Well, if you are willing to experiment a bit more, there are reports (conflicting, I have to say), about using a relatively small shared cathode resistor, non-bypassed, in a push pull output stage to reduce odd harmonics. I have tried simulations, and it does indeed reduce distortion, but it is not a clear cut, the effect varies with the output power level, signal frequency. It probably makes no sense, especially taking into account that speaker impedances dances around a lot. But if you feel adventurous...🙂
For example: https://www.diyaudio.com/community/threads/distortion-neutralizer-for-odd-harmonics.111501/
Yep keeping the 0V separate for the two channels and just joining at the inputs will solve the differential hum issue. If you don't connect 0V directly to chassis then use a couple of back to back diodes (1N540X) + 15R + 22nF all in parallel for each 0V to chassis. The 22nF is for RFI the diodes in case you short HT to chassis. That or the hum buck input. I would still leave spaces to fit a dominate pole on the plate of the EF86.
@jcalvarez,
I'm game! I'll whip that up and do some measurements, however if the effect is small it may not show up at all.
I'm game! I'll whip that up and do some measurements, however if the effect is small it may not show up at all.
@jcalvarez,
Alright, I have 2R2 from the cathodes to a common 10R up and running now, not seeing an improvement thus far, slightly raised THD from 10-30 Watt output. I'll try some different values and a parallel capacitor, also of different values, to see whether there's an optimum somewhere.
Alright, I have 2R2 from the cathodes to a common 10R up and running now, not seeing an improvement thus far, slightly raised THD from 10-30 Watt output. I'll try some different values and a parallel capacitor, also of different values, to see whether there's an optimum somewhere.
@jcalvarez,
Please have a look, the darker plot is two individual 10R cathode resistors, the lighter plot is 2R2 from each cathode and a common 10R bypassed with a 470uF capacitor, I would say this is a significant difference?
THD vs. power 1kHz - Shared cathode resistor
Please have a look, the darker plot is two individual 10R cathode resistors, the lighter plot is 2R2 from each cathode and a common 10R bypassed with a 470uF capacitor, I would say this is a significant difference?
THD vs. power 1kHz - Shared cathode resistor
Okay, tried a few different combinations, the best compromise seems to be 4R7 from each cathode to a common 4R7 bypassed with a 470uF capacitor. Below plots show two individual 10R cathode resistors (dark plot), 2R2 from each cathode and a common 10R bypassed with a 470uF capacitor (light plot), the 2nd darker plot is the 4R7 from each cathode to a common 4R7 bypassed with a 470uF capacitor, which shows the best results thus far.
THD vs. power 1kHz - Shared cathode resistor
THD vs. power 1kHz - Shared cathode resistor
This image shows the difference more dramatically, the dark plot is the original individual 10R cathode resistor, the light plot the the 4R7 from each cathode to a common 4R7 bypassed with a 470uF capacitor.
THD vs. power 1kHz - Shared cathode resistor
THD vs. power 1kHz - Shared cathode resistor
Could you measure the open loop behavior (THD vs. power, frequency response @1W, gain)? How much NFB is used?
I think it is great, probably will be difficult to go significantly lower than that at 40W.
Sure, I can do that tomorrow. You just want the OL behavior with the feedback from the output disconnected right?
I am too interested in the gain difference between open and closed loop. I would imagine its around 20dB.
One thing I not seen (or missed it) is the 1KHz square wave response (closed loop). Could you check and pop on the chat just to check HF stability while your at it.
One thing I not seen (or missed it) is the 1KHz square wave response (closed loop). Could you check and pop on the chat just to check HF stability while your at it.
Ssassen,
You said:
"The fact that some tube amplifiers that have an LTP also have a AC balance pot probably wasn’t a mistake?"
True, sometimes a pot is necessary:
1. The CCS is not high enough impedance, versus the impedance of the phase splitter cathodes. (And in regards to plate loading . . . you can purchase 0.1% plate loads, and 0.1% next stage Rg resistors, how close do you need to be?)
2. The output tubes are not gain matched for the circuit they are in.
3. A not so good output transformer is not AC matched (unlikely for any transformer most of you would purchase).
Other than that, the only other cause is the Witches Pot and Curses in Shakespeare's McBeth.
You said:
"The fact that some tube amplifiers that have an LTP also have a AC balance pot probably wasn’t a mistake?"
True, sometimes a pot is necessary:
1. The CCS is not high enough impedance, versus the impedance of the phase splitter cathodes. (And in regards to plate loading . . . you can purchase 0.1% plate loads, and 0.1% next stage Rg resistors, how close do you need to be?)
2. The output tubes are not gain matched for the circuit they are in.
3. A not so good output transformer is not AC matched (unlikely for any transformer most of you would purchase).
Other than that, the only other cause is the Witches Pot and Curses in Shakespeare's McBeth.