noise from new output transistors


I am using a old Hafler DH200 power amp. It is working fine.

Not long ago I came across and bought a dozen of Hitachi 2SK1058/J162 pairs.The 1058/J162 has higher Vds ,160v vs.140v,compare to the 2SK134/2SJ49 in the DH200.Using the Hafler schematics,I constructed a output stage with parts of identical values.At power up ,use a variac,noises that sound like statics started to come at 70v. If continued to increase voltage then the noises become like waterfall(or like FM white noise in air).

I subsequently connected a tuner and speakers to the setup. The amp sounded dynamic,quiet,foucused at any voltage below 70v.I can listen to it for more than an hour with no problems.

I understand that negative feedback may play a role here.But I do not know which resistor to change.

I think bias current may also be a factor but I do not know what the original setting is. I turned the adjustable resistor but did not observe any change in noise level.

Can anyone give me a hand on this? I have the Hafler schematics if needed. Thanks!

You might have picked up an oscillation. The gate capacitance of the new MOSFETS is very likely higher. I've got the schematics for the DH-200, but not with me--check to see if there's a resistor in series with the gates for the output devices. If so, you might try increasing the value of those resistors.

Hi Grey,

Thanks for the tip. The 2SK1058/2SJ162 does have higher output capacitance as compare to the 2SK134/2SJ49 pair.
I changed the resistors in series with the output devices
increasing from the original 460 ohms to 1000 ohms and then 1500 ohms. It's much quieter now at no input connected situation. But once connected to the tuner voices and music were accompanied by the same noisy background (in smaller magnitude). It would sound correct(no noise,dynamic and clean) if I turned the variac back to about 70v. Or any voltage below 70v for that matter.

Any further thoughts?

It would be helpfull to find the negative feedback resistors and see whats happening there. Usualy with these mosfets from hitachi you have oscilations. Something that helps sometimes ( I don´t really know for this situation ) is small ceramics capacitors, to cut these oscilations. The elektor amp small crescendo uses similar fets in the new edition. One thing to do is to find the feedback resistor from output to input and put parallel to it a capacitor from 4p7F to 33pF. That usually helps. If you have long cables to the mosfets, connecting them of board you should make them shorter if possible and twist them.
as the other guys say this is hard situation and you need an osciloscope if nothing works. Good luck !
According to the schematics I've got here, Hafler went source to source. I have two different vintages of schematics. I haven't taken the time to trace down what they changed between the two, but the output configuration didn't change.
I know JFETs will run 'backwards,' i.e. with source and drain reversed. I think (but am not sure) that the same is true of MOSFETs. Seems to me that I recall that there's only a difference in the capacitance between the two hookups.
Good luck.

Hi guys,

I've got it fixed. One of the resistor was defective in the output stage.I replaced it with a good one and the amp now sounds just fine. And ,I experimented it, the source of the N channel MOSFET should go to the source of the P channel MOSFET - not the drain.

Thanks for everyone's help! I appreciated all the ideas put forward on this subject matter.