pushing the limits.

In the horsepower wars, the DX100s claimed 100 – 125 watts OUTPUT on phone and 120 – 140 watts OUTPUT on CW. This is only possible if operated above ICAS ratings at 180 watts INPUT. That would be an implied efficiency of 69.5%. The same output tubes in a Viking 2 were claimed at 100 watts output phone and 130 watts output CW on 10 meters, with comparable results on lower bands. The manual tells you to run it at 230 mA at 620 volts or 142 watts input less 12.5 watts for the screens for an RF INPUT power of 130 watts, within ICAS specs. That is 76% efficiency. Both efficiency figures are realistic for class C operation. The DX100 claims up to 125 watts phone output; the Viking 2 claims only 100 watts phone output. Maybe you would buy the one with the higher claimed power, especially if it is about $100 cheaper. But if you read tube specs and can do math, maybe not.

I said I might do some math, and let you decide which is the truth.

In any event, no matter what you run or how you tune it, I strongly recommend a cooling fan and never setting anything on top of the DX100 or any other AM rig that would impede air flow.

Adding a frequency counter. I did something similar, and it was a big advantage, given the tendency of the DX100 to wander off frequency. http://home.comcast.net/~n8itf/dx100mod.htm

As with all transmitters from this time, the 4 watt drive pot is subject to failure. These are getting hard to find, and there is a better way to address the problem. Here is the drive pot modification: http://www.amwindow.org/tech/htm/drivepot.htm

A well done photo essay showing comprehensive restoration and modifications: http://www.k3msb.com/dx100b/dx100b.html

The finals work better on 15 and 10 meters when they are neutralized. I highly recommend this modification. Note that it will NOT WORK WITH A SB-10 modified DX100. Here are the parts I found important:

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.

Please see: http://www.amwindow.org/tech/htm/dx100.htm

I closely paralleled WA1HLR's work. It is difficult to work on the modulator low level stages, but you can do it with patience. I added low voltage PTT (push to talk) powered by the filament supply to a solid state relay. It drove a 3 pole 115 VAC relay which controlled the VFO keying for AM, the 807 screens (which now run off the LV supply instead of the HV bleeder midpoint tap), and the LV to the audio driver stages, as well as the receiver muting. The PTT circuit actually drove the control side of a solid state relay which sent 115 VAC to the HV transformer primary, the Dow Key T/R antenna relay, and the 3 pole relay coil mentioned above. MOVs protected the solid state relay from the 3 pole relay and HV transformer surges when going back to receive. This mod also allowed replacement of the 115 VAC bulb to indicate transmit mode in front of the VFO with a jumbo LED in series with the control side of the solid state relay.

The tune up instructions in the Heathkit manual do not observe the proper ICAS ratings for the RF final tubes. A 6146 is rated for 62 watts input each for a total of 120 watts input for a pair. Do not run them at 180 watts input (the CW ratings) in PHONE, or they will be damaged. Measure your DC plate voltage with a calibrated meter (not the front panel DX100 meter). Allow for about 15 to 20 mA screen current. The DX100 displays CATHODE current, which is the sum of the control grid current (negligible), the screen grid current (15 to 20 mA for two tubes) and the plate current. The increase in strain by operating them out of spec will not provide a noticeable S meter increase at the other end, but it will measurably reduce the life of your final tubes. It possibly does look good in an advertisement. The honest specs published for the Viking 2 appeal to me more. But maybe the sales department would not agree. (Click on thumbnail below to enlarge.)

BOTTOM LINE: This is a good legacy AM and CW rig. While it has some quirks, as do all the equipment from that era, it is an affordable way to experience AM or nostalgia operation. If you can obtain one reasonably, buy it.

I have tried to gather together in one place the best of the best of the modifications. I have tried the mods I have recommended here. I added some of my own ideas. You may have some of your own favorites. But if you do the best ones, you will have a solid performer for a fair price.

Do something about the 5R4 sockets. Add insulation by increasing the size of the space between the sockets and the chassis. Use some hi voltage insulating paint too. When solid stating this power supply, be sure to protect from arcing with HV values that are even higher than the stock unit.

Add a separate fuse for the LV transformer primary. If the electrolytics fail, the fuses in the stock cord are too big to protect it from damage.

The 1625s can be replaced with 807s by rewiring the filaments in series rather than parallel. This ensures that if one tube fails, they both go off and no modulation occurs, alerting the operator of the failure. The 807 is more available than the 1625 these days. However, you will have to replace the tube sockets from 6 pin to 5 pin. Other than the filament ratings, all specs are identical.

The internal VFO has some known issues, which are widely reported. The long shaft of the VFO switch inside the VFO enclosure has some play in it, which can expand with heat from the HV ON indicator and the 1625 modulators nearby. Even the ceramic switch wafers can shift with heat. This results in a sudden jump in frequency. The 500 pF mica capacitors contribute to this as well. The 6AU6 can add its own drift and jump. You can do something about this, but it is a huge job. It involves replacement of the entire VFO switch with reed relays glued to the case, driving it from a new switch mounted directly to the DX100 chassis. Also, the mica capacitors can be replaced by polystyrene high Q capacitors of the same value. This eliminates most of the source of the problems. The coil and tuning capacitor are surprisingly good. You can solid state the VFO, leaving the 6AU6 can remain as a buffer stage to get the voltage you need. All that remains is the thermal drift from the heat during warm up. You can add a power resistor which stays on even though the rig is off, or better yet, a temperature controlled heater in the VFO case. With the difficulty involved in working on the internal VFO, I doubt anyone is interested in all this effort. I do plan to write up a TURBO VF-1 article later which gives the details of this work, but doing it to an internal VFO like the DX100 seems unlikely. You might consider a modern synthesized VFO with the original VFO left in place for appearance. The crystal VFO switch would serve to switch sources from internal to outboard synthesized VFO. Or you could defeat the internal VFO entirely, and use it as a buffer amplifier for the synthesized VFO.

This is one of the reasons I prefer the Viking 2 above the DX100. The VFO is external, eliminating most of the drift problems, and much easier to work on. The other reason is nearly continuous coverage of all frequencies in the 1.7 to 30 MC range, allowing WARC band operation with the right VFO frequency.

I have a separate article on the Apache, and some of the comments here apply to that rig as well. If you need the SB10, get an Apache. Stock, it does not have 160 meters. But that CAN be fixed! However, I hate the 6CA7 modulators in the Apache, as well as the modulator tube sockets arcing.

Here's a source for NEW replacement front panels: http://swap.qth.com/search-results.php?keywords=dx-100&fieldtosearch=TitleOrDesc.

Also, here's a source for synthesized VFO: http://www.pongrance.com/.

• Remove the original VFO box and gut it, except for the 6AU6 socket and output coils.
• Design a new switch to attach to the linkage for bandswitching.
• Use the 6AU6 as a buffer amplifier.
• Install the display in the original VFO window.
• Put the shaft encoder in the original VFO frequency control hole.

I have not done this, but the complete fix for under $100 is a clear choice over trying to get the original parts to be truly stable at 40 meters and up.

Lastly, I do want to reiterate: DO NOT DO the SB10 modification, or the 1956 QST time sequence keying modification. Maybe the April 59 QST keying mod will work; I have not tried it.

FOR BRAND NEW SILKSCREENED FRONT PANELS FOR THE DX100 FOR A COMPLETE RESTORATION - from Electric Radio magazine. (http://www.ermag.com/) (You really need a subscription to this great journal if you are a builder).

For DX-100 new panels contact: W3FNZ@verizon.net