NAD C270 (C272) Power Amp Protection Mode - Stray Negative Voltages

Hi everyone,
I bought a non-working NAD C270 power amp a couple of weeks ago. I've made quite a bit of progress but there is still a problem somewhere (I believe) in the LEFT channel Power Amp Board. I’ve already repaired the oft-failed 68V power supply (see photo).

68V supply area problems.jpg

I wouldn't say I am at the end of my tether, but I am out of my depth understanding the fault on this one. I've read a number of forum entries here for fixing the NAD C272, which is very similar and I am using the C272 schematics because I can't find ay C270 ones. Clearly many people here on diyAudio understand schematics waaaay better than me and also some have first hand experience of these models. So any guidance or insight would be fantastic.


Current Symptoms
Amp stays in protection mode unless L channel Power Amp Board disconnected. Offset bias test points (TP2 and TP8) giving more than 56V on L channel compared to 25mV on R channel. Looking in more detail shows that the Power Amp Board comprises of a 68V end and a 64V (power transistor) end. These two ends are joined by only two connections; R328 (100 Ohm) and R356>R319+R322 (a L-FDBK call out on the schematic). All around R328 I am seeing high NEGATIVE volts compared to the right hand Power Amp Board (relative to ground).

Volts - Left PAB FAULTY_L.jpg

Volts - Right PAB WORKING_R.jpg


I can't work out exactly which end of the board (68V or 64V) these volts are coming from. Initially I suspected Q311 as it was showing -64V at the collector compared to -0.03V on Q411 (same transistor on R Power Amp Board), so I replaced Q311 and then Q310 (positive side equivalent) and then the 4 diodes (D303. D304, D305, D306). Just about every other component (save a few slightly low capacity electrolytics) seemed to test ok in circuit.


TLDR;
I've got high NEGATIVE volts running along the signal path (e.g. at the offset bias test points) on the LEFT channel only. The amplifier runs correctly with just the right channel in circuit. I can't work out how to begin finding the shorted/failed/open component on the L Power Amp Board. I've marked some schematics with my measured voltages (with CZ719 disconnected). Can anyone read my numbers into the schematic logic and suggest where I look next to find the source of these high negative volts? Main schematic attached. Any help appreciated!
 

Attachments

  • hfe_nad_c272_service_schematic.pdf
    409.7 KB · Views: 138
Hi pumptronics,

The circuit in your second post is a Zener regulator and is probably operating properly, given the -16VDC you note at the base of Q307.

As you've observed, there are many voltages that are wildly abnormal and it's difficult to determine what's cause vs. effect--- and what collateral damage may have occurred after the originating failure. So after a few voltage measurements, I'm going to suggest opening the feedback loop to help isolate problems.

I think you've zeroed into an interesting node. There appears to be 14VDC across R328, or about 2W dissipation. Inspect it for damage. Similarly, R334 is likely to be extremely hot; you mentioned -56V at output; if shunt regular IC301 survives, there would be about -51VDC at the bottom of R334, and roughly 108 V across it, and producing 5.3W dissipation.

So I suggesting lifting one end of R328 to open the feedback loop, and tacking a 10k resistor from the bases of Q313/Q314 to ground. The output stage should bias to about 0V but probably won't. This should allow easier troubleshooting for failed transistors in the outputs. Post some voltages if you get stuck.

Also, the input section is now open loop and should should saturate at an appropriate rail--- it should try to correct for error voltage at the output. But there is likely a problem in this circuitry: I note +16V at the base of Q309, suggesting it is failed. Cause or collateral causality?

Let us know how things go, what you encounter. Good luck!
 
Last edited:
Thanks so much for getting back and for giving me some direction!
The circuit in your second post is a Zener regulator and is probably operating properly, given the -16VDC you note at the base of Q307.
Ok yep that makes sense it being a 18V Zener.
So I suggesting lifting one end of R328 to open the feedback loop, and tacking a 10k resistor from the bases of Q313/Q314 to ground. The output stage should bias to about 0V but probably won't. This should allow easier troubleshooting for failed transistors in the outputs. Post some voltages if you get stuck.
So I did something similar to this already by removing R328 (volts in green) however I will try it again with the 10k resistors on Q313/Q314 today.

Volts - Left PAB R328 Removed_excerpt.jpg

There appears to be 14VDC across R328, or about 2W dissipation. Inspect it for damage
I have measured R328 in and out of circuit and it is measuring around 100 Ohm as per spec.

Interesting you have considered IC301 to possibly be stressed, I have replaced IC302 but it was my last new TL431 so am awaiting more in the mail.

But there is likely a problem in this circuitry: I note +16V at the base of Q309, suggesting it is failed.
I have now replaced Q309 (new SA1015GR) and Q308, Q312 (new SC1815GR). I have also swapped Q310 (pre-used SB649A) and Q311 (pre-used SD669A). None of these made any noticeable difference.

If it fleshes out the picture at all, I have done some voltage measurements without the 64V rail connected. I still see -12.6V at R328 but -0.03V at R428 (the working right channel side). I will go and get my hands dirty now and let you know what I find.

Thanks!
 
So I suggesting lifting one end of R328 to open the feedback loop, and tacking a 10k resistor from the bases of Q313/Q314 to ground. The output stage should bias to about 0V but probably won't. This should allow easier troubleshooting for failed transistors in the outputs. Post some voltages if you get stuck.
Ok, I've just done this....I don't quite understand what we've done there, but I like the results. All the supply cables are connected to the faulty channel board. R328 is removed and a 10k resistor goes where the R328 was (64V end) to ground. At the offset test point TP2 I am now seeing about 120mV and the relays have clicked; the unit is no longer in protection mode!

So what does this mean? D307 fault? Or IC301 as you suggested? I have some BAV21's in the mail also...can I try a 1N4148 temporarily?

It's so rewarding to see these tiny steps of progress. Thank you again.
 
Volts - Left PAB R328 Removed_and_earthed_excerpt.jpg


So I suggesting lifting one end of R328 to open the feedback loop, and tacking a 10k resistor from the bases of Q313/Q314 to ground. The output stage should bias to about 0V but probably won't.
So this is how I did it. And the full -68V are coming through to R328 when it is disconnected from the other end. As shown this gives a bias offset of about 120mV and the amp is no longer in protection mode.

So what does this mean? D307 fault? Or IC301 as you suggested? I have some BAV21's in the mail also...can I try a 1N4148 temporarily?
I already tried a 1N4148 in place of BAV21 this afternoon. I also tried another IC301 (a pre-used TL431). There was no change. Could the fault lie in Q305, Q304 or Q307 to give this high negative 68 volts?
 
I hadn't realized that R328 was out of circuit, so my inference of 14V across R328 was bogus.

A bit of explanation: With bases of Q313/Q314 floating, the voltage you observed there was governed by transistor leakage currents, so more or less random. The 10k to ground provided bias voltage and I suggested a relatively large value as a precaution against defective transistors. The circuits between Q313/Q314 and the output have essentially unity voltage gain, so whatever voltage is presented at Q313/Q314 is reproduced at the output.

So 120mV out is very encouraging, but there's still the possibility of damage in the output stages. So I suggest a few measurements while the feedback loop is still open. If IC301 and IC302 are operable, there should be 5VDC across C311 and C312. I also suggest testing the output stages now, as it will be easier to troubleshoot with the feedback loop open. I suggest the following: lift that 10k resistor from ground and instead drive it with a few volts--- perhaps a 9V transistor battery if you don't have a more convenient adjustable source. You should see about 9V at the output. Reverse battery polarity, and you should see -9v. Repeat both tests with 1k load on output, then 100 Ohm. You may have to work around intervention of protection relay. If all this looks good, the output stages are probably OK.

Now focusing on the input section, with replacement of Q309, its base voltage should be only about 0.6 V below the the positive rail. With 120 mV present at output, open loop gain from the input should drive R237 toward the negative rail, and definitely so with +9V at the output. -9V at output should drive R237 toward the positive rail.

Again, please advise if actual results are different, and good luck!
 
Just discovered your earlier post. -68V is very encouraging. Now you have to nudge the input section to confirm that it can also swing to the plus rail when circumstances dictate. When R328 is reinstalled, that voltage will go where it needs to to deliver the correct drive voltage to the output stage. We just want to be confident that all else is correct before closing the feedback. Test no load at first.

The fact that your experiment yielded 120mV at the output pretty much absolves the diode.
 
Last edited:
Thanks BSST and Wakh

Going to start working through your suggestions....conscious of time zone differences. And apologies for the many, short posts.
So I suggest a few measurements while the feedback loop is still open. If IC301 and IC302 are operable, there should be 5VDC across C311 and C312
Just set this up (I still have the 10k resistor to ground). Volts across C311 = 5V, so IC301 is ok? Volts across C312 = 0.01V so IC302 is not ok? Both capacitors have been replaced. IC302 was a new component.
 
Now focusing on the input section, with replacement of Q309, its base voltage should be only about 0.6 V below the the positive rail.
So I didn't flag this as a fault on the L side (Q309) because the R side (Q409) has similar volts but works fine.
With 120 mV present at output, open loop gain from the input should drive R237 toward the negative rail,
Do you mean R327? Thanks
 
Yes, I meant R327. Sorry for typo.

I had not noticed the similar measurement in the other channel. Please double check that measurement point in both channels.

No volts across C312 seems like a problem. You can compare with the working channel for reassurance/comparison.

Voltage at Q318 emitter should be near 0V and there should be about -5V at R344; R344 should be hot. If issue not discovered, with power off, inspect work in that region, especially since you've done repairs in that area. Do resistance checks in in that region for shorts, especially across C312. If you find this path is shorted, you may have remove C312 and/or IC302 till short is found. Make sure Q318 base-emitter isn't shorted.
 
I had not noticed the similar measurement in the other channel. Please double check that measurement point in both channels.

Returning to this issue of the base voltage at Q409: +66V at R424 implies there should be about +67.2V at emitter of Q409, and hence about +66.6V at base of Q409. But you report +17V, so something isn't making sense at this node. Same issue seems to be present in left channel.
 
I suggest the following: lift that 10k resistor from ground and instead drive it with a few volts--- perhaps a 9V transistor battery if you don't have a more convenient adjustable source. You should see about 9V at the output. Reverse battery polarity, and you should see -9v. Repeat both tests with 1k load on output, then 100 Ohm.
I've just done these measurements. I noticed that with the circuit open at R328 the machine makes a kind of fizzing/pressurizing inrush noise. It doesn't do this with the 10k resistor earthing it. I didn't think much of it but then I noticed that the main heatsink got quite hot within only a few minutes of off-on-off testing. I don't know what that indicates because I think these machines run hot anyhow. Here are the numbers:

+3V @ base of Q313/Q314 produces +2.2V @ TP2
-3V @ base of Q313/Q314 produces -3.8V @ TP2
+5V @ base of Q313/Q314 produces +4.2V @ TP2
-5V @ base of Q313/Q314 produces -5.7V @ TP2
+12V @ base of Q313/Q314 produces +11.3V @ TP2
-12V @ base of Q313/Q314 produces -12.4V @ TP2

Now with 1.2k load (I applied this load across CZ719 as that is upstream of the speaker relays):

+5V @ base of Q313/Q314 produces +4.2V @ TP2
-5V @ base of Q313/Q314 produces -5.5V @ TP2
+12V @ base of Q313/Q314 produces +11.2V @ TP2
-12V @ base of Q313/Q314 produces -12.1V @ TP2

Now with 100 Ohm load across CZ719:

+5V @ base of Q313/Q314 produces +4.2V @ TP2
-5V @ base of Q313/Q314 produces -4.9V @ TP2
+12V @ base of Q313/Q314 produces +11.2V @ TP2
-12V @ base of Q313/Q314 produces -11.6V @ TP2

Only pattern I can see from that is that the negative inputs give lower values as the load is applied compared to the positive which stay the same. So do these figures mean the output transistors are behaving normally?
 
Returning to this issue of the base voltage at Q409: +66V at R424 implies there should be about +67.2V at emitter of Q409, and hence about +66.6V at base of Q409. But you report +17V, so something isn't making sense at this node. Same issue seems to be present in left channel.
Thank you for making me re-visit these...clearly there was some fluffery afoot in my measurements.
Voltage values are:
Base of Q409 = 65.8V
Base of Q309 = 66.3V

there should be about -5V at R344; R344 should be hot
So the voltage across R344 is 5.5V as you predicted. However the voltage across it's opposite friend (R334) on the positive rail is 115V DC. That's why that one becomes skin-searingly hot within seconds.

Now the voltage measurements are done I will whip the board out of the machine again and begin inspecting the C312/IC302 area as well as Q318.
Замените конденсатор С301 на входе усилителя

Я также проверю этот конденсатор и поставлю новый, если у меня есть запасной ... C301 показал довольно низкие значения.
I'll also check this capacitor and fit a new one if I have spare....C301 was reading quite low.
With 120 mV present at output, open loop gain from the input should drive R327 toward the negative rail, and definitely so with +9V at the output. -9V at output should drive R327 toward the positive rail.
So I don't fully understand this statement...or more importantly what I can test to prove/disprove it. Do you mean I should put a test supply volts (e.g. 9V) into the left input (LIN @ CZ301) and measure the volts at the top of R327? Or is this testing to be done with R328 reconnected?

Many thanks to you both.
 
I’ll spend a bit more time studying the data, but numbers look ok at first blush. Is IC301 behaving now? I won’t trust the output stages if it’s not showing 5V. I’ve not found a bias current procedure, but would guess 20mV across each of the eight 0.2 Ohm emitter resistors. That would lead to toasty heat sinks. You might look at the working channel for guidance.

Any luck with mystery voltage in the input section?

You’re making good progress. :)
 
Just found your latest post.
Good volts at Q309/409.

I assume power output rails are about +/- 64V. Assuming about 0V at amp output, low voltage end of R334 should be about +5V and low voltage side of R344 should be about -5V. If you’re seeing 115V, that’s scary out of whack. Be careful! I’m surprised there isn’t escaping smoke. As a sanity check, you might make analogous checks on the working channel. I’m hoping that a meter probe is the wrong place.

re the behavior of the input section: note that feedback is still connected back to the front end via L-FDBK. When you have the output driven to +12V as in your previous tests, there should be large negative voltage at R327. Conversely, with -12V at output, a large positive voltage should appear at R327.

Past my bedtime I’m afraid. TTYL

Steve
 
You’re making good progress. :)
Thanks...I've had my head in it for a few weeks now and I keep feeling the fix is just up ahead!

So I'd better write this up before I get myself (and you) confused. I spent some time inspecting the back of the board for shorts/solder bridges etc with the continuity audio mode on the meter. I paid specific attention to the IC302 area, C312 area etc. I couldn't find anything beeping that shouldn't have been.
Make sure Q318 base-emitter isn't shorted.
I also checked this and couldn't find any shorts.

I didn't like the R342 and R343 measurements in circuit so I took them out and put in new ones (they measured fine out anyway). I also checked R339 out of circuit and it was good. Checked R341, R332, R338, R330, R339, R331, R357, R347 in circuit; seemed fine.

Found a 10uF capacitor available so I replaced C301 anyway as the board was apart.

Is IC301 behaving now?
Back to these regulator? things....At IC302 I removed my new TL431CLP regulator and replaced it with one that was taken previously from IC301. This is a bit risky I'm sure, but some previous measurements implied I had at one point a voltage drop of 4.5V across it. This is the problem with replacing too much stuff, too quickly, I guess! I now have 5V across C311 and 2V across C312. So it is better but still not right.

So onto the big hot resistors. R344 and R334 are definitely not at the right volts. The rails are ok at +/-64V.
Assuming about 0V at amp output, low voltage end of R334 should be about +5V and low voltage side of R344 should be about -5V
Does 0V output mean open at R328 to keep the idle around the mV range? That's how I did the L channel measurements:
R working channel R444, voltage across 59.5V giving -5V output, as you predicted!
L broken channel R344, voltage across 19.3V giving -44V output
R working channel R434, voltage across 59.5V giving +5V output
L broken channel R334, voltage across 98V giving -34V output

re the behavior of the input section: note that feedback is still connected back to the front end via L-FDBK. When you have the output driven to +12V as in your previous tests, there should be large negative voltage at R327. Conversely, with -12V at output, a large positive voltage should appear at R327.
Yeah I think I understand now; that the feedback circuit attempts to counteract the volt swings somewhat. I probably just oversimplified that, didn't I?! I tried it anyhow:
0V through output section gave -68V at junction of R327/R328
+12V through output section gave me -67.7V
-12V though output section gave me -63.3V, which is +4 volts so I can see it is trying to do what you said!

So I guess the main questions are:
From the earlier voltage level testing and reversing with speaker load, do you think the output transistors are ok?
Where can I look next to find out why I have only 2V across C312?
Definitely don't understand my huge/wrong voltages through the 3W resistors.
btw I'm amazed how you can work out what volts should be.....it's like you can see the future....or do Ohm's law maths in your head!

Thanks again
 
TL431 are adjustable shunt regulators--- think of them as adjustable Zener diodes. eg. https://www.ti.com/product/TL431#pps. The part is available from several manufacturers, perhaps with different lead orientations--- not sure about. Track down pin assignment for the exact part you're using and make sure it's installed according to schematic. You may have to scramble lead orientation, but don't worry about that. They're relatively rugged/forgiving. The original part may still be viable. There should be 2.5V across each of the 10k resistors, yielding 5V.

I don't know if the power transistors have survived, but the fact you can bias them to near 0V plus the experiments you report in post #14 are very encouraging. That you observed 115V leads me to suspect the meter somehow saw rail-to-rail supply voltage. Ground the black meter lead and measure each end of the resistor for insight.
 
Last edited: