NAD 7140 Intermittent cut-out - power supply issue?

[DJSteve]

Bought an NAD 7140 off ebay a few months ago... and apparently didn't read the listing closely enough, as there was a note in the listing it would occasionally fail to turn on. Wasn't a real big deal during summer/fall, as if it failed to kick on initially you could just leave it on and it would start working within a few minutes. As we've gotten further into winter (house is a few degrees cooler, generally, so maybe temperature related?) it's been happening far more often, and on at least a couple occasions took a lot longer to warm up (like 30+ minutes) before starting to work. Sometimes it will kick on and off a bit as it transitions from not working to working... once working it seems fully functional but it's mainly getting used in the living room to play music for the kids so not pushing it very hard or anything.

I have a moderate amount of experience with general electronic stuff, however have not done a lot with discrete BJT/FETs and what I have done was a couple decades ago... so I understand the general theory of operation but am pretty rusty on specifics and am not real knowledgeable on testing/troubleshooting. Schematics in the service manual I found are taking some effort to figure out what's what and how everything actually works (for example I'm still baffled what actually drives output relay as I can't find that anywhere).

Opening up the case, electrolytic caps all look OK visually (no bulging/leaking). There is some heat discoloration of the pcb (not scorched/black but definitely darker discolored area on the component side) that seems to emanate from Q903, and to a lesser extent also some discoloration around Q901 and Q902. All these relate to the +/- 43V supplies, which seems to run the power amplifier section. Screenshot of relevant section of the schematic with a red circle on Q903 and measured voltages included below. +43 rail measured a little bit low at 40.8 but not sure if that matters(?). -43 was much closer at -42.95, however checking in AC mode (using Fluke 87V) showed 45mV AC--which seemed like something not working properly, especially given that Q903 is in that part of the circuit.

When it cuts in/out I *think* I can hear relay click, but am having trouble reliably reproducing the failure mode with the receiver opened up. Some stuff I've read seemed to indicate that the relay in these was just related to power on/off rather than having any protection function? I haven't been able to catch any of the power supplies doing anything weird, HOWEVER there have been a couple times where the receiver was cut out and when I tried to measure -43 supply traces it would cut in and start working... so possibly putting the DMM in the circuit is somehow changing things enough to stabilize whatever is going wrong? At this point I'm also not ruling out something weird like a flaky solder connection or a trace problem, as I've seen a few reports of those types of issues on these.

Any advice or insight from someone more knowledgeable / that's been into one of these before on how to proceed diagnosing / what to try to replace? Pull Q903 out and see how it measures? Also partly looking for speculation of what's cause and what's effect--is failure of Q903 likely to be source of the problem, or is that running hot more likely a result of something wonky with Q902 or elsewhere? Would power supply flakiness lead to complete cut-out of the output relay or is that more likely a separate issue?

I'm probably going to keep trying to reproduce the issue with the case open for a bit, so will add replies if/when I find anything further. I do have an oscilloscope I can get out, also, and could potentially set that up on the -43V supply to see if I can tell more about what's going on at power-up or with the AC/noise if warranted.

 

[OldJBL]

I own two NAD Monitor Series receivers, a 7400 and 7100. Both had cracked solder runs that needed reflowed which is a fairly common issue with NAD components. Your problem might be the same since the problem gets worse as the ambient air temperature drops. When it is warmer the solder expands and can bridge the gap and when it is cooler the solder contracts and the gap gets larger. When turned on, the receiver will start to heat up internally. What might be happening is it takes longer to heat up and close the solder gap when the ambient air temperature is cooler. Or it could be something else causing the issue.

 

[DJSteve]

I own two NAD Monitor Series receivers, a 7400 and 7100. Both had cracked solder runs that needed reflowed which is a fairly common issue with NAD components. Your problem might be the same since the problem gets worse as the ambient air temperature drops. When it is warmer the solder expands and can bridge the gap and when it is cooler the solder contracts and the gap gets larger. When turned on, the receiver will start to heat up internally. What might be happening is it takes longer to heat up and close the solder gap when the ambient air temperature is cooler. Or it could be something else causing the issue.

That definitely seems possible. It seems like there is quite a bit of flux residue on the bottom side of that board in general, and some connections that look a bit suspicious. Pin 5 on IC901 that runs +43v supply (and by proxy -43v) for one example. Were you able to isolate specific spots that were causing problems, or just reflow anything that looks questionable and see if it fixes it? It looks like someone has messed with / tried to reflow a couple of the pads for the output relay but nothing else looks obviously reworked.

I did get it into failure mode while measuring correct voltages on +43 and -43 supplies. It kicked on maybe 10-15 seconds later.

Running it for a while and then checking for heat in that power supply area, Q901 seems hottest but not terrible. Q903 does not feel warm so discoloration being centered in that area may just be heat conduction through the traces from the larger Q901/Q902 components? Maybe I don't actually have a power supply issue at all... just seemed weird seeing a lot more AC on -43 than any of the others.

 

[OldJBL]

That definitely seems possible. It seems like there is quite a bit of flux residue on the bottom side of that board in general, and some connections that look a bit suspicious. Pin 5 on IC901 that runs +43v supply (and by proxy -43v) for one example. Were you able to isolate specific spots that were causing problems, or just reflow anything that looks questionable and see if it fixes it? It looks like someone has messed with / tried to reflow a couple of the pads for the output relay but nothing else looks obviously reworked.

I did get it into failure mode while measuring correct voltages on +43 and -43 supplies. It kicked on maybe 10-15 seconds later.

Running it for a while and then checking for heat in that power supply area, Q901 seems hottest but not terrible. Q903 does not feel warm so discoloration being centered in that area may just be heat conduction through the traces from the larger Q901/Q902 components? Maybe I don't actually have a power supply issue at all... just seemed weird seeing a lot more AC on -43 than any of the others.

I am not skilled, or knowledgeable, enough to do repairs on anything that requires any decently complex soldering. The tech I had working on them told me about the cracked solder joints. This isn't just an issue with NAD components but a lot of other brands too. My 7400 had other issues besides cracked runs but the sole problem with the 7100 was cracked solder runs to the speaker terminals. There were some other cracked runs in the 7400 but I never enquired about where they were located.

 

[DJSteve]

Update after doing some further investigation / testing in between other tasks today... I followed traces from the relay to figure out related components and finally found the driving circuitry in the schematic. Definitely looks to me like it has some sort of comprehensive speaker protection circuit... IC801 is TA7317p that appears to be an IC for that specific purpose, which seems to have inputs from L and R output and some power supply stuff, as well as thermal switch on the big heat sink. Was able to catch it in fault condition and found about -3v on output coming from R power amp section into relay (L chan real close to 0). Once output relay kicks on that R output DC voltage is lower but still not zero.

I hadn't messed with amp alignment stuff previously, however with the revised understanding that the output relay has some protection function beyond just avoiding turn on noise I decided to go through those procedures--with the new theory being that something is off and protection circuit is doing its job. That was possibly informative, but not particularly fruitful. I see offset on R channel speaker ouputs and can adjust that somewhat, however best case seems to be around 400mV (spec is 0 +/-50mV). It seems real sensitive about how that is adjusted in terms of tripping protection/output relay. I can set L channel offset within 5mV and that seems to act like it's supposed to.

Neither bias adjustment yields the desired behavior... spec there is ~25mV. I never see anywhere close to that on either channel with volume down/no load, and adjusting the relevant trim pots doesn't seem to really change anything. Right channel is around 11mV with speaker connected / 0.2mV open. Left is around 0.9mV speaker connected and 0.25mV open.

Any hints what to check/try first in terms of further troubleshooting? It seems like I'm finding a few things that seem not quite right, but despite that it seems functionally fine once warmed up.

 

[petes'picks]

Most caps I have run across that tested low in value, high in df and high in esr have looked perfectly fine so I would suggest at least getting one of them cheap Chinese cap testers that measure df or esr. NAD is known to have used lesser quality caps durring that time frame but are worth fixing.

 

[Namlook]

On older NAD units I always check the resistance of the power switch when you engage it (while unplugged). Then I would start looking at capacitors in the protection circuit. NAD was notorious for (1) running garbage quality capacitors in their units and (2) having poor thermal management in all those BEE designed units.

 

[DJSteve]

Most caps I have run across that tested low in value, high in df and high in esr have looked perfectly fine so I would suggest at least getting one of them cheap Chinese cap testers that measure df or esr. NAD is known to have used lesser quality caps durring that time frame but are worth fixing.

Do you have specific guidance on cheap cap testers that give reliable df/ESR? I went down the rabbit hole researching a bit, and it seems like a lot of that stuff is based on open source firmware and Atmel processor, but a lot of them (namely FNIRSI LCR-P1 that seems to sorta be the flavor of the week) use different cheaper/more readily available processors that either don't give reliable ESR or just do a table lookup based on capacitance.

On older NAD units I always check the resistance of the power switch when you engage it (while unplugged). Then I would start looking at capacitors in the protection circuit. NAD was notorious for (1) running garbage quality capacitors in their units and (2) having poor thermal management in all those BEE designed units.

Power switch resistance looks good--0.1-0.2 ohms

I assume even with a capacitor tester I'm going to have to pull caps off to reliably test? Assuming I want to fix this up and keep it for a while, is my best course of action just to go ahead and do a full recap (all electrolytics?) and see where that gets me?