In the 1940s, Frank Jones was a prolific writer on anything radio related and published many books and designs.
One was the Supergainer, and later, the Ultragainer
Years ago I had a crack at making one, and was suprised at its lively performance, especially considering its low parts count. Using modern high slope TV valves in my case made for a very compact and nice performing radio.
The original did have some quirks, I think caused by the choice of components back in the 40s. The Ultragainer used not only the regenerative IF, but also a regenerative mixer stage, I would say a nightmare to tune and to remain in tune.
The regenerative IF was also dodgy by design, the feedback obtained by "lifting" the cathode from RF ground and relying on Anode to Grid, and Grid to Cathode spurious valve capacitances to make an artificial tap on the IF coil.
This is VERY poor design practice, usually capacitances are added to swamp these random and variable unknowns. I do wonder how many of the original designs that were built actually worked correctly in the day. Just changing the valve, or using a valve from a different manufacturer would likely have failed the project.
A quick video on my version supergainer in action, and compared to the JRC, same aerial..... the difference is..... um.... quite suprising I thought.
Using an ECH84 gated synch separator valve for the HF Oscillator and mixer, and the ECF80 , the pentode section being the regenerated IF/detector, and the triode section being the audio output stage.
It works very well on SW AM, but leaves much to be desired on CW/SSB.
After some thought and experimenting, the answer seems to be CIO injection via the screen grid of the detector.
The screen grid is used as the regeneration control, the DC voltage here is varied from around 20 to 30 volts to increase/decrease the pentode gain. By injecting an external oscillator signal at the 1700kc IF frequency,and superimposing it on the DC voltage, the detector doubles as a regenerative product detector.
By doing this, the detector can remain in its most sensitive spot, giving maximum selectivity, not being required to go into oscillation.
Also, the detector locks on to the stronger BFO signal, and if this is crystal controlled, will keep the receiver exactly on frequency.
Why this is so is a little unclear, there must be some slight interaction between the detector and first oscillator, even with the BFO offset frequency required for SSB reception.
Although I hope to use the audio output triode to double as a crystal controlled BFO to save adding another valve, I dont fancy my chances of this working correctly, being in the same envelope as the high gain IF. Changing from a ECF80, or 6BL8 to a ECF82/6U8 may help, I think that one has an internal shield?
Anyways, thats my next move.
'regenerative product detector' ... um ... I had the same idea a while ago, but I aimed to achieve the same by using a 2nd ECH84, and in the following way: 1) signal from 1st mixer, fed to 1st control grid, 2) regeneration control by setting the DC voltage on the screen, and 3) injecting the 'CIO' (i.e. the 'gating signal' from the Crystal Osc.) on the 2nd control grid.
I suspect the 'poor performance' as SSB detector in your experimental setup being cause by instability of 'CIO' induced by current variations on the screen grid, that's working as the plate of the regenerative detector. The easy 'locking' you mention is perhaps in support of my explanation.
The use of a 5 grid converter tube (ECH84 can be seen as such, imho), with separate grids for separate inputs, should keep the interactions against the CIO a bit lower.
Currently I have a few of such ideas I would like to experiment with, but unfortunately I'm occupied into a big campaign of components recovery from broken radios and pieces of old electronic stuff. Once I finish ... I'ill start joyfully using all the saved goodies ....
Very cool. The performance is crazy good. Nice writing too !