IBM PS/2 Model 25 - CRT has no high voltage

[Jimmy422]

Hi all,

I have an IBM PS/2 Model 25 all-in-one computer that has a non-functional built-in CRT. There is no high voltage or heater voltage that I can see on the CRT.

Here are some things I've tried so far to fix the issue:

• Verified all voltages coming out of the SMPS are correct - 5v, 12v, 115v outputs are all perfect. The PC part of the system runs off the same power supply and I can confirm it works perfectly, so the lower voltages are definitely OK.
• Replaced all capacitors across the SMPS, flyback board, and CRT neck board since many were leaking
• Tested the horizontal output transistor - tested OK but replaced to be safe
• Tested the horizontal oscillator IC - someone else told me the outputs looked incorrect and it could have been bad, so I replaced it, but it did not solve the issue

https://imgur.com/a/svZKIND here are some measurements of the oscillator, associated transformer, and HOT I took with my scope.

https://imgur.com/a/hHLXU5q here are some photos of the front/back of the 3 boards.

There are no schematics that I have been able to find so I've been flying blind here. Any help would be much appreciated - I want to get my dining room table freed up again

Thanks!

[bob91343]

The high voltage is generated by an oscillator running at a frequency somewhere between 15kHz and 150 kHz.  The resultant wave is stepped up by a ferrite core transformer and then rectified by a diode.

Locate the transformer and the oscillator circuit.  Apparently yours isn't oscillating, perhaps due to a failed component or connection.  Examine each part carefully and each solder joint as well.  There may be a  cracked PC board causing an open circuit somewhere.

There should be reasonable voltages on each component.  Measure the resistors and capacitors and  semiconductors and transformer windings.

[Jimmy422]

The oscillator should be working - take a look at the waveforms coming out of it / going into it here: https://imgur.com/a/r9xr8AI

I have verified the solder connections all look good and I don’t see any cracks on the board.

All voltages going into the oscillator IC are exactly in spec according to its data sheet.

The transformer and its associated transistor also tested OK.

All resistors in the area also tested OK.

[bob91343]

If the oscillator is working, the transformer should raise its voltage, the rectifier should change it to dc, and that dc should be applied to the CRT.

[floobydust]

With the monitor powered off, I would make sure the 115VDC rail is discharged and measure ohms from the HOT collector (tab) to that rail. To see if there is an open circuit in the flyback primary to power path.
The HOT emitter (continuity) should be near DC common.  I don't see much drive to the HOT at the base and it's not even close to a square wave. There should be a small drive transformer going to the HOT base. I can't see a MC1391 in your pictures, where is it? Its datasheet gives a good schematic for a typical horizontal section.
It's just a matter of finding the open circuit, you can measure parts esp. power resistors nearby for high value.
There is a safety x-ray protect SCR next to Q6 and if it trips/shorts the HV gets shut down.
You should hear the V-defl buzzing at the yoke as well?
The board markings API-6310, 8509218, 03570-011 not sure which is the monitor part number. Would be worth some more pictures or more hunting for the service manual.

edit: the pics are just of the power supply where's the flybackkkk

[Jimmy422]

The collector of the HOT to the 115v rail is showing as having continuity according to my meter with a reading of 1.2 ohms.

Here’s some more photos, let me know if you want to see anything in particular: https://imgur.com/a/gGxZ2gy

The flyback board is marked as PWB-1298, but I haven’t been able to find any schematics for it. The computer itself has the model of 8525-004, since the monitor is built in it’s hard to find anything for the monitor alone.

I don’t hear any noise at all from the CRT when I turn it on. Not a peep.

[bob91343]

That might be the flyback transformer shown in part on the right.

[decoder_82]

When you see no heather voltage, where did you take the measurement ? at the CRT neck ? or directly from the FBT ?

[inse]

The drive signal for the HOT does not look valid, please make your measurement with more than just a few pixels resolution.
Check the drive circuit, Q401 seems to drive T401, which is the signal transformer for the base voltage.

[Jimmy422]

I checked the heater voltage at the neck - should I be checking it at the flyback?

I can re-measure the signal at the HOT but it seemed to be complete gibberish when I measured it last time, the scope reading was jumping all over the place and it wasn’t a consistent signal.

I pulled Q401 and it tested OK with a basic multimeter test as described here: https://vetco.net/blog/test-a-transistor-with-a-multimeter/2017-05-04-12-25-37-07 - is there another test I can do on it?

How can I test T401? This is my first time testing anything beyond a basic resistor/diode so forgive me on being a little rusty!

[inse]

Please verify that Q401 drives the drive transformer.
Check the horizontal control signal path from the IC to the driver across the transformer at the base of Q403.
Check continuity on both windings of T401 and verify the magnet core is intact, it is glued together and might have fallen apart.
One of the legs of T401 has to be connected to a supply (R416?)
Q403 needs a good amount of base current to switch properly.
And when its switching, you have a high amplitude on its collector, several hundred Volts!
The presence of heater voltage should be observable, in a dimly lit room you can see the filament glowing.

[decoder_82]

In the absence of glow of the heather filament, I would check heather voltage on the FBT itself. I had the case on a monitor where there was a failure between the FBT and the CRT neck board.

[Jimmy422]

When I check continuity of T401, the two points next to R421 have continuity, but the two points next to Q401 and C409 do not. This happens even after I pulled the transformer out of the circuit.

Does that mean this transformer is shot?

[inse]

Then its unfortunately not a transformer any more...
Depending on your commitment, you could try rewinding it unless you are able to source a spare part.
Examine whether the winding possibly broke close to the terminals and can be reattached

[floobydust]

Sometimes it's possible to repair a transformer, check with a magnifier or microscope where the windings are terminated at the bobbin pins.
Corrosion can happen there or varnish made a poor soldering joint. It's a common place for problems. A pic of the underbelly of the transformer would help too, before it's written off.
I have also seen the power resistor go open-circuit, the one from the drive transformer to power.

[Jimmy422]

A quick look and it looks like they are connected at the pins - I’m getting continuity between the wire and the pins, just not between the two pins.

Here are the best pictures I could get of the bottom https://imgur.com/a/RjLTNgL

[inse]

What a pity...
Do you consider rewinding it?
Hopefully the broken winding is on the top.
Peel off the tape and see if the core comes apart.
If it is glued, one more hurdle.

[floobydust]

OP you did a great job replacing the capacitors and I would not get discouraged. Many times vintage gear has weird failures with the unique parts. Patience and tenacity is needed.
The transformer, it's only copper wire but it either gets bashed during manufacturing leaving a thin spot, or it corrodes and breaks or both.

I'm assuming the break is on the primary side, with the very thin wires? If you are very gentle you can touch the solder-tinned parts of the leads with your multimeter probes to see if you get continuity there, which would rule out a bad connection at the pins. Any sharp angle or bend is also a place for breaks.

It's entirely possible to disassemble the transformer, unwind it counting # turns and rebuild it. I use my oven or hot air gun to heat up the transformer to get the (glued) core apart. DiodeGoneWild does it all the time in many of his videos.

Or, find another drive transformer. What's the part number written on it? I don't recognize the monitor's logo but the same part would be used in many different models.
Could look at an IBM 5153 and use that one?

[fzabkar]

I have repaired about 6 or so PS/2 Model 25s. In 5 of those cases the fault was the SMOS Systems graphics controller IC. One or both of its H and V sync outputs failed. The pulses were still there, but they no longer transitioned between 0V and 5V. I "repaired" these boards by interposing a comparator to buffer the faulty signal.

The other motherboard had a faulty crystal that oscillated at a higher frequency. This frequency was not an integral multiple of the design frequency, so it suggested to me that the quartz element had fractured and become lighter. One other possibility which I considered was that there had been a "dag" when the crystal was cut, and this had fallen off in normal usage.

Anyway, this SMOS IC was unobtainable at that time (1990/1991), either from IBM or SMOS.

Just something to watch out for if you intend to keep this computer as a museum piece ...

[Jimmy422]

You know that’s kind of interesting - I made an adapter to go from the onboard video port to a VGA connection (I plan to just retrofit an LCD to fit in the casing and hook it up through VGA), and the monitor said the frequency was out of range. This should have worked because the video output connector is the same as on the PS/2 model 30, which just outputs to a standard VGA connector. I wonder if that chip is shot…

Regardless I’m going to throw in a VGA card which I know works, so there’s still a solution.

Out of all of my vintage computers I’ve repaired, this has by far been the worst…

[inse]

That means, you have written off the CRT monitor?