PS
Just like the Jpe Tritschler 75uS roll-off followed by the 3180/318uS shunt feedback design , the JL Hood derived design where the shunt feedback 75uS roll-off follows the 3180/318uS feedback stage, it's difficult to get the response curve right because of the interaction between the two stages. With the full passive design it was relatively easy to get below 0.2dB response curve , while the other two designs were strugling to get the response curves right. In the fully active design , I wanted to lower the feedback resistances involved , but the interaction with the second stage made it too much of a hassle if I went below 49k9 gain setting series resistor. Split filtering like these two designs are harder than when you have one feedback or passive network with fully known interactions and readily available calculators on the net, but I think if you put voltage follower stages in between it should be easier for the calculations and conformance thereby.
Just like the Jpe Tritschler 75uS roll-off followed by the 3180/318uS shunt feedback design , the JL Hood derived design where the shunt feedback 75uS roll-off follows the 3180/318uS feedback stage, it's difficult to get the response curve right because of the interaction between the two stages. With the full passive design it was relatively easy to get below 0.2dB response curve , while the other two designs were strugling to get the response curves right. In the fully active design , I wanted to lower the feedback resistances involved , but the interaction with the second stage made it too much of a hassle if I went below 49k9 gain setting series resistor. Split filtering like these two designs are harder than when you have one feedback or passive network with fully known interactions and readily available calculators on the net, but I think if you put voltage follower stages in between it should be easier for the calculations and conformance thereby.
PPS , I remember , the first Joe Tritschler design I made , I had to use a 2n1 C as the 75uS roll-off cap with the 36k series resistor , but the shunt feedback network consisted of the 49k9k series to the grid input , with 300k in series with 1n1 coming from the output to the grid of the second stage. That gave the closest conformance to the riaa curve, except for the bass lift in the 50 to 20 and lower Hz region. I added, I think it was a 5M6 R parallel to the 1n1 C to remedy the too high bass lift and shifted the coupling cap to before the 49k9 gain setting resistor. To prevent interaction from the grid leak resistor , recommendation is to have the series resistor be 50 times less than the grid leak resistor, being that 1M , the series resistor should be 20k or less, which I think is the combined series resistors of the first stage and Zout and the second gain setting resistor to the grid of the second triode, but that resulted in too much interference between stages , besides loading the output of the first triode too much probably. Increasing the grid leak resistor to 10M results probably in grid-leak-biasing of the second triode, which is unpredictable.
Grid leak resistor for ecc83 is 2M max if I remember correctly, but for the 6J11P is 500k I think,at least for the e280f it is, which means series resistor should be no more than 10k.
Grid leak resistor for ecc83 is 2M max if I remember correctly, but for the 6J11P is 500k I think,at least for the e280f it is, which means series resistor should be no more than 10k.
Last edited:
You've given a ton of detail above. Thanks. It will take a little longer than a few minutes to take it all in but may come back with a comment or question later.
Yes, his article has it "shunt derived/shunt applied" - terminology I have yet to understand in the context of tube circuits, though some way more knowledgeable than I am have expressed not being 100% comfortable with it either.
@Hearinspace
some more to read about series-shunt :
https://www.vintage-radio.net/forum/showthread.php?t=133276
This is the original series-shunt circuit , where I got the name giving from :
The series feedback ecc83-pmosfet first stage is the discrete stage in JL Hood's original design performing the 20dB bass lift at 3180 uS with shelving at 318 uS and the shunt feedback pentode second stage is his opamp second stage with the 75 uS high roll-off. Shunt feedback is chosen because the shunt feedback has no gain limit of 1 but allows actual attenuation , gain lower than 1, of the input signal, something the series feedback can't , The difference is o.c. that the inverting input is used as virtual ground at the inverting input, which is unlike the "shunt" feedback around the pentode and the series feedback is to the inverting input of the input LTP , where the input signal goes to the non-inverting input , while with the ecc83 - p-mosfet circuit of my implementation , the "series" feedback goes to the cathode of the ecc83, which is not the same. So maybe it's right to speak of series-shunt feedback with opamps , but not with tube circuits.
With the tube circuit , the feedback signal to the cathode must be in phase with the input signal and when the foodback goes to the grid it has to be inverted in phase, the same as with opamps inverting-non-inverting inputs, but there is no virtual ground , so maybe no shunting, but feedback to the catrhode is called series-applied voltage feedback and I have seen feedback to the grid been called parallel-applied voltage feedback.
Systux
some more to read about series-shunt :
https://www.vintage-radio.net/forum/showthread.php?t=133276
This is the original series-shunt circuit , where I got the name giving from :
The series feedback ecc83-pmosfet first stage is the discrete stage in JL Hood's original design performing the 20dB bass lift at 3180 uS with shelving at 318 uS and the shunt feedback pentode second stage is his opamp second stage with the 75 uS high roll-off. Shunt feedback is chosen because the shunt feedback has no gain limit of 1 but allows actual attenuation , gain lower than 1, of the input signal, something the series feedback can't , The difference is o.c. that the inverting input is used as virtual ground at the inverting input, which is unlike the "shunt" feedback around the pentode and the series feedback is to the inverting input of the input LTP , where the input signal goes to the non-inverting input , while with the ecc83 - p-mosfet circuit of my implementation , the "series" feedback goes to the cathode of the ecc83, which is not the same. So maybe it's right to speak of series-shunt feedback with opamps , but not with tube circuits.
With the tube circuit , the feedback signal to the cathode must be in phase with the input signal and when the foodback goes to the grid it has to be inverted in phase, the same as with opamps inverting-non-inverting inputs, but there is no virtual ground , so maybe no shunting, but feedback to the catrhode is called series-applied voltage feedback and I have seen feedback to the grid been called parallel-applied voltage feedback.
Systux
Last edited: