Hi there,
I am in the process of repairing an old Sony ES amplifier (very nice build ) and there are some problems with stability in one channel. It seems like HF-instability or oscillation. It is only present ones in a while, and makes the amp shut down (protection).
I think I have solved it (might be some weeks before I know for sure), by adding 220R gate stoppers right at 2SK389 pins. But if not, my next thought will be to make the VAS compensation capacitor bigger (from 6.8pF to 22pF?).
But a more attractive thing, would be to modify the gain stages more radical. The gain of the amp (CL) is 45dB and much higher than needed (I need perhaps 10dB less). So looking at the circuit, some options are present:
A: Change the feedback resistor to lower CL-gain 10dB. This might not be good as the amp is already instable and adding feedback level might make things worse.
B: Lower open-loop gain by:
-Adding source resistors to 2SK389 diff. amp (from zero to 47-100R)
-Changing current mirrors in top of diff. pair to resistors
-Changing the source resistor on the VAS 2SJ313 from 33R to 220R (seen on alike Sony models)
-Replacing the VAS current source from active (2SC3514) to boot-strapped resistors (AKSA style)
Might use all of the above options?
My problem is basically that I only have a multi meter instrument, so the possibility to monitor the result is limited. So I will have to work on safe grounds and follow advises.
Given an amp that is designed stable, with correct HF-compensation as a starting point:
1) Does lowering the loop gain always means more stability?
2) How should the feedback network be changed? If closed loop gain must 10dB down and the network is made for this, will lowering the openloop gain by 10dB or more ensure same or better stability as standard?
I am in the process of repairing an old Sony ES amplifier (very nice build ) and there are some problems with stability in one channel. It seems like HF-instability or oscillation. It is only present ones in a while, and makes the amp shut down (protection).
I think I have solved it (might be some weeks before I know for sure), by adding 220R gate stoppers right at 2SK389 pins. But if not, my next thought will be to make the VAS compensation capacitor bigger (from 6.8pF to 22pF?).
But a more attractive thing, would be to modify the gain stages more radical. The gain of the amp (CL) is 45dB and much higher than needed (I need perhaps 10dB less). So looking at the circuit, some options are present:
A: Change the feedback resistor to lower CL-gain 10dB. This might not be good as the amp is already instable and adding feedback level might make things worse.
B: Lower open-loop gain by:
-Adding source resistors to 2SK389 diff. amp (from zero to 47-100R)
-Changing current mirrors in top of diff. pair to resistors
-Changing the source resistor on the VAS 2SJ313 from 33R to 220R (seen on alike Sony models)
-Replacing the VAS current source from active (2SC3514) to boot-strapped resistors (AKSA style)
Might use all of the above options?
My problem is basically that I only have a multi meter instrument, so the possibility to monitor the result is limited. So I will have to work on safe grounds and follow advises.
Given an amp that is designed stable, with correct HF-compensation as a starting point:
1) Does lowering the loop gain always means more stability?
2) How should the feedback network be changed? If closed loop gain must 10dB down and the network is made for this, will lowering the openloop gain by 10dB or more ensure same or better stability as standard?
Have you actually had a scope on it to see if its unstable ? I see you only have a meter so I guess not. The protection circuit can trip for many reasons such as simple DC faults.
I wouldn't advise changing any component values tbh. If you lower the gain by altering the feedback then generally the amp will be less stable and closer to oscillation.
Although its possible to alter a design greatly you need far more than just a meter to even attempt anything like that. Even then all you are doing is altering for alterings sake.
Fix the fault properly Dries and intermitent semiconductors are most likely (such as drivers or devices that run hot)
I wouldn't advise changing any component values tbh. If you lower the gain by altering the feedback then generally the amp will be less stable and closer to oscillation.
Although its possible to alter a design greatly you need far more than just a meter to even attempt anything like that. Even then all you are doing is altering for alterings sake.
Fix the fault properly
Dries and intermitent semiconductors are most likely (such as drivers or devices that run hot)- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.