Amplifier Distortion, half wave.

[afourtrackmind]

Im currently trying to figure out why an Inter-m QD-4960 four channel amplifier is distorting. 

When the output of one channel is unloaded, the signal is clear. Once loaded, the negative side seems to phase shift into clean distortion.

If I swap the boards (there's two identical boards), the problem doesn't carry over. So the fault is on the board. When I probe around, it appears the fault lies somewhere in the first driver stage. I've swapped all of the three legged semicons with the other board. The fault stays and the other board continues to operate correctly.

When measuring voltages, Q311 2sc3298 (2sc4370 is actually on both boards) for example, there appears to be -6.1 to -7.1 vdc on the base, where the operational board has 1.1vdc at its base. At the collector of the good board I measure a half volt DC whereas the malfunctioning one is far closer to 0.005 vdc. When I remove the load, this offset is gone.
?
I've thrown the scope at and can sort of trace the fault. I'll attach some pics from the scope. Is there a big lesson here that I haven't learned yet? Ok, that's obvious, can someone help me move forward? Am I a dunce and missing something fundamental?

Any thoughts or discussion would be greatly appreciated. Thanks.

service manual link-
http://62.49.91.212/interm/audiosm1/QD-4960.pdf
 

[cncjerry]

Yes. You'll get driver distortion if the final isn't loading the driver correctly.  Also look for burned or open resistors around there. Another problem could be bias if it has a separate bias supply.  Sounds to me like one of the finals in a push pull arrangement blew.  I didn't look at the schematics.

[afourtrackmind]

Thanks for the reply.
There doesn't appear to be much in the way of heat damage or extended overheating other than the heatsink compound being dry. This thing has very well laid out cooling.
I went through and think I've checked all the resistors, after so many I can't remember if I've checked all that are in that section of the circuit. Been trying to mark them as I check them but get distracted by my noisy mind. I did find a large .47 5 watt ceramic resistor that was open and have replaced it. It didn't appear to be the cause of the distortion. At least not anymore.

If there are separate biasing for the driver and the final, would I expect to second trimpot? Or could it be set with resistors?

The outputs have been tested using the diode check method. Now that I think of it, I may have even swapped them from the other board. I'll check them again in the morning, with that, I'm reasonably confident that the outputs themselves are in working order. 

Thanks cncjerry for throwing your brain at my problem. Much appreciated. 

[Anks]

Look at the signal going into and out of the drive transistors if there is a issues here it wil be obvious. I'm unsure what the scope shot are suppose to be showing but we need to know if the drive is good.

[flynwill]

With the amp idling (no signal) and no load, is the output actually at zero like it should be?  When you attach the load does it say at zero or change?

Idling the base of Q311 should be at about 1.2-1.5V.  (Actually it should always be about 1.2-1.4V positive with respect to the amplifier output).

Q309 and it's associated components appear to be a current-limit circuit.  If your scope can accurately difference signals I would look at the difference between the base of Q309 and the amplifier output.   It looks like it's designed to cut in if the output current exceeds something in the range of 2.5-3A.  It's sensing the voltage across the 0.47 emitter resistors and when that exceeds the forward voltage of Q309's BE junction (~0.6V) Q309 cuts the drive voltage to Q311.  Q310 and it's associated parts are a mirror image of the same design protecting the negative side drivers.

Q308 and all of surrounding resistors (including the pot and the thermister) form the bias network.  It should effectively act like a constant voltage source between the collectors of Q306 and Q307.  If you measure between the collector of Q306 and the collector of Q307 there should probably be about 2.5-3V.  Note that the manual includes directions for setting the bias current with VR301.

I'll admit I can't think of any fault that would cause these symptoms, but maybe this will get you thinking about the problem.

[cncjerry]

since you have two channels, just measure the power transistors from leg to leg, forward and reverse.  I've fixed about 100 stereo amps and it is almost always a power transistor that takes out a driver or two as well as a low value current sense resistor like yours. Did you check that .47 ohm you put in to see if it is still good?  I would just order the power transistor (or IC pack) as well as the driver(s) and throw them in.  I assume you checked the power rails? Look for a diode or two in that area as well.   Make sure you didn't burn out a land pattern in the process.


Some have a separate regulated bias supply.  My krell amps are build that way.  They have a just diodes and caps on the high voltage and then a very finely regulated bias supply.  Sometimes what happens is the high voltage opens and the amp tries to draw from the bias supply burning it out.

[dom0]

base.BMP shows what exactly? Looks like half-cycle oscillation to me. Oscillating driver stages cause all kinds of crap at the output...

since you have two channels, just measure the power transistors from leg to leg, forward and reverse.  I've fixed about 100 stereo amps and it is almost always a power transistor that takes out a driver or two as well as a low value current sense resistor like yours.

+1 99 % of all faults are thermal destruction of power stage transistors.

[mzacharias]

Voltages readings while in operation are often the best way. Actual troubleshooting rather than blind replacement of parts. This is especially true if you have another functioning channel with which to compare.

For example - if on one channel you see a +4 volts measured between the base and emitter of an NPN, by definition that transistor has an open B-E junction.

A failure to develop a voltage drop across a resistor can point to another component open circuit, etc.

That said - if the outputs check OK, you could try just replacing the drivers and pre-drivers.

[afourtrackmind]

Hi all, thanks for your interest and help. I really do appreciate it.

Anks, thank you and yes, I did probe in and outs. The outputs appeared to be doing what the should have. They were amplifying the signal going in. I traced it back to the drivers and they were all behaving the same. The scope shot of the base was of q311. The other scope shots were at the speaker output, with a speaker not connected and connected.


Flynwill, thank you. You may have nailed it. What I didn't see the first few goes throught this is that a large chunk of the resisters are fusible resistors. And quite a few in parallel to boot. Rather than looking for some sort of continuity, I went over it again checking the values. It didn't take long before I realized I made a big rookie mistake. There were a number of those fusible resistors that were open or really high in value. Because their legs were effectively in parallel, I had to remove a number of them before I could accurately measure them.

I've replaced them with resistors of the same value for duration of this repair. I've gone through and checked the drivers and outputs using the diode check method. Turned it on and ta-daa, clean signal. I shut her down to get things in a state to check that everything was honkey dorey. Upon reapplying power, both channels are now in protect. Measure and a number of the outputs caved into shorts or near shorts.

cncjerry, thanks. I have checked those, and all those diodes, zeners included. In retrospect, I think this was a correct prognosis.

dom0, thanks. Yep, probably right.

mzacharias, thank as well. This is actually what I'm trying to achieve. Spending these last few days working this out, without understanding it, is a waste of time. This whole exercise has been worth it so far.
I did go over all the silicon measuring the voltages. THou shall check voltage. But, I didn't think for one minute to check voltage drop. Thanks for that little piece of kit for my batbelt. It didn't ever occur to me.

For anyone who cares:
r335- open
r340 - open
r331 - measured over a meg-ohm
r342 - open
r334 - open

After power up then subsiquent fail...
q313 and/or q315 are shorted.
I'm going to order these parts and post results.

Thanks to all.
aftm.

 

[flynwill]

Sounds like you're on the way to getting it running again.

If you have one available you might use a variac to bring it up slowly next time.  It's a lot easier and safer to check that everything ok with the rails at +/- 10-15V rather than the normal 70 or whatever that amp uses.  I don't think there is anything that will be harmed by the under voltage, but it also looks like you can totally disconnect the pre-amp supplies and still test the amp.  You might also want to adjust the bias network to it's minimum while testing then set it per the factory procedure.  (The amp should run just fine with the bias all the way down, it will just have more cross-over distortion than normal).