Help to elminate A75 front end oscillation

[ckwong99]

Dear friends,

After completed A75 front end found both channels voltage levels and specified voltage check point can reach acceptance level. Also, use oscilloscope check both channel properly amplified input signal to want . Somehow the problem be found was there only one channel can adjust DC level near to zero when without any input signal. No matter how to adjust P1 & P2 can’t get stabilized DC offset value. In additional I could measure the output by DMM between the center point Z3 and Z4 to GND have AC signal with frequency about 350K(Hz). I though there must have an oscillation happened on this channel.
Can anyone tell me how to counter-act this effect (only single channel) ?

First let me state my additional setting on A75 front end both channels :
1) 10K resistor between signal input and GND.
2) 100n parallel with 100u between +Reg 50V and GND.
3) 100n parallel with 100u between -Reg 50V and GND.
4) No folded-cascade (R21, R22, C5 and C6).
5) C9 and C10 added value are 30p.
6) Unbalance input only.
7) No R81 100% feedback.
8) C3 change to 47u.
9) Q11 change to 2SK216 (due to I use Hitachi MOSFET as my back end)

CK

[BobEllis]

You might try adding around 1 uF film caps to your rail decoupling bank.

Check/change the way your power leads run - mine were sensitive to lead dress. Twist them together.

Does shorting the input to ground as NP suggests in the setup phase kill the oscillation? I have one channel (of 4) that oscillates unless connected to a low impedance source. Since it doesn't oscillate in the system I never bothered trying to fix it.

DC stability is VERY sensitive to air currents. Be sure to measure in dead air - even at the bottom of a cardboard box would make life easier. If measuring in your case, adjust, put the top on and wait a few minutes before checking. I wish I'd used multi turn pots in mine.

[ckwong99]

Hi Bob,

Thanks for your hints.

Now the oscillation was gone , I just short signal input to GND. As well the DC level can be sticky controlled. But not sure when it be connected by pre-amp the problem will occur again :q

My front end is only under testing phase individually on my bench.

BR,
CK

[BobEllis]

Just keep an eye on it when you put it in the case. Test and adjust the front end again before powering up the output stage. Its offset and bias will be quite different than sitting on the bench.

As long as you have a preamp connected and powered up before the amp you should be fine. Not great practice, but in my case I know that nobody will disconnect the preamp with the amp on.

[jupiterjune]

My A-75 oscillates if turned on when not connected to a preamp. I read somewhere that using a lot of NFB can cause an amp to oscillate. I never investigated the problem, I just make sure I don't turn it on when not connected to a preamp.

JJ

[ckwong99]

Oh , I see that and I will not let my wife use it, otherwise it may fire my home :>

I am also a lucky one because my only front end's single channel have this problem .

Next , I will construct the output stage. Can you tell us any reminding ?

On a single channel total have 6 pairs of output FETs. I divided them into two rails mount on the heat sink. The front end output planned to connect at output FET stage in a single entry point from the leftmost of the output rail. Is it the best way ? How to prevent the output stage oscillate ?

CK

[Nelson Pass]

Originally posted by jupiterjune
My A-75 oscillates if turned on when not connected to a preamp. I read somewhere that using a lot of NFB can cause an amp to oscillate. I never investigated the problem, I just make sure I don't turn it on when not connected to a preamp.
JJ

Is this a balanced connection, or single ended?

Oh, I see that it's single-ended. It seems possible that the output
of your channel is being picked up by the input. Do you have in
input cap to ground, and/or have you tried lowering the value
of the input resistor to ground? Do you have a resistor in series
with the input connection?

[BDP]

I had a similar problem with oscillation with no input attached. It first showed up when I originally tested the amp with the output stage laying on the bench. The output network let out a gasp of smoke along with my enthusiasm. In those days the output MOSFETS were mighty expensive and I thought I had smoked them!
For a long time I was not concerned because the source equipment was balanced and with the switch in the balanced position the impedance was low enough to prevent it from oscillating. No problem, as long as I didn't forget! As time went on I experimented with different input values and settled on 47.5K ohms. I adjusted the feedback and other input values to reflect the change.

[ckwong99]

Posted by Nelson : Do you have in input cap to ground, and/or have you tried lowering the value of the input resistor to ground? Do you have a resistor in series with the input connection?

According to the circuit input cap to GND is C1=39p , input resistor in series is R4=475 ohm and input resistor to GND is R30=75K.

Already lower R30 to 10K but no help on oscillation. How do you think R4 increase to 2K ?

I prefer only to connect this AMP in UNBAL mode so I omitted R28 and R5 . Not sure R2 and R29 should be removed from the circuit ?

CK