Fixed bias has changed after volume position is middle

[nataphong_audio]

Since you have three gain stages you need to invert phase in the output transformer.
That's not always indicated in a schematic from a producer, but for a DIY it should.

Look at this and think about it, since I suspect you have made some errors in your schemetic.
I doubt that 8 ohm tap shoulf be grounded. 16 ohm tap is more likely to be grounded.
By inverting phase you get this:
COM will be new 16 ohm tap
4 ohm will be new 8 ohm tap
8 ohm will be new 4 ohm tap
16 ohm will be new ground tap

View attachment 1341207

As seen in the schematic the phase are inverted between primary and secondary side.
Have you checked if you connected the right EL84 to the phase indication of the output transformer?
If you mix them you will not get the necessary signal inversion.

View attachment 1341214

This is my output transformer spec and wiring spec.

 

[nataphong_audio]

It seem can fix this problem but how to calculate the properly value of capacitor ?

 

[flex2]

It can be a bit confusing when there is no clear phase indication on the transformer.
My best guess is that P1 and 0 is IN phase since they are to the left and have a "-" infront of them
that might indicate the phase.

Then this coupling would be "IN" phase without necessary inverting. Green arrows indicates phase.

So I suspect that you should connect cables like sketch below to get an inverted signal.

With your suggestion you are inverting the secondary output side instead.
Since you are grounding the 8 ohm tap you will end up with COM and 4 ohm tap IN phase and 16 ohm tap OUT of phase.
Probably will work but it's not the right way to do it and feedback will not work properly since the load, the loudspeaker, will be driven
and loading the output transformer OUT of phase. You can't use 16 ohm tap.
I think COM or 16 ohm tap should be grounded if you want to invert the secondary side as per your suggestion.

I would start to check how you have wired the primary side...

 

[flex2]

It seem can fix this problem but how to calculate the properly value of capacitor ?

View attachment 1341217

Regarding the capacitive load on the anode, you will only benefit from that if VR1 and R9 are lowered to something around 20k for VR1 and 10kohm for R9.

The original have a open loop bandwidth of around 18kHz when VR1 in end positions and 9kHz when VR1 are in center.
These figures will double with each 6dB of NFB.

If VR1 are 20kohm and R9 are 10 kohm, and a 100pF cap is placed between anode and ground you will get this:
Open loop bandwidth of 19kHz when VR1 in end positions and ~16kHz when VR1 are in center position.
So less dependence on position of VR1, and in general a better capacity to drive the first tube which means slightly less distortion.
Even if better I think the original seems quite OK even if I personally would aim for a open loop bandwidth of at least 20 kHz regardless
of position of VR1.

I think you can stay with the original or maybe lower VR to less than 25 kohm. That will give almost the same effect.
But then you should increase C9 from 0,22u to 1uF to not loose bass performance.

 

[nataphong_audio]

OK let me try as your suggestion.

Thank you so much.

👍

 

[nataphong_audio]

Regarding the capacitive load on the anode, you will only benefit from that if VR1 and R9 are lowered to something around 20k for VR1 and 10kohm for R9.

The original have a open loop bandwidth of around 18kHz when VR1 in end positions and 9kHz when VR1 are in center.
These figures will double with each 6dB of NFB.

If VR1 are 20kohm and R9 are 10 kohm, and a 100pF cap is placed between anode and ground you will get this:
Open loop bandwidth of 19kHz when VR1 in end positions and ~16kHz when VR1 are in center position.
So less dependence on position of VR1, and in general a better capacity to drive the first tube which means slightly less distortion.
Even if better I think the original seems quite OK even if I personally would aim for a open loop bandwidth of at least 20 kHz regardless
of position of VR1.

I think you can stay with the original or maybe lower VR to less than 25 kohm. That will give almost the same effect.
But then you should increase C9 from 0,22u to 1uF to not loose bass performance.




For the 1uF can I use the ECAP(electrolite) ?

 

[flex2]

"For the 1uF can I use the ECAP(electrolite) ?"

Yes, but it would not be my first choice.
Try it, and if you are sure that you have no DC at input you can short it to hear what kind of sonic impact it has.
If you can't hear a difference between 1uF ECAP and peice of wire you should contuniue using ECAP.