VOGONS


Reply 20 of 46, by whaka

User metadata
Rank Newbie
Rank
Newbie

yes, a bit off topic, but still useful i guess 😀
they're also not custom made for ibm. you can still buy 3 legged tantalum from digikey for exemple. they're only not widely used, it's different.

but to confirm what i said : https://i.stack.imgur.com/oqEyK.png

Reply 21 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

whaka -

So I see the one on that data sheet is positive center, but do they make them with the negative in the center as well?

I've got them all out now - so does anyone think it will cause any major problems to power up the planar with all of them out so I can see which ones are getting voltage, and the polarity?

Reply 22 of 46, by maxtherabbit

User metadata
Rank l33t
Rank
l33t
Deunan wrote on 2022-02-12, 22:12:

you can't be sure that in each and every case both outer leads are connected on the PCB. If they are - great, I was worried for nothing. But it has to be checked for every capacitor that will be replaced, to be sure. And it's not like this is a difficult or time consuming task to test that.

on the 5150 and 5160 they are all connected, you're right to recommend testing for certainty though

Reply 23 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

I'm going to check out the Sams photofacts and see if they don't have what I'm looking for - the more I look at this the more it looks like 2 different polarities on these. I really don't want to explode any tantalums in the house!

I have a message out to them and will update when I get more info.
Thanks!

Reply 24 of 46, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

Also remember the 5V line is negative to anything above 5V... then if they're on the -5 or -12, the ground rail is positive.

What I can't believe though, is that the component packaging and marking would be absolutely identical as in the pics of side by side ones in OP that are apparently oppositely connected.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 25 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

BitWrangler - that's my big problem! I've checked out the traces from top to bottom, and if you assume one is the ground and it connects to the two outside pins, but on the one next to it, that 'ground' now connects to the center pin!

I don't wanna smell any exploded tantalums!
😉

Reply 26 of 46, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie

Does it help if you check which pin on the capacitors goes to which pin on the J7 power connector? http://ohlandl.ipv7.net/8530/8530_Power.html says that the power connector has:
Ground - pin 2,5,6,7,8
+5 - pin 10,11,12
+12 - pin 3
-5 - pin 9
-12 - pin 4

Hopefully that'd cover most of them.

Reply 27 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

snufkin - that's probably a good way to do it! I wasn't thinking that, but that's what I'll do! It may be a few days before I get it done, but I will make it a point to get it done this week!

Thanks for the tip!

Reply 28 of 46, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Perhaps thig got lost in the OT noise I've made, or you want a second opinion, but let me restate that it should be safe to power the mobo without these caps, to measure voltages. At worst some of them are reset delay and the mobo will not boot properly but nothing should smoke. Again, I would check all the outer pads for continuity on the PCB, just in case. And let me repeat - do not assume that the center pin is negative just becuase it's connected to GND - it might be a filter cap for negative voltages. You should be able to figure that out by testing vs the power connector and then confirm what you found by measuring voltages on powered mobo.

Reply 29 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

Deunan - correct! I did get all of that - at first I was only going to pull the blown cap and power up the board from there, but I just decided to pull all of them while I was at it. I marked every cap with the board # and took pix of each one so I would know the orientation as well. Hopefully they'll all just be 1 type and I can get on with tracking down and ordering the caps!

Reply 30 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

Well, I've checked them to the power pins on the planar, and it's still soup as far as I can tell.. Here are the results using my trusty Fluke 77 multimeter's continuity test:

C84 - Center +5

C93 - none shorted to any pin on the power connector

C94 - Center +5

C99 - Center +5

C100 - all to GND

C101 - center to +12V red (this was the one with the blown top)

C104 - Center +5

C111 - all to GND

C113 - Center +5

C114 - Center GND

I'm probably going to have to break down and get the Sams Photofact for the model 30 and hope it's really similar enough to have these same caps on it.

Reply 31 of 46, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie

Ok, if you're measuring resistances to the power connector then you need to check all the pins on the capacitor. The ones with centre connected to +5V might have their outers connected to Ground. The one with centre connected to Ground probably have their outers connected to -5V or -12V. The one with all pins connected to ground suggests there's still a short somewhere, probably on -12 or -5. Don't know about the one not connected to the power connector. For that one you can use Deunan's method of powering the board up and measuring the voltages instead, but probably best to do that once you're sure there are no other shorts.

Remember that when describing the capacitor as centre +ve or -ve, that's relative to the other pins, nothing to do with the board voltages. Ground (0V) is more positive than -12V, so with -12V on the outer pins and 0V on the centre pin, that's still centre +ve.

Reply 32 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

snufkin - thanks for the info - I'll get more time next week hopefully to do more checking. After finding a couple of them with all 3 to GND I agree that powering it up may not be the best idea at this point.

On the list above, if I noted it as Center +5 then the outers went to GND, and if Center to GND, then the outers went to +5.

Reply 33 of 46, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

It's usually the 12V ones that short and explode. Typically the mobo has caps rated at 16V but I'd recommend replacing the 12V rail ones with 25V rated caps, just to be on the safe side. As for the reasons they short, it's age and possible defects they had in the first place, thermal cycling causing cracking and allowing moisture and oxygen in. Also most "new" builds of old PCs do away with HDDs and don't use floppy drives much, so 12V rail can perhaps spike during power-on events more than it used to.

Anyway, you should measure all the pins, not just the center ones. Caps with all pins seemingly shorted to each other can be an indication of mobo damage, or perhaps there's more tantalums on that PCB? Not 3-lead but standard 2-lead variety? Those too can be shorted. Inspect any ceramic disc caps for damage, those usually don't short on their own at such low voltages but physical damage can cause that.

In fact the very first test you should do now, with the tantalums removed, is check the power connector - each voltage rail vs GND. There should be no shorts. That is a requirement to even try and power up any motherboard.

Reply 34 of 46, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie

I think I found a photo of the board, and it looks like C93 might be connected to Q2, which on the photo I've found is marked RIA or R1A, which could be an NPN transistor or a small adjustable voltage reference. If it's an NPN transistor then it might be used to connect the outer pin of C93 to Ground. If true, then it's not directly connected to the main ground, so that's ok that you didn't find a connection. Worth checking that there's no short between centre and outer pins. I wonder if it's part of the keylock control, it's near J10 which I think is the keylock header. I'm slightly surprised there's no connection to one of the power supplies from the centre pin and wonder what it's connected to.

If you follow Deunan's advice and test for shorts on the J7 power connector then I guess you'll find pin 4 (-12V) shorted to ground. Looking at the traces near the J7 power connector, it looks like -12V goes to one of the outer pins of C100, and the centre pin goes to Ground. If that's right, then that means there's a short from -12V to Ground somewhere (so all C100 pins connect to ground). If that's true, then C100 is also centre +ve (-12V -> Ground, probably true for C111 as well).

I think that then leaves just C114 as the odd one out. All the others look to be centre +ve. Would you mind double checking the measurements for that one, once you've cleared any shorts?

The capacitor values look odd. From your earlier information then C113 is rated for 25V, but is used between +5V and Ground. But C101 was only rated for 16V and was used between +12V and Ground. I really would have expected those values to be the other way around. As Deunan said, 25V is better on the +/-12V lines.

Oh, last odd thing, on your first measurements you listed C84 as shorted, and now it's listed as connected to 5V. Did something change with that one?

Reply 35 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

Deunan - I'll see what I can do to get your checklist done! I appreciate the help! I'm pretty sure that there are only 10 tantalum caps on the board, but I'll go over it again just to triple check - I've been doing this later at night and usually I'm pooped by then.. 😉

snufkin - Thanks for the help - C84 was shorted when the cap was installed. I've pulled all 10 of them to do this last batch of testing. Also, C101 is the one that blew the top off, and of course the markings are on the destroyed side - but it's exactly the same size as 99 and 100 - the 25V ones are noticeably larger (see the 2nd picture in my original post of 99, 100, 101).

Reply 36 of 46, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie
TheRealAnubis wrote on 2022-02-20, 01:34:

C84 was shorted when the cap was installed. I've pulled all 10 of them to do this last batch of testing. Also, C101 is the one that blew the top off, and of course the markings are on the destroyed side - but it's exactly the same size as 99 and 100 - the 25V ones are noticeably larger (see the 2nd picture in my original post of 99, 100, 101).

That's a bit odd. If there was a short between +5V and Gnd, which is what you've measured for where C84 connects, then you should have measured a short between +5V and Ground for all of the +5V capacitors (C84, C94, C99, C104 and C113). All of the +5 points are connected together, as are all of the Ground points, so a short anywhere between +5 and Ground will show up as a short everywhere between +5 and Ground.

If you still have them, check if any of the capacitors you removed measures short from centre to outer. Also, check if the middle pad of C84 is shorted to the centre pad of C94 (should be), and that the outer pads between C84 and C94 are also shorted.

Reply 37 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

snufkin - I have them all in a ziploc baggie - I'll check each of them as well to see what I get. Who knows, maybe initially I fat-fingered it and the probes were touching? The pins are pretty cramped.

check if any of the capacitors you removed measures short from centre to outer

None show a short.

Also, check if the middle pad of C84 is shorted to the centre pad of C94 (should be), and that the outer pads between C84 and C94 are also shorted.

Yes and yes.

Reply 38 of 46, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie

Ok, so maybe the probes were touching originally. Have you been able to check the PSU connector yet? If you measure the resistance from J7 pin 2 to every other pin you should find shorts on pins 5,6,7,8 and not on any other pin. That will check there aren't any power-ground shorts. Pin 3 (+12V) should not be short to any other pin. Same with pin 4 (-12V) and pin 9 (-5V). Pin 10 (+5V) should be short to pins 11 and 12, but not any other pin.

Want to check these because on C100 and C111 you measured that all pads connected to ground, which doesn't sound right.

Incidentally, someone here: http://www.ibmfiles.com/pages/ps2model25.htm found that all the 3 leg tants on their board had centre +ve. They don't actually specify which revision they had, but it must be a little different as they only had 9 tants on their board.

Reply 39 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

snufkin - I was on that page a bit ago and totally missed that - looks like their board had one less 10-16 than mine.

Pin 2 on J7 shorts to 4,5,6,7,8.

Pin 3 does not short to any of the other pins

Pin 4 shorts to 2,5,6,7,8

Pin 9 does not short to any of the other pins

Pin 10 shorts to 11 and 12

Looks like pin 4, the blue (-12V) is grounded out.

I have had an offer to swap this board for a working one, so if you think it'd be too much to figure out this problem, let me know - I'm fine with trying to find the problem as this planar is likely the original one to the system and I'd like to keep it like that, but when it comes down to it, I'd like a working model 25 even more!