VOGONS


"The Beast" - IBM PS/2 P70 8573-121

Topic actions

Reply 20 of 40, by Brickpad

User metadata
Rank Member
Rank
Member

[Quick Update}

I have completely isolated the power supply and got a completely different reading:

+5v is still holding out at 5.07v (in spec - no load), no change.
+12v line dropped to +10.35v (out of spec - no load), down from +11.64v
-12v line puts out -11.98v (in spec - no load), up from -10.62v

My previous measurements were based on the block connector disconnected from the board, but power was still being supplied to the display. This time I have removed the power supply from the unit entirely as to put it in a no-load condition.

Reply 21 of 40, by luckybob

User metadata
Rank l33t
Rank
l33t

Pics? I'm interested in the caps.

It is a mistake to think you can solve any major problems just with potatoes.

Reply 22 of 40, by Brickpad

User metadata
Rank Member
Rank
Member
luckybob wrote:

Pics? I'm interested in the caps.

Haven't gotten the shield off yet, but I got the security torx screws out. Probably going to drill out the rivets tonight to get a good look inside. I'll get pictures of it ASAP for you.

Reply 23 of 40, by Brickpad

User metadata
Rank Member
Rank
Member
luckybob wrote:

Pics? I'm interested in the caps.

Here you go. The four caps that are clustered together are 1800uF 10v. I believe this is the output side? It's hard to see from these pictures, but those four caps do have a slight bulge in them, and I suspect they need replacement. I'll pull them off tomorrow and test them all.

FEWJrJq.jpg
cDmjkh9.jpg
klkD3ZR.jpg
aE3lPsY.jpg

Reply 24 of 40, by luckybob

User metadata
Rank l33t
Rank
l33t

Shit... 105C HFZ Nippon chemi con caps.

And they look in perfect condition.

Well, I see several blue potentiometers, I would adjust the rails until they are all slightly over spec. like 12.1v or 5.2, Then test in the system. You should be able to trace the -12v line to the little standing board that controls its regulation.

Past that, if the psu holds up after being readjusted, I'm out of ideas. There might be other fuses in other locations. But I honestly don't know at this point.

It is a mistake to think you can solve any major problems just with potatoes.

Reply 25 of 40, by Brickpad

User metadata
Rank Member
Rank
Member
luckybob wrote:
Shit... 105C HFZ Nippon chemi con caps. […]
Show full quote

Shit... 105C HFZ Nippon chemi con caps.

And they look in perfect condition.

Well, I see several blue potentiometers, I would adjust the rails until they are all slightly over spec. like 12.1v or 5.2, Then test in the system. You should be able to trace the -12v line to the little standing board that controls its regulation.

Past that, if the psu holds up after being readjusted, I'm out of ideas. There might be other fuses in other locations. But I honestly don't know at this point.

They do look good in those photos, but if you could physically see them, you will see that they do have a slight bulge and not dead flat like the rest of the caps in the power supply, which is why I kind of think they're suspect. This guy here ( https://youtu.be/EExhdqfKPMw?t=5m50s ) had a similar problem, albeit a completely dead PSU, with the same caps which were not holding their true capacitance.

What I can't understand is why I got two different reading from the -12v and +12v under two different circumstances.

When I first measured the output, the power supply was isolated from everything except for the plasma display. My readings were +11.64 (+12v rail), which was in its proper tolerances, however the -12v rail gave me a reading of -10.64v, which is below minimum spec of -11.0v. +5v was +5.07, which was fine.

Now, this is where things make a change. When the power supply is completely isolated from everything, that is standing on its own disconnected from everything and pulled from the unit, I took another measurement. This time the +12v rail was below its minimum tolerance and giving me a reading of +10.36v. However the -12v rail was up to -11.98v, which was in the proper margin. Again the +5v rail was reading +5.07

Reply 26 of 40, by luckybob

User metadata
Rank l33t
Rank
l33t

You got it all apart. you might as well re-cap it. Check for bad solder joints like the guy in the video. I'm grasping at straws at this point.

It is a mistake to think you can solve any major problems just with potatoes.

Reply 27 of 40, by Brickpad

User metadata
Rank Member
Rank
Member
luckybob wrote:

You got it all apart. you might as well re-cap it. Check for bad solder joints like the guy in the video. I'm grasping at straws at this point.

Yeah, it's got me scratching my head too. Thanks for your help, by the way! Think I'll replace them with Rubys or Nichicons.

Reply 28 of 40, by Brickpad

User metadata
Rank Member
Rank
Member

I am an idiot. I read the meter wrong. The +12v line is normal, but the -12v is still showing below minimum tolerance. They didn't flip as I first thought.

Reply 29 of 40, by luckybob

User metadata
Rank l33t
Rank
l33t

If I was in your shoes, I'd recap the board and go from there. A partially shorted cap can be causing the voltage to go low like that.

While you are on Mouser or Digikey, go ahead and see if you can also get new mosfets. I can't image them being very expensive, and as long as you already are re-capping the board, just in case do the mosfets.

It is a mistake to think you can solve any major problems just with potatoes.

Reply 30 of 40, by Brickpad

User metadata
Rank Member
Rank
Member

That's the plan at this point. Going to replace the output cluster first. So far I have bookmarked these particular caps. Any suggestions or recommendations would be greatly appreciated.

Panasonic(1) 330uF 3 5v
http://www.digikey.com/product-detail/en/pana … 07CT-ND/3831074

Nichicon (4) 1,800uF 10v
http://www.digikey.com/product-detail/en/nich … 69-1-ND/3129206

Both of these were the closest to the appropriate dimensions, capacitance, and voltage rating.

For future reference:
1,800uF 10v - 34mm (h / seated) x 10mm (dia.)
330uF 35v - 29mm (h / seated) x 10mm (dia.)

Reply 31 of 40, by luckybob

User metadata
Rank l33t
Rank
l33t

Replace every cap. And I will always choose nichicons.

It is a mistake to think you can solve any major problems just with potatoes.

Reply 32 of 40, by Brickpad

User metadata
Rank Member
Rank
Member

I stopped in at a local electronics repair shop. I gave the guy all the specs for the caps I've pulled so far, and he gave me the following:

(1) Nichicon: 470uF 40v
(2) Rubycon 1500uF 16v
(2) Panasonic 2200uF 16v

Now according to this guy, it doesn't matter so much as they're all in series (My knowledge is limited at the moment), so I'll trust his judgement, but I am still skeptical.

[EDIT]

Replaced the caps, but still got the same readings. I tweaked the potentiometer and brought the -12v line up from -10.35v to -11.32v. Still won't POST. I'm now betting on that fact that it's the 3A SOC fuse at this point.

Reply 33 of 40, by Brickpad

User metadata
Rank Member
Rank
Member

Well, I'm fresh out of ideas. I've got the power supply recapped, trimmed the voltages according to specifications, and replaced the 3A fuse. Still no POST. I suppose at this point the board is fried. I can't think of anything else that I am missing.

Reply 34 of 40, by luckybob

User metadata
Rank l33t
Rank
l33t

well-shit-moments-dumpaday-18.jpg

It is a mistake to think you can solve any major problems just with potatoes.

Reply 35 of 40, by Brickpad

User metadata
Rank Member
Rank
Member
luckybob wrote:
http://www.dumpaday.com/wp-content/uploads/2013/07/well-shit-moments-dumpaday-18.jpg […]
Show full quote

well-shit-moments-dumpaday-18.jpg

That pretty much says it all. There is one other thing I haven't done yet and that is to test the voltages with everything attached under load and see if perhaps the voltages are dropping still.

Reply 36 of 40, by luckybob

User metadata
Rank l33t
Rank
l33t

have you replaced the regulators? Also the caps he sold you are high quality, so your good.

It is a mistake to think you can solve any major problems just with potatoes.

Reply 37 of 40, by PcBytes

User metadata
Rank l33t
Rank
l33t

Another thing you could try is repairing some traces if the original battery was one of those leaky stuff ones.

"Enter at your own peril, past the bolted door..."
Main PC: i5 3470, GB B75M-D3H, 16GB RAM, 2x1TB
98SE : P3 650, Soyo SY-6BA+IV, 384MB RAM, 80GB

Reply 38 of 40, by luckybob

User metadata
Rank l33t
Rank
l33t

no battery in this model, but good info!

It is a mistake to think you can solve any major problems just with potatoes.

Reply 39 of 40, by Brickpad

User metadata
Rank Member
Rank
Member

I wish it were that simple to fix, but thanks for the heads up. I did take a look at the IBM HMM and saw that the spec voltages are different than the ones I found online from another PS/2 site. The -12v should be around -11.5v to -12.6v, not -11.28v as I have it right now. I'll have to disassemble it again and fudge with the trimmer. Thankfully I didn't replace the pop rivets.