Asking for help repairing TOA KD-2 Power Section

[kalimusic]

Please see the attached schematic.
KD-2schematic.pdf
I've been working to restore this unit and find there is no idle bias voltage. the unit will pass signal with no load. It will also pass signal when loaded with an 8Ω load, but at a bias voltage of 16mv, Q711 begins to run away and overheat, leading to failure.

I have replaced these components which either failed or read bad on DVM:
Q710, 711, 708, 718, 719, 720, 721
R 759, 757, 729, 751, 749

I've also had to hand-wire a number of connections due to pads lifting from the board.

All the silicon components read good in circuit and resistors and diodes are within spec. I expected to find some idle bias voltage measured across R751 or R749, but no voltage appears until a signal is inserted with the amp under load.
I reckon I've overlooked something, likely on the Q711 side of the unit.
I'd be very grateful for any help offered.
Regards,
Steve

[Enzo]

It passes signal into load, OK, does it do it cleanly? And does it sound OK?

Where did you get a 16ma target? Across a 0.2 ohm resistor, I get about 80ma of idle current. Seems like a lot to me.

The parts you replaced are the outputs, drivers, and predrivers, but R729 is way over in the balance control. How did that burn up??? In any case, do check that trim pot next to it for total health.

If ti runs, turn the bias to the cold end of the settings. Scope the output and run a sine wave into it. Look for the crossover notch, then dial up the bias slowly watching for the crossover notch to shrink. Once it just disappears, back it off a hair. Now nice clean sine wave, no notch. Assuming that doesn;t blow it up, what voltage is dropping across the 0.2 ohm resistor?

Are you monitoring mains current draw while working with bias? I do, and I can see the draw start to climb when I get to the bias point, any further advance just heats up the outputs.

Also scoping it will tell you if it is falling into RF oscillation.

[kalimusic]

Hi Enzo,
Thanks for your reply.
1. I mistakenly included R729. I did not replace that. I did replace R762, 1k which is between Q711 and Q714.
2. Idle Bias target is probably around 2.5 ma, not 16ma.
3. It will pass signal into 8 Ohm load, but at 16ma (very low output) Q711 begins to run away and if I continue to increase the signal it will overheat and blow up.
4. Yes, I am constantly monitoring the mains current draw and the temperature of Q711

The 2 issues are:
1. there is no measurable idle bias current across the .2 ohm resistors no matter where VR702 is.
2. Q711 goes into run-away as signal is applied.

[Enzo]

The goal of bias is distortion, or lack thereof, not a particular current. The current thing is secondary, most times at the specified current, the amp goes out of crossover distortion. Have you checked for that?

I am also confused. You say it shows no voltage at idle - assuming you mean across the ballast resistors. Yet you then say at 16ma it starts to run away. Where did the 16mv come from if there is no voltage? And if I am missing something and you can sweep from zero up through 16mv, why are we turning it up that high?

Is everything solid on its heat sink? Especially the temp sensing diode or transistor? Look on the other side of the outputs, what is the voltage relationship between opposing bases?

[kalimusic]

Thanks for you reply.
I erred in writing "ma". I meant to write "mv". There is no voltage at idle (no signal) and the voltage increases across the resistors when signal is applied and the amp is under load. I am not adjusting the bias trim pot. The fact that the voltage increases is really not the issue, which is that Q711 wants to blow up. I just used the 16mv number to indicate at what point Q711 begins to run away.

All the components are solid, tight on the sink.

I can bring the signal and output way up on the scope with no load attached. It seems Q711 has no bias (?) and is the only component that overheats. Can you be more specific about which transistors you want me to measure? I'll check the voltages you're asking about and get back to you.

[Enzo]

OK, so you can crank the amp with no load and it works? Then run a sine wave through it, scope the output, and turn up the gain on the scope. We don't care if the top and bottom of the wave are off the top and bottom of the screen, we only care about the crossover notch. How large is it, and does it change with bias adjusting. With a load, we want to know if the amp breaks into RF.

Just left of Q720 is a 3.3 ohm resistor, is it open or burnt?

Q719 is the bias transistor, it adjusts the voltage between the bases of Q710 and Q720. Ought to be very roughly 2 volts. What do you get? And does adjusting the bias trimmer change it?

Between the bases of Q711 and Q721, there ought to be about half that voltage.