Any chance of some further general discussion on this point? I’m really struggling to understand the schematic as I have no electronics training. What I’m trying to understand is whether C15 is in anyway tied to the output transformer type or just sets a safe operating parameter for the EF86. I also realise the amp has a negative feedback loop, but I don’t really understand how these work (my guess would it would be R12 & C7, but I don’t really know what it does!). What exactly is likely to be oscillating, the EF86 or the whole circuit?
Another thing that puzzles me is why there is so much commonality between so many values around the small signal valves in both the TL12 Plus and the Stereo 20? I don’t get this as there is a substantial gain difference between an ECC81 and ECC83, and an EF86 isn’t even a triode!
I don’t need to know this stuff, I’m just curious. All I actually need to know what to buy and I’ll then solder it in and see if it works! I can see some 220pF silver mica caps at HiFiCollective (where I’d have to go to get a matching Takman 4.7k resistor). I could buy some other mica values too if there is any doubt as to the right answer here (I’ll get some 300pF for the S20). Given HFC’s postage prices I’d far prefer to bung an ‘every eventuality’ order in than have to make several.
As mentioned before C15 and R23 across the 82K load resistor for the EF86 stage roll off the gain at high frequencies by shunting the load. This reduces the tendency to oscillate at HF caused by phase shifts in the rest of the circuit, particularly the output transformer. Obviously if overdone the response at audio frequencies will be affected but if too little the amplifier will become unstable. This is the reason for the square wave tests so popular in the old HiFi news reviews, the squareness of the output and the presence or not of ringing on the transition gives a good indication of stability.
The " goodness" of the output transformer is the major variable and is one of reasons that the various correction networks differ for different build versions of the amplifier. Major advances were made in magnetic materials during the war which is why Williamson, Leak and Walker amongst others were able to produce such improved amplifiers shortly after the war.
C15 / R23 are not the only components involved, R12 / C7 and R12 / C1 , the main feedback components are also involved, as well as setting the overall gain so there are many variables and I freely admit that the deeper maths is beyond me, especially as my valve theory was 53 years ago and I haven`t worked on valve kit for forty years or so.
There is more to the gain of a stage than the gm of the valve, the current the valve is run at and the consequent component values are also very important and affect the final result.
There is a popular belief that you can reduce the sensitivity of an amplifier such as the the Leak by replacing the high gain ECC83 with a lower gain ECC81 or even lower gain ECC82 - this is not so as the global feedback will take care of this at the expense of frequency response and distortion, even if the valves are operating somewhere reasonable on their operating curve.
I`ve dredged my memory for all the above, if any of it is wrong hopefully smarter people will say so.