Is this a differential pair? McIntosh MC2300

McIntosh 2300 

driver board  Q102  Q104

looks like a diff pair drawn different?

color coded transistors not matched very good, hfe = 260 and hfe = 344

2nd board , hfe = 270  hfe = 400

should i get some better transistors?  thanks!

> Is this a differential pair?

Good eye. (I had to stare the first time I saw this.) Bad drawing.

I don't see that hFE matters much here. And that would be in line with Mac conservatism.

Is it working OK?

right side oscillates past 90 volts AC on the power cord

common problem from what i hear,

new MJ15003G power transistors installed by previous owner,

there is some talk about dropping emitter resistors down to .18 ohms and making sure they are wire wound for squelching inductance,

also a ceramic cap is removed  on driver board as per a service blog but we do not know which one.

Post on DIYAudio.com . There are a few Mac mavens there. (Also a lot of clowns--- do not feed them!) The ones who have honestly worked-on and know this amp will be clear from their writing.

> transistors not matched very good

The currents are un-matched, by design. 1mA in one side, 2.5mA the other. So there's gonna be >20mV of output DC offset, except the autotransformer (not seen here) will keep it off the speaker.

Whether hFE is 100 or 500 makes little more difference. In fact R112 to R104+R106 marginally reduces the built-in offset, less with high hFE, but in any case insignificant.

Oscillation is not here. Replaced output devices changes compensation, and "fast" (modern) devices may be as bad as slow devices. Someone will know but it sure aint me.

gnd for the diff amp goes to the same terminal as the shield,  wonder if we should change this, possible gnd loop could be cause of oscillation,?

should be OK, where does the ground from the speaker return connect? You don't want speaker return current corrupting + input.

I would review the amplifier compensation scheme(s) for possible stability concerns, especially if modern devices are faster than old ones.

JR

PS: Operating amp at higher closed loop gain should be more stable, so maybe experiment with that, to determine how marginal the stability is. That said I am not a big fan of redesigning presumably stable designs...  unless they never were all that stable.

thanks for the help!

we hooked up some external power so we could test the driver board and a set of power modules and it works fine, very stable,

we injected an AC signal into the driver input, generator maxed out at 5 volts and the amp was not distorting at all, 

here are the C volts for the driver section, not a lot of gain going on, only puzzling thing is how Q106  gets so much collector voltage swing without a lot of base swing, must be a current type of deal,

there are a lot of corroded push-on terminals that we are replacing so that might be the cause of the osc. problem on the rt channel.

> how Q106  gets so much collector voltage swing without a lot of base swing

The output stage is high current gain, so not much loading on Q106.

Then superficially the gain of Q106 is about 3,600 / 22 or pretty large. Indeed 0.080V at base may be 13V at collector. 10.5V measured means the output stage loads, but not a lot (output stage input Z may be near 14k).

thanks for the help!

here is a DC voltage chart we measured of the driver stage for left and right channels, one board had more DC offset, it is the one with the different V-eb volts for the input pair,

McIntosh has -40 volts on both sides of the 22 ohm resistor so they were giving approximate volts,