Sony PVM-8044Q Shutting Off Problem Update: Fixed

[dgame]

I have a Sony PVM-8044Q; it is a portable 8” CRT monitor that supports 15 kHz arcade resolutions.

When it works the picture is beautiful and everything is fine.

However, after awhile the screen goes a little crazy on the horizontal sync and then turns off.

The green power button is still pressed in but the power light is off. If release the power button and wait awhile it will turn on again when I press it back in.

Eventually it will turn off again in the exact same way. Jittery horizontal collapse from the edges to about two inches in while it turns off.

What could it be? What should I check?

[dgame]

Just before it [automatically/unintentionally] powers down there is a loud hissing noise coming from the power supply area.  It sounds like a tire losing air, not buzzing or humming.  I took out the power supply and physically, I don't see any obviously bad [swollen] capacitors on the board. There is some discoloration [brown] on the other side of the PCB under the positive leg of a big capacitor, both legs of a smaller capacitor, and under all of the legs of the transformer.  Could one of the good [looking] capacitors make that sound, or could it be the transformer?

[dgame]

I ran the monitor with the exposed power supply board hooked up.

Turns out the hissing was coming from a transformer on the power supply board.
With the board exposed it sounds like a buzzing bee.
When I press down on the transformer's windings the pitch of the buzz changes.
A replacement transformer [1-450-760-12] costs about $60+, which is more than I paid for the monitor.

However, the monitor can also be powered directly from 12VDC which is a separate from the AC/DC power supply.
An eBay seller has a 12V 4 pin XLR power supplies for around $80.
I decided to convert an external AC adapter to power it from the 12VDC XLR connector instead.
I ordered a 4 pin XLR connector and 12V 5A 60W laptop style power adapter for less than $20.00.
Solder two wires and save $60. Waiting for the parts to get here to try it.

[MonMotha]

If you care to try to repair that board, a few suggestions:

Touch up the solder joints.  That may be the only problem.  Those transformers vibrate a little during operation at fairly high frequencies (several kHz) due to the way the circuit works, and a cold solder joint can work its way loose over time.  Audible noise is not an uncommon result.

Audible noise during normal operation also happens, and the frequency and harmonics will change if you press down on the transformer since you'll change the electrical characteristics (stray capacitance, mostly).  This does not necessarily indicate failure (though it may hint at faulty or incomplete design).

If got a shorted winding somewhere in there or the core has come loose, then that can be fixed.  If the core is loose, you're in luck, since getting the core apart is often the hardest part of re-winding them.  Just get some similar magnet wire, and re-wind the transformer exactly how it was wound before (count turns and note direction as you take the old wire off) then re-epoxy the core together.  If the core hasn't come loose, you can often make it loose (to work on the transformer) by heating up the whole assembly (obviously apply heat to the ferrite core itself) via a hotplate or electric skillet until the epoxy hits the glass transition temperature, at which point it will let go.

Why yes, I do design switch mode power supplies

[lilshawn]

Why yes, I do design switch mode power supplies

you'd better not say that, cause i could have some choice words.... 

i'll just shut up now. 

[MonMotha]

Why yes, I do design switch mode power supplies

you'd better not say that, cause i could have some choice words.... 

i'll just shut up now. 

If it helps any, I overspec the hell out of my capacitors, and I actually look at the secondary characteristics (beyond just nominal rated capacitance and working voltage).

The linear supplies are simpler, and they have their place, but switchers can do some pretty amazing things.  I actually built a supply that did 24VDC out at about 12W (0.5A) from 22VDC-180VDC or 22VAC-125VAC input.  Anything in that range.  It didn't care.  It had its quirks, but achieving that range is pretty cool.

Of course, linears are not without their problems.  I've certainly had to debug my fair share of oscillating LDOs.

[lilshawn]

Why yes, I do design switch mode power supplies

you'd better not say that, cause i could have some choice words.... 

i'll just shut up now. 

If it helps any, I overspec the hell out of my capacitors, and I actually look at the secondary characteristics (beyond just nominal rated capacitance and working voltage).

The linear supplies are simpler, and they have their place, but switchers can do some pretty amazing things.  I actually built a supply that did 24VDC out at about 12W (0.5A) from 22VDC-180VDC or 22VAC-125VAC input.  Anything in that range.  It didn't care.  It had its quirks, but achieving that range is pretty cool.

Of course, linears are not without their problems.  I've certainly had to debug my fair share of oscillating LDOs.

FINALLY somebody gets it... now if we could only get the monitor engineers and designers and everyone else to follow suit.......

[dgame]

If you care to try to repair that board, a few suggestions:

Touch up the solder joints.  That may be the only problem.
...

I think it’s working now!
I changed out all of the electrolytic capacitors and touched up most of the solder connections.  
On my first attempts I only changed the capacitors that had some discoloration on the solder side of the board.

I think it was replacing the last cap (680µF 50V next to transformer) that completed the fix. I was reluctant to change it to the generic looking one I got to replace it. This cap had some epoxy sticking it to the board and the solder side didn’t have any discoloration or anything. In retrospect, once removed this capacitor was the only one with any physical symptoms of failure. Looking at all of them only the 680µF 50V had subtle signs of leakage under the bottom of the capacitor.

Thanks for the advice. It motivated me to try again and now the little monitor works!