This page last revised September 9, 1999.
The following is copies (admittedly, edited) of some correspondence which I exchanged with a visitor to my website. I've tried to contact the other person for permission to post this information publicly, but have not had any response, so I'm not including the other person's name (I'll just refer to him as "John Doe"):
The correspondence began when I received the following email. For the following email, the writer was making references to some of the text on my R392 web page as it existed at that time:
email from John Doe to Mike Dinolfo dated April 5 1999:
>Replacing audio stages with solid state line-level driver: .....
I've already done this -- I'll dig up my notes and send them to you. No soldering required. All mods replace existing tubes.
>Remove V601, V602, and V603; replace the diode sections with solid state diodes (perhaps 1N914 units?).....
I've done this mod too..... After you replace the AM detector and AGC Detector you can add a mod to increase the BFO drive level. I'll dig this mod up also. I may have to open up my '392 since it has been 15+ years since I installed these mods.
The result is a great reduction in 28v supply requirements.
Both the first and second mixer tubes can be replaced with 2n3819 or 2n3823 jfet. Both should be available at Radio Shack. FET should "plug" right in -- no additional mods or parts required.
The 26A7 can be replaced with a pair of darlington NPN power transistors. I used bargain basement to-202 types of unknown origin. Bias resistors were added.
The 6aj5 audio stages can be replaced with 2n2222 or 2n3904 NPN types with a couple of resistors added to set correct bias.
The detector diodes should be replaced with germanium types -- 1n34, etc.
The result of of all these mods was much less heat and current with only slight increase in audio distortion at low freq -- probably due to output transformer or phase invertor imbalance. If the r392 is working correctly with the tubes there will be no loss in performance after switching to solid state.
If you are interested I'll scan my notes and EMail them along.
end of April 5 1999 email from John Doe to Mike Dinolfo
email from John Doe to Mike Dinolfo dated April 6 1999; a lot of additional information is provided. Again, references which John makes are to some of the text of my R392 web page:
(1) FET Mixers - The 26c6 tubes are poor at best. I couldn't find any difference replacing the tubes with 2n3819 fets. In fact, the 26c6 mixer tubes look exactly like the 2n3819 when modeled at 28v. I soldered pins from a cannibalized db25 male to the fet leads and just plugged er in. I have a 2n3819 at v204 and a 26c6 at v203. Must have run out of 2n3819!
(2) 26A7 - I used a pair of Ge D43D Darlington to-220 power transistors. 130 ohm bypassed with .1 uf in the emitter leads. Bias resistor between collector and base of 2 meg ohms. To reduce chance of oscillation, the bias resistor has 220 pf across it. I built the thing on a piece of proto board and glued it to an old octal tube base.
(3) More Audio - I built an external 10w HiFi audio amp to really liven things up. I chose not to modify the radio to increase the audio power.
(4) V607 6aj5 - 2n3904 with 1 meg bias resistor from base to collector.
(5) V606 6aj5 - 2n3904 with 2 meg bias resistor + 1.5k unbypassed emitter resistor.
(6) V602 and V603 12au7 - Germanium diodes. I used what I had in my junk box, 1n101. Silicon diodes may work better for the AGC rectifier I just didn't try them. The detector coils will have to be repeaked with the new detector diodes.
(7) V601 12au7 - I have a germanium diode on pins 1 and 3. Nothing for the other half of the 12au7.
(8) V605 6aj5 - empty - I have tried a 2n3819 fet here. Works OK but really not satisfied with squelch action.
(9) BFO output mod - 220 pf from pin 5 of v604 to ground. Value was determined by trial and error!
RF and IF stage mods -
It's hard to replace the RF and IF tubes without messing things up. The solid state replacements have too much gain and cause oscillations. After some experimentation, I gave up! The RF and IF stages are what "makes" the R390/392 receivers. I couldn't justify further work.
Ramblings.........
It's real easy to get too much gain with solid state replacements for the 28 v tubes. My brother and I built a test jig so we could evaluate the 28 v tubes as fets on the curve tracer. That's how we came up with the 2n3819 replacements. Matching the gain and agc action of the 26a6 used in the IF and RF stages is nearly impossible. It's like a juggling act!
I once ran the b+ on 3 Kodak Ultralife 9v batteries.
Tube equivalents -
26a6 <-> 6ba6
26d6 <-> 6be6
26c6 <-> 6BF6/6BU6
The 28 v tubes are characterized at 28 v b+ and 250 v b+. We tried a 6ba6 at 28 v b+ and it behaved just like a 26a6 did at 28 v. Surprise!It should be possible to replace the 26a6 and 26d6 with their 6v or 12v equivalents and run the filaments on 6 or 12 v. I entertained ideas of running the entire radio on 12 v by replacing the tubes and building a 12v to 24v converter for the b+.
Somewhere, many years ago, I received a case of 26d6 tubes. The person I got them from did not know what they were used in. Neither did I at the time!
This stuff was done around 1980. I have not done any further work on the r392 since then. A few years ago I retired the r392 -- I got a TS450 with a general coverage rx.
My r392 is a Stromberg-Carleson made unit. The tube pullers and and all of the special tools were still inside.
(portion of text deleted)
My feelings are that the mods should be easily reversible and should be nearly as good as the original receiver. I understand there are other r392 out there that have been modified to the extreme and are not reversible.
>From a purely engineering perspective, it should be possible to replace every tube in the r392 by simply plugging in transistors or fets. The tubes operated at 28 v have no advantage that I can see over solid state parts. Properly done there would be no difference between the original and modified receiver.
In many ways, my r392 sounded better than the r390a in my shack. I'm sure the 10w external amp helped some!
end of April 6 1999 email from John Doe to Mike Dinolfo