I'm going to be making a number of modifications to the circuit of this headphone amplifier, to achieve a pleasantly saturated type of sound.
My first questions:
How would one remove all of the negative feedback?
And how will the output impedance be affected by that change?
A schematic is here:
Dropbox - PA1502A APPJ Schematic.png
Photos of the PCB:
Any opinions / reviews on Gemtune APPJ PA1502A | Page 16 | Headphone Reviews and Discussion - Head-Fi.org
Thanks
My first questions:
How would one remove all of the negative feedback?
And how will the output impedance be affected by that change?
A schematic is here:
Dropbox - PA1502A APPJ Schematic.png
Photos of the PCB:
Any opinions / reviews on Gemtune APPJ PA1502A | Page 16 | Headphone Reviews and Discussion - Head-Fi.org
Thanks
The global nfb can be removed by deleting R35 and E05,
but the output impedance will certainly increase considerably.
The amplitude response will vary quite a bit more, also.
Last edited:
I somehow doubt that E05 is in fact a 0.1µ, that would sound muffled. (Maybe 0.1n = 100p.)
NFB at 112:1 (L for low feedback?) is only about 6 dB anyway (rough estimate), so I don't see much point in removing the feedback path.
I have no idea what the quality of output transformers in this amp is like, but I would be tempted to convert it to parafeed output. In a classic SE output as shown here, the output transformer does two things at once: It serves as a plate load and also does impedance transformation and AC coupling. As a result of the first function, the plate current flows through it, meaning it is always partially saturated, which is detrimental to both low-end frequency response and distortion. This makes SE output transformers unusually big and heavy and usually gapped.
In push-pull amplifiers, the magnetic flux from both legs neatly cancels. If PP sounds a bit excessive, you can also go parafeed, which splits up the function of the output transformer: The actual transformer goes behind a coupling capacitor that keeps all kind of DC off, and the now missing plate load is replaced by either a suitable plate choke or some sort of active load (like a gyrator). This can be a relatively cheap option if you get the active load sorted out, plate chokes aren't as inexpensive - still much easier than a PP hack though.
NFB at 112:1 (L for low feedback?) is only about 6 dB anyway (rough estimate), so I don't see much point in removing the feedback path.
I have no idea what the quality of output transformers in this amp is like, but I would be tempted to convert it to parafeed output. In a classic SE output as shown here, the output transformer does two things at once: It serves as a plate load and also does impedance transformation and AC coupling. As a result of the first function, the plate current flows through it, meaning it is always partially saturated, which is detrimental to both low-end frequency response and distortion. This makes SE output transformers unusually big and heavy and usually gapped.
In push-pull amplifiers, the magnetic flux from both legs neatly cancels. If PP sounds a bit excessive, you can also go parafeed, which splits up the function of the output transformer: The actual transformer goes behind a coupling capacitor that keeps all kind of DC off, and the now missing plate load is replaced by either a suitable plate choke or some sort of active load (like a gyrator). This can be a relatively cheap option if you get the active load sorted out, plate chokes aren't as inexpensive - still much easier than a PP hack though.
I have this amp and has done a some major modifications.
First, thank you to haidar for posting the schematic. I have check the value on the capacitor E05 is 100pf, E03 is 100nf.
In the beginning, I converted the circuit to use 12au7, replacing R19~22 to 24k for plate resistor, and R33 to 360R for cathode. There is still too much gain for my Denon D7200 which has 25ohm impedance.
Few mod that is worth doing is
1. cut the input rca trace and directly connect a good wire from RCA into the driver input grid
2. replace the coupling cap E07 with better one (I use Mundorf supreme SIO)
3. Replace the cathode ecap C10 and bypass with film cap.
For those who still has too much gain and the input dac can control volume, my final mod was I completely bypass the driver section and connect output directly into 6v6 grid adding a grid resistor.
This made a nice improvement to the sound as there is no cap in the signal path and the pot is also bypassed. And I still have more than enough gain to drive the headphone.
I ordered an output transformer to replace the existing, we'll see how that goes.
One question is the screen voltage is 13V higher than the plate, is this optimal?
First, thank you to haidar for posting the schematic. I have check the value on the capacitor E05 is 100pf, E03 is 100nf.
In the beginning, I converted the circuit to use 12au7, replacing R19~22 to 24k for plate resistor, and R33 to 360R for cathode. There is still too much gain for my Denon D7200 which has 25ohm impedance.
Few mod that is worth doing is
1. cut the input rca trace and directly connect a good wire from RCA into the driver input grid
2. replace the coupling cap E07 with better one (I use Mundorf supreme SIO)
3. Replace the cathode ecap C10 and bypass with film cap.
For those who still has too much gain and the input dac can control volume, my final mod was I completely bypass the driver section and connect output directly into 6v6 grid adding a grid resistor.
This made a nice improvement to the sound as there is no cap in the signal path and the pot is also bypassed. And I still have more than enough gain to drive the headphone.
I ordered an output transformer to replace the existing, we'll see how that goes.
One question is the screen voltage is 13V higher than the plate, is this optimal?
Last edited: