VOGONS


First post, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

Hi,

I just got a new to me IBM Model 25, and it shows nothing at boot, no beeps, nothing. Power supply is outputting the correct voltage.

I pulled it completely apart today and found one of the 10uf 16v tantalums with no top. There are 7 total of these, and three 22uf 25v ones. They all have 3 legs.

Does anyone have a schematic or parts list for the model 25, or can someone 100% sure tell me if the polarity is (- + -) or (+ - +) ?

I'm just going to replace them all since at least one other is shorted.

Thanks!

Attachments

Reply 1 of 46, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

Taking a psychic reading off the pictures, the voices say the middle leg is ground.

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 2 of 46, by darry

User metadata
Rank l33t++
Rank
l33t++
BitWrangler wrote on 2022-02-11, 04:48:

Taking a psychic reading off the pictures, the voices say the middle leg is ground.

If they are similar to these, https://electronics.stackexchange.com/questio … this-a-dual-cap , it would be the opposite .

Measuring resistance between a known ground point and each pin should give an idea (at least once the shorted one is taken out of the equation).

Disclaimer : I use Google and my limited experience. Please don't just take my comment at face value and try to corroborate/validate with more knowledgeable people .

Reply 3 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

I thought middle ground as well - but I'm going to probably do some testing as well. I figured someone would have a schematic, but I haven't run across one yet, so maybe not!

Thanks!

Reply 4 of 46, by stamasd

User metadata
Rank l33t
Rank
l33t

Best way to figure out is with a multimeter in continuity mode. Test each pin vs. GND and +5V. Then you'll know.

I/O, I/O,
It's off to disk I go,
With a bit and a byte
And a read and a write,
I/O, I/O

Reply 5 of 46, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Ah yes, it wouldn't be IBM without custom, unobtainium parts in it. I wouldn't put it past them to create special tantalum pass capacitors either. So even if it has 3 leads, two (most liketly outer ones) might actually be internally connected. Do check that because it will affect how you install a replacement. Middle pin would normally be GND or (-) but I agree that it would be best to just test that with an ohm meter vs power connector. Anything is possible with custom parts.

EDIT: Is that a burnt via, on the first photo, left of the crystal resonator?

Reply 6 of 46, by pentiumspeed

User metadata
Rank l33t
Rank
l33t

These type of 3 legged design is omni-directional. Works either way because if i remember correctly. The outside legs is a wire which connects to one end of capacitor, on other side of capacitor is connected to middle wire.

Cheers,

Great Northern aka Canada.

Reply 7 of 46, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie

Polarity will need to be measured. This: http://minuszerodegrees.net/failure/failure%20-%203lt.htm says IBM have used - + - before.

[I think they get used because it means you can't fit them the wrong way round. Still have to get the right type in the first place though.]

Reply 8 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

Snufkin - right - I read that page as well, which then got me to this point. Schematic anyone?!?

Deunan - oof! I'll have to get a better look at that as well! Was too busy focusing on the cap markings.

Reply 9 of 46, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

I disagree with the "Replacing 3-legged with 2-legged" on that page. You have 3 leads because it's a pass capacitor after all. Might not always be used or wired this way in the circuit but assuming that both outer pins will have electrical connection between them, without the capacitor in place, is asking for trouble. Always check that before installing 2 lead replacement. If necessary add a wire, this will of course reduce the filtering effect for higher frequencies - but it's better than nothing.

Reply 10 of 46, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

I am curious now what those cheap component testers say if you plug one into it. ... but I don't think I've got any loose.

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 11 of 46, by whaka

User metadata
Rank Newbie
Rank
Newbie
Deunan wrote on 2022-02-11, 17:25:

I disagree with the "Replacing 3-legged with 2-legged" on that page. You have 3 leads because it's a pass capacitor after all.

no, they're not "pass" or anything else special.
3 leads just give the ability to not care about the way you mount them. in any way, you can't be wrong. that's all.
i replaced thoses in a model 30 8086 by regular aluminium without any issues.

Reply 12 of 46, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

I thought he was saying that they might be used instead of wire links or "zero ohm" resistors to bridge tracks over so you can avoid having to have another PCB layer. If they are also doing that duty as well as capacitoring, then you need to make sure you join the dots.

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 13 of 46, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
whaka wrote on 2022-02-11, 22:25:

no, they're not "pass" or anything else special.
3 leads just give the ability to not care about the way you mount them. in any way, you can't be wrong. that's all.
i replaced thoses in a model 30 8086 by regular aluminium without any issues.

You seem to think there were some new guys on the assembly line so IBM was nice and designed the PCB to use custom, and definitely more expensive (if for no other reason then economies of scale) parts so that they wouldn't have hard time putting them in right. Well, these things are automated, and even if for some reason there was a human doing that job, they would be trained and checked. These days even the work done by machines is visually inspected by other machines.

I don't buy this "easier to mount" argument. It makes no sense in mass production where cost is everything. Somehow no other manufacturer, ever, had so many issues with this that they used special capacitors. These are not "proper" pass capacitors, sure, but I still think that was the main reason for the design. I would suggest replacing them with good brand ones like AVX or Kemet, the cheap no-name ones can easily have twice the ESR (comparable with standard electrolytics).

Reply 14 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

Thanks for all the information everyone! I'm planning on replacing them with 3 leg Kemet or Sprague (which I think these are).

I'll test all of them (since there's only 10) and report back if they're center negative or positive.

Reply 15 of 46, by TheRealAnubis

User metadata
Rank Newbie
Rank
Newbie

Well - this is interesting - here are my test results:

C 111 - center negative (22-25)

C 113 - center positive (22-25)

C 114 - center negative (22-25)

C 99 - center positive (10-16)

C 100 - center negative (10-16)

C 101 - blown (center pos?) (10-16)

C 93 - none test negative (10-16)

C? - by the keyboard fuse F1 - center positive (10-16)

C 104 - center positive (10-16)

C 84 - shorted (10-16)

PS - Deunan - I checked that area and I think it may be where a piece of that exploded tantalum landed - there's nothing there, and nothing on the other side either, but there is a rusty orange residue in that area that smells like burnt hair.

Reply 16 of 46, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

In my defense I took my psychic reading mostly off C114 and C100 😜

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 17 of 46, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

It's weird that IBM used two different kind of capacitors, but then again, it's IBM. However, do not assume that the center pin is negative just becuase it's connected to GND - it might be for negative voltages. Since you tested them all, remove the shorted ones, check the connection between outside pins in case you do need to add a wire, then just power the mobo up. If one of those is for some reset delay or someting it won't boot properly, but nothing bad should happen. Now check the voltages to confirm which pin is what, that should confirm your current findings. If not, investigate.

Reply 18 of 46, by whaka

User metadata
Rank Newbie
Rank
Newbie
Deunan wrote on 2022-02-12, 01:10:
whaka wrote on 2022-02-11, 22:25:

no, they're not "pass" or anything else special.
3 leads just give the ability to not care about the way you mount them. in any way, you can't be wrong. that's all.
i replaced thoses in a model 30 8086 by regular aluminium without any issues.

You seem to think there were some new guys on the assembly line so IBM was nice and designed the PCB to use custom, and definitely more expensive (if for no other reason then economies of scale) parts so that they wouldn't have hard time putting them in right. Well, these things are automated, and even if for some reason there was a human doing that job, they would be trained and checked. These days even the work done by machines is visually inspected by other machines.

I don't buy this "easier to mount" argument. It makes no sense in mass production where cost is everything. Somehow no other manufacturer, ever, had so many issues with this that they used special capacitors. These are not "proper" pass capacitors, sure, but I still think that was the main reason for the design. I would suggest replacing them with good brand ones like AVX or Kemet, the cheap no-name ones can easily have twice the ESR (comparable with standard electrolytics).

don't buy this argument if you want, but they are nothing less than "no possible polarity mistake".
maybe you never saw it, but wrongly mounted capacitors on automated chain is a reality.

Reply 19 of 46, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

This is getting off topic and I don't want to derail this thread any further - so fine, IBM found a "solution" to a problem everybody else dealt with without using custom parts. But my point stands - 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.