Pioneer SX-880 Protection Relay Dropping Out

[Brivan]

Good afternoon,

I have a Pioneer SX-880 with a protection relay issue I've been battling with for the last few days. Typically, the protection relay will briefly drop out after the receiver has been powered up for a minute or two. This may happen a few more times - about at 30-second to one-minute intervals, after which the relay would stay engaged.

Originally, I couldn't set the DC offset in the right channel to 0v - it would end up somewhere between -50 and -100mV at the minimum. I ended up removing the Q jumper to be able to obtain 0v. This was odd - I've never had to do this on any Pioneer X80 receiver, before. After this, I was able to set DC offset and idle current in both channels to factory spec (although I now had to remove the P jumper to obtain 0v on the left channel). But the sporadic relay action continued.

Monitoring DC offset on both channels, I've seen spikes up to 40 volts when the protection relay kicks out.

I checked voltages and have 13.5v from Q26 and about 41v at both main filter caps. After reading a few AK posts, I replaced Q20 (and Q19 at the same time). Relay cycling problem still present. I checked output transistors and they test fine. I tested R255, R256, R261 & R262 for high resistance / open, but they tested fine. From my notes, I tried replacing Q33, Q34 and Q36, with no good results (this was before I finally saw 40v spikes on both channels). I've monitored the 13.5v, 41.5v and -41.5v supplies and they don't budge during a protection relay drop-out.

I've pressed on the main board and wiggled components, but it doesn't make any changes. Whatever is happening appears to be happening as it warms up and stops when warmed up fully (I think).

At one point, I was convinced it had something to do with function change. Whenever I would switch to a new function after a one- or two-minute arm-up, the protection relay would drop out "sometimes" when I would switch from one function to another, but it would only do it once per function. This could be coincidence, but I was able to repeat this a few times, so I thought it was worth mentioning. Although, the relay would still drop out on its own without me touching anything. This is when I started looking into the power supplies.

The spiking DC offset in both channels should be a big clue as to what's happening, but I don't know where to go next. Any suggestions you could provide would be appreciated. Thanks!

 

[markthefixer]

Let's start at the beginning, the power supply regulators.

They ALL key off the D28 14v zener diode. The 13.5v supply is a DIRECT read of D28's 14v to the base of Q26 the 2sd712. It also powers the PA3004 protect chip.

The + 41.5v regulator is fed + 52~56 volts off of c407 470uf.
Jfet Q24 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q30 2sd325 pass transistor runs through R405 a 2.2 ohm resistor. Could be damaged...
Q28 2sc945 forms the feedback element in the regulator, with it's emitter at the zener voltage.
Q28 base senses the R410 15k, R409 8.2k voltage divider and wants to balance at 14.6 volts.

The - 41.5v regulator is fed - 52~56 volts off of c408 470uf.
Jfet Q25 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q27 2sb536 pass transistor runs through R406 a 2.2 ohm resistor. Could be damaged...
Q29 2sa904 forms the feedback element in the regulator, with it's emitter at GROUND.
Q29 base senses the divider R411 15k and R406 300k in parallel make 14285 ohms,
with R412 15k makes a voltage divider that wants to balance at - 0.6 volts.

Thus the negative voltage is HIGHLY dependent upon the positive voltage.
Glitch one, glitch them BOTH.

The negative supply can glitch on it's own, independently.

These two voltages BOTH directly influence the DC offset of BOTH of the power amplifiers.

Those 2.2 ohm 1/4 watt resistors take a beating. A quick test would be to short across them, they are just intended to be a "last chance" type fuse.

 

[Brivan]

Let's start at the beginning, the power supply regulators.

They ALL key off the D28 14v zener diode. The 13.5v supply is a DIRECT read of D28's 14v to the base of Q26 the 2sd712. It also powers the PA3004 protect chip.

The + 41.5v regulator is fed + 52~56 volts off of c407 470uf.
Jfet Q24 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q30 2sd325 pass transistor runs through R405 a 2.2 ohm resistor. Could be damaged...
Q28 2sc945 forms the feedback element in the regulator, with it's emitter at the zener voltage.
Q28 base senses the R410 15k, R409 8.2k voltage divider and wants to balance at 14.6 volts.

The - 41.5v regulator is fed - 52~56 volts off of c408 470uf.
Jfet Q25 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q27 2sb536 pass transistor runs through R406 a 2.2 ohm resistor. Could be damaged...
Q29 2sa904 forms the feedback element in the regulator, with it's emitter at GROUND.
Q29 base senses the divider R411 15k and R406 300k in parallel make 14285 ohms,
with R412 15k makes a voltage divider that wants to balance at - 0.6 volts.

Thus the negative voltage is HIGHLY dependent upon the positive voltage.
Glitch one, glitch them BOTH.

The negative supply can glitch on it's own, independently.

These two voltages BOTH directly influence the DC offset of BOTH of the power amplifiers.

Those 2.2 ohm 1/4 watt resistors take a beating. A quick test would be to short across them, they are just intended to be a "last chance" type fuse.

Thank you, Mark, for you in-depth explanation. I like to learn about this stuff and your explanations are second to none. Also curious as to why glitches like these cause 40v to show up in the audio outputs. I'll dive into this, again, tomorrow morning and will report my findings back to you. Thank you!

 

[markthefixer]

dc offset of the power amp is balanced between the two rails by Q18 & Q19's transistor base current. Upsetting that could force the amp to rail.

It really isn't hard to rail an amp.

 

[Brivan]

Let's start at the beginning, the power supply regulators.

They ALL key off the D28 14v zener diode. The 13.5v supply is a DIRECT read of D28's 14v to the base of Q26 the 2sd712. It also powers the PA3004 protect chip.

The + 41.5v regulator is fed + 52~56 volts off of c407 470uf.
Jfet Q24 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q30 2sd325 pass transistor runs through R405 a 2.2 ohm resistor. Could be damaged...
Q28 2sc945 forms the feedback element in the regulator, with it's emitter at the zener voltage.
Q28 base senses the R410 15k, R409 8.2k voltage divider and wants to balance at 14.6 volts.

The - 41.5v regulator is fed - 52~56 volts off of c408 470uf.
Jfet Q25 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q27 2sb536 pass transistor runs through R406 a 2.2 ohm resistor. Could be damaged...
Q29 2sa904 forms the feedback element in the regulator, with it's emitter at GROUND.
Q29 base senses the divider R411 15k and R406 300k in parallel make 14285 ohms,
with R412 15k makes a voltage divider that wants to balance at - 0.6 volts.

Thus the negative voltage is HIGHLY dependent upon the positive voltage.
Glitch one, glitch them BOTH.

The negative supply can glitch on it's own, independently.

These two voltages BOTH directly influence the DC offset of BOTH of the power amplifiers.

Those 2.2 ohm 1/4 watt resistors take a beating. A quick test would be to short across them, they are just intended to be a "last chance" type fuse.

Good morning, Mark

I tested R405 & R405 in-circuit and both measured 2.3 ohms. I pulled both and re-measured them at the same values. Since I had them out, I went ahead and replaced both with some standard 2.2 ohm resistors I had in stock. I assume the originals are a fuse-type flameproof resistor and should probably be replaced with the same type? I can likely reinstall the originals once this thing (hopefully) comes to a conclusion. I still have the protection relay single-cycling after about a 1-minute warm-up, then another single cycle about 30 seconds later. I powered off after this. While the receiver was on, I set my DMM to a fast sampling rate and monitored the +41.5v supply. It had little to no deviation when the protection relay dropped out. If anything, I observed a 0.1v voltage gain during the drop-outs. I'll keep looking for more clues, but any further advice you have to offer is welcome! Thanks!

 

[Brivan]

Let's start at the beginning, the power supply regulators.

They ALL key off the D28 14v zener diode. The 13.5v supply is a DIRECT read of D28's 14v to the base of Q26 the 2sd712. It also powers the PA3004 protect chip.

The + 41.5v regulator is fed + 52~56 volts off of c407 470uf.
Jfet Q24 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q30 2sd325 pass transistor runs through R405 a 2.2 ohm resistor. Could be damaged...
Q28 2sc945 forms the feedback element in the regulator, with it's emitter at the zener voltage.
Q28 base senses the R410 15k, R409 8.2k voltage divider and wants to balance at 14.6 volts.

The - 41.5v regulator is fed - 52~56 volts off of c408 470uf.
Jfet Q25 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q27 2sb536 pass transistor runs through R406 a 2.2 ohm resistor. Could be damaged...
Q29 2sa904 forms the feedback element in the regulator, with it's emitter at GROUND.
Q29 base senses the divider R411 15k and R406 300k in parallel make 14285 ohms,
with R412 15k makes a voltage divider that wants to balance at - 0.6 volts.

Thus the negative voltage is HIGHLY dependent upon the positive voltage.
Glitch one, glitch them BOTH.

The negative supply can glitch on it's own, independently.

These two voltages BOTH directly influence the DC offset of BOTH of the power amplifiers.

Those 2.2 ohm 1/4 watt resistors take a beating. A quick test would be to short across them, they are just intended to be a "last chance" type fuse.

Another update: I pulled R419, R422 & R423, but was still getting protection relay drop-outs (while reading a constant 2mV on the collector of Q36), so I think this means that the problem is in the protection circuit and not in the amplifier section since (I think) I've de-coupled the protection circuit sensing by pulling those three resistors. As stated earlier, I already replaced Q33, 34 & 36. Next I kept my meter probe on pin 4 of the PA3004 IC and measured about -10mV. The protection relay dropped out (and back in, as it always does), but the -10mV reading didn't fluctuate much - maybe a millivolt or two. I should have monitored pin 3 of the PA3004 (base of Q37), as well, but I skipped ahead and replaced Q37, but still getting the relay drop-out. With the probe on the base of Q37, I'm reading about 650mV with the protection relay engaged. When it drops out, I see a reading of near 0v. Now, it's looking like the PA3004 might be bad.

Or not. If you may recall, I'm the guy with crappy power in his shop, who has to rely on a DBT to "filter" out probable harmonics from mains power in order to keep any Pioneer receiver with a PA3004 above an SX-780 from having protection relay cycling issues. Usually, if I'm going to have relay issues related to this, the relay will cycle on and of, on and off - continuously. But this one's different. It's more random and less cyclic. There still may be a problem with the PA3004 and I may replace it with a new spare, but I'm going to power it up on a single 60-watt bulb (I usually use 2 or three bulbs in series on my DBT - depending on what's connected to it) and see what happens. A single bulb (or two) usually does a better job of filtering the power than all three.

I'll report back with any revelations.

 

[Brivan]

Let's start at the beginning, the power supply regulators.

They ALL key off the D28 14v zener diode. The 13.5v supply is a DIRECT read of D28's 14v to the base of Q26 the 2sd712. It also powers the PA3004 protect chip.

The + 41.5v regulator is fed + 52~56 volts off of c407 470uf.
Jfet Q24 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q30 2sd325 pass transistor runs through R405 a 2.2 ohm resistor. Could be damaged...
Q28 2sc945 forms the feedback element in the regulator, with it's emitter at the zener voltage.
Q28 base senses the R410 15k, R409 8.2k voltage divider and wants to balance at 14.6 volts.

The - 41.5v regulator is fed - 52~56 volts off of c408 470uf.
Jfet Q25 provides a constant current to run the regulator while servoing out the hum.
The DC to the Q27 2sb536 pass transistor runs through R406 a 2.2 ohm resistor. Could be damaged...
Q29 2sa904 forms the feedback element in the regulator, with it's emitter at GROUND.
Q29 base senses the divider R411 15k and R406 300k in parallel make 14285 ohms,
with R412 15k makes a voltage divider that wants to balance at - 0.6 volts.

Thus the negative voltage is HIGHLY dependent upon the positive voltage.
Glitch one, glitch them BOTH.

The negative supply can glitch on it's own, independently.

These two voltages BOTH directly influence the DC offset of BOTH of the power amplifiers.

Those 2.2 ohm 1/4 watt resistors take a beating. A quick test would be to short across them, they are just intended to be a "last chance" type fuse.

I powered up the receiver using only 1 60-watt bulb in my DBT. This changed the relay drop-out behavior. It took from five to ten minutes for a drop-out to occur. At least I'm seeing a result when making a change. I'm betting when I take this receiver off-site, the protection relay will stay engaged.

One thing I don't understand through all of this is, given my problem is stemming from the PA3004 (whether induced by mains harmonics or not) is why I'm seeing 40v spikes on pins 5 & 12 when the relay disengages? I haven't yet monitored the speaker terminals to see if the protection relay is disengaging fast-enough to keep the voltage spikes from being output to any speakers. Or, if 40v is making it to the speaker terminals, is it easily snubbed by the low resistance of the speakers?

 

[Brivan]

dc offset of the power amp is balanced between the two rails by Q18 & Q19's transistor base current. Upsetting that could force the amp to rail.

It really isn't hard to rail an amp.

After a bit of a delay, I'm back at working on this problem. My theory of "bad power" at my workshop was debunked when I took the receiver home and tried it, there, with the same results - brief protection relay disengagement after about a 1-minute warm-up, then again about 30 seconds later. After that, it usually works fine, but may have one more relay disengagement before settling down. Today, I pulled R237 and R238 in an effort to separate the preamp section from the amp section. The relay doesn't pull in at all, now and I have some odd voltage n=measurements at pins 65 & 68. Pin 65 (right channel) = 36v and pin 68 (left channel) = 6v. I wasn't expecting this at all and it doesn't make sense to me. I'm at a bit of a loss as to where I should go from here. Any suggestions are welcome!

 

[Brivan]

dc offset of the power amp is balanced between the two rails by Q18 & Q19's transistor base current. Upsetting that could force the amp to rail.

It really isn't hard to rail an amp.

Some other observations and measurements taken since my last message:
I patiently took measurements at each pin of the PA3004 (which has been replaced) while the protection relay drop-outs were occurring. here's what I found:
Pin 1) 13.3v - doesn't vary
Pin 3) 650mV; decreases to 0.1mV when the relay drops out
Pin 4) -9mV; decreases to -4mV when the relay drops out
Pin 5) 0.54mV; increases to 0.75mV when the relay drops out
Pin 6) -7.4v; decreases to 0.75v when the relay drops out
Pin 7) 6.7vac; decreases to 40mvac when the relay drops out
Pin 8) 7.1v; decreases to 1.6v when the relay drops out

All the above voltages recover within a few seconds and the relay re-engages. This usually happens within a minute or two of power-up. It may happen a second time about 30 seconds later. After that, the relay remains energized.
C417, 418 & 419 have all been replaced.

 

[Brivan]

dc offset of the power amp is balanced between the two rails by Q18 & Q19's transistor base current. Upsetting that could force the amp to rail.

It really isn't hard to rail an amp.

Another update: Today I started tracing the -41.5 supply starting at the emitter of Q27, where I saw the voltage drop significantly when the protection relay dropped out. I tested both the collector and base of Q27 after that and didn't notice any fluctuation when the protection relay disengaged. Looks like Q27 is suspect. The solder connections of Q27 didn't look terrible, but didn't look great, either. I resoldered all three and applied power. Protection relay is now staying engaged. Hmm. You'd think it would have dropped out on one of the many occasions I pushed on the PC board looking for just such a problem. Guess not. I left power applied for an hour and it's still working fine. I'll test it further and, if you don't hear from me, consider it fixed!

 

[petehall347]

Pin 7) 6.7vac; decreases to 40mvac when the relay drops out
not checked schematic yet but losing AC here the relay will drop out .
edit .. clean up fuse 2 and the problem should be fixed . clean all fuses whilst in there .

 

[Brivan]

Pin 7) 6.7vac; decreases to 40mvac when the relay drops out
not checked schematic yet but losing AC here the relay will drop out .
edit .. clean up fuse 2 and the problem should be fixed . clean all fuses whilst in there .

Interestingly-enough, I did pull the fuses and check them for continuity while being tapped on. I also wiped them off and checked the fuseholder clips. All good.

 

[Brivan]

Another update: Today I started tracing the -41.5 supply starting at the emitter of Q27, where I saw the voltage drop significantly when the protection relay dropped out. I tested both the collector and base of Q27 after that and didn't notice any fluctuation when the protection relay disengaged. Looks like Q27 is suspect. The solder connections of Q27 didn't look terrible, but didn't look great, either. I resoldered all three and applied power. Protection relay is now staying engaged. Hmm. You'd think it would have dropped out on one of the many occasions I pushed on the PC board looking for just such a problem. Guess not. I left power applied for an hour and it's still working fine. I'll test it further and, if you don't hear from me, consider it fixed!

Never mind. I let it sit for an hour or so and powered it back up with speakers connected for an extended trial session and the relay dropped out within about 15 seconds of power-up. Back to the drawing board. I'm going to follow that -41.5v supply drop-out issue - it's either Q27 or beyond that. I need to do more testing around Q27 to make sure. The problem is that the relay drop-outs are starting to be inconsistent, making me think I'm onto something when, in reality, I'm not.

 

[suzook]

Stupid question...Have you replaced the protection relay? These are known to go bad, or act strange after time.

 

[Brivan]

Stupid question...Have you replaced the protection relay? These are known to go bad, or act strange after time.

I haven't replaced the protection relay. Voltage checks at the PA3004 indicate the relay is being "told" to open up. I was trying to figure out why. The worst part about the problem is that it was intermittent and wouldn't always rear it's ugly head. I'll follow up in the next post.

 

[Brivan]

Yesterday, I did some more voltage checks while the relay disengaged. I was concentrating on Q27, since that's where I left off. Oddly-enough, when I held my meter probe on the base of Q27, the problem would never happen. I knew the emitter voltage would drop out when the relay de-energized, and that was way wrong. The collector voltage remained nearly constant during drop-outs, so it was down to testing the base. And that's where I couldn't test any further due to the relay remaining properly engaged as long as my probe was on the base of Q27. After that, I replaced both Q27 and Q29, hoping one of them was causing the problem. So far, so good. I've powered the receiver up several times and have no relay drop-outs to report. I'll do more testing next week and hope that I got it. Q27 and Q29 both tested fine, but I'm hoping one of them (probably Q29) was breaking down with power applied to it.

 

[suzook]

That's great news! Hopefully you can finally enjoy it!

 

[larryderouin]

Just as a PM measure, check D35. Or just replace it with a 1n4004 or better. (Flyback Diode).

 

[Brivan]

Just as a PM measure, check D35. Or just replace it with a 1n4004 or better. (Flyback Diode).

Okay, will do. I'll just go ahead and replace it. I'll report my test results back here in a day or two. Thanks, Larry!

 

[Brivan]

Reporting back on this - the receiver is still working fine. The replacement of Q27 & Q29 eliminated the sporadic dropping of the protection relay. Thanks to all those who provided advice!