Macintosh Quadra 840AV Post Issues

[jmacz]

I've been taking a look at @techstep 's Quadra 840AV which has been having some boot/post issues. I need some additional suggestions.

Setup: the motherboard, PRAM battery installed, memory installed, connected to speaker, connected to original power supply, connected to 13" Apple monitor, no drives attached.

Symptoms: push the power button, PSU fan spins up, voltage seems to be coming to the motherboard, but no chime, no video. @techstep mentioned it does post every so often (chimes/video/blinking disk icon). When I tested it, I could not get it to post at all after 30 attempts.

Visual Inspection: caps were toast, they had leaked and there was cap goo around the caps and nearby chips. The leakage actual seems to have eaten the printed board labels (cap IDs, resistor IDs, etc).. those printed labels were falling off the board like water based decals.

Recap: I took off all the electrolytic caps, replaced all of them with equivalent tantalum caps, cleaned up the board as much as I could including gently cleaning the legs of the surrounding chips that had signs of corrosion and green gunk.

Post Recap Test: the problem persists BUT with a change. Whereas before I was not able to get the board to post at all, now I can get it to post more often than not. So more often than not, I get a chime, I get video, and I get a blinking disk icon. I also fully booted the motherboard with an external SCSI disk no problem.

Current Symptoms:

• Machine is cold (has been powered off and disconnected for over an hour).
• Connect the power to the power supply and push the power button. Nothing.
• Wait a second, and push the power button again. Chimes, get video, get blinking disk about 50% of the time.
• If the second attempt didn't work, wait another few seconds, and push the power button again. Chimes, get video, get blinking disk about 75%+ of the time.
• Once I get the machine to post (ie. it chimed), it will ALWAYS chime if I restart or even power down the board but as long as I power it again within 10 minutes or so.

All the new tantalum caps were good (tested capacitance) before they went on the board. Each new cap is solidly on the board and has good connectivity. I did not touch the existing surface mounted (non-electrolytic, tantalum?) capacitors. I would assume they should be fine but I have not taken them off the board to test them.

Given it works after the second or third time, and works after it's been successful, it seems like a capacitor somewhere? I had thought it might be temperature but it's not, because even on a cold boot, the second or third attempt works, before anything heats up.

Another possibility might be the power supply. It's the original power supply. I checked the various pins (5V, 12V, -12V) and those all check out and have proper voltage. But maybe there's some timing thing going on after the PSU's sat for a bit?

I know these 840AVs are notoriously hard to deal with once caps have leaked. So hoping to see if anyone has some additional suggestions on things to try before I return this to @techstep.

 

[3lectr1cPPC]

Can you post a photo of the board?

 

[jmacz]

Here's a picture.

 

[jmacz]

And of course, after I post this, it's been posting like a champ.

 

[jmacz]

FWIW, I put it under a microscope and looked for breaks, didn't see any visible ones. Also checked all the worn or gunky pins for connectivity from the top of the leg to another component, and all had continuity.

 

[cheesestraws]

When it is powered on but not apparently doing anything, does anything happen when you press the reset switch?

 

[3lectr1cPPC]

Ah boy, brown board. The traces on those are extremely fragile compared to others and very small. Absolute nightmare to deal with. It looks like you’ve checked over everything with a fine tooth comb but I’d do so again.

 

[jmacz]

When it is powered on but not apparently doing anything, does anything happen when you press the reset switch?

Hmm, I have not tried that. Once I get it back into that state, I will try pressing the reset button.

 

[jmacz]

Ah boy, brown board. The traces on those are extremely fragile compared to others and very small. Absolute nightmare to deal with. It looks like you’ve checked over everything with a fine tooth comb but I’d do so again.

LOL, that's exactly what I told @techstep. If there's a trace issue, going to be a pain to fix it. Spent way too much time in the garage today and my eyes are tired so will take another pass later. But nothing obvious stuck out to me and once it's booted, it will continue to work fine even after multiple restarts so that doesn't seem like a break (?) and also again doesn't seem to be a thermal issue since it fixes itself before the temperature gets warm.

 

[3lectr1cPPC]

Some connection somewhere is weak, otherwise that shouldn’t be happening. If there are no electrolytic caps left then it shouldn’t be that I’d think.

 

[jmacz]

Some connection somewhere is weak, otherwise that shouldn’t be happening. If there are no electrolytic caps left then it shouldn’t be that I’d think.

Yup, no electrolytics left (except in the PSU).

 

[jmacz]

When it is powered on but not apparently doing anything, does anything happen when you press the reset switch?

Umm, hmm. I only have the motherboard. What does the reset switch connect to? And do I ground whichever pin it is to perform the reset? Does it use one of the J29 pins?

 

[jmacz]

Hmm, I have not been able to repro the issue this evening. Go figure.

The only change since I made the post was I removed the PRAM battery. I left it out and I've been powering up the machine without it. I can't imagine that's related but as of right now, it's posting every time. Will try again tomorrow morning.

 

[jmacz]

Tried again this morning and the machine booted/posted immediately on the first try. Again, the only difference is I have left the PRAM battery out. I just put it back in and will try again later and see what happens.

 

[jmacz]

I don't know what to make of it. It posted way more than not today (kept trying every time I passed by). With the PRAM battery out, it posts on the first try. With the PRAM battery in and if it has sat for a bit, the first attempt fails, but the second attempt always succeeds. The PRAM battery itself is good (I tried with another PRAM battery that's good and same result). The battery is holding doing its thing as the system clock stays correct even when the PSU is disconnected.

Out of ideas at this point. The unit is powering up most of the time as opposed to when I got it when I couldn't get it to chime or show video at all.

All I can think of is that the PRAM battery thing is a coincidence. And maybe something's up with the power supply at this point.

 

[3lectr1cPPC]

Probably a bad or weak trace or connection in the PRAM circuit somewhere. They hide really well on this particular board.

 

[dalek]

I'd suggest recap PSU - old caps may be weak / reforming when powered on for long periods of time. Also check/clean power connector - oxidised contacts in the PSU plug can lead to poor power distribution (measuring ok at PSU connector end but not at motherboard power connector end).

In particular, with my 660av there was weak supply on the + and -12V lines which held the board in reset state.

 

[jmacz]

I had measured the +12/-12/+5 lines on the motherboard side. But yeah, still a possibility.

 

[joshc]

The 840av PSU uses the same pinout as the IIci/Q800/IIcx etc I believe, so if you have any working ones of those that would be another way to test.

 

[jmacz]

This particular 840AV board has a 22-pin PSU connector (of which 21 pins are used, one is not connected). I'd have to rig something to test this with a 10-pin IIci PSU.