• Please review our updated Terms and Rules here

Troubleshooting Compaq Portable III power supply failure

aperezbios

Member
Joined
Feb 26, 2016
Messages
36
Location
Santa Rosa, California
Hey folks,

I picked up a Compaq Portable III at VCF West this past weekend, and it came with a bad power supply. If I power it on with no load, the built-in fan powers up, but none of the DC output voltages are anywhere close to correct. If I power it on with a load, it sits there and ticks, and the fan never spins up.

There are no obviously blown or leaking electrolytic capacitors, so i'm trying to identify the first component, U2, on the PCB, which does not really have any meaningful markings on it, besides what vaguely looks Cyrillic, and "8908"

Compaq_Portable_III_PSU_U2.jpg

I've placed an album of other photos of the PSU at https://photos.app.goo.gl/6BhjECaeyDwiskFq5

Any advice would be appreciated.
 
...trying to identify the first component, U2, on the PCB, which does not really have any meaningful markings on it, besides what vaguely looks Cyrillic, and "8908"

It's a full-wave (i.e. "Bridge") rectifier (note the "sine-wave between the 2 center pins and a + and - on the edge). It converts AC in to full wave rectified voltage (likely the next component would be an inductor or large filter capacitor) The writing isn't Cyrillic---it's likely "house-markings" to determine the type and lot-code number. "8908" indicates it date of manufacture (8th week of 1989). While I've seen them fail, it's not common for them to fail without obvious signs. Remove it from the circuit and check with an ohmmeter to be certain (should not read anything less than roughly 500 ohms between any pins).

Heres a sample of another one (but it may or not be within the specs of yours:

https://www.digikey.com/product-detail/en/on-semiconductor/KBL06/KBL06FS-ND/1054251
 
Last edited:
U2 looks like a full bridge rectifier to me, but I could be wrong.
8908 is likely "8th week of 1989".

ESR meters are a good way to check the condition of electrolytic caps.
I'm not educated enough to help you diagnose the fault, but will be watching the thread, I enjoy learning about switch mode power supply diagnosis.
 
It's a full-wave (i.e. "Bridge") rectifier (note the "sine-wave between the 2 center pins and a + and - on the edge). It converts AC in to full wave rectified voltage (likely the next component would be an inductor or large filter capacitor) The writing isn't Cyrillic---it's likely "house-markings" to determine the type and lot-code number. "8908" indicates it date of manufacture (8th week of 1989). While I've seen them fail, it's not common for them to fail without obvious signs. Remove it from the circuit and check with an ohmmeter to be certain (should not read anything less than roughly 500 ohms between any pins).


Given that the DC fan spins up fine, it seems safe to say that the rectifier is doing its job. There are two Sprague manufactured 330uF filter caps, rated at 105C, which look fine when removed, but need to be tested. I unfortunately can't locate my ESR meter.
 
Actually, the fan is connected to the mains in most of those PSU so it will spin no matter how dead the power supply is.

They were a bad design for the whole series of portables that Compaq made and, if you find one of the portables, chances are VERY high that the PSU will be toast with about 20 different causes.
 
Actually, the fan is connected to the mains in most of those PSU so it will spin no matter how dead the power supply is.

That doesn't explain why the fan fails to work when the DC rails are loaded, not to mention he says unloaded he gets a DC output but extremely poorly regulated.
My guess is the caps are visually fine but are unable to deal with the inrush current due to age. I've seen this a few times. Replace them and it should be fine.
 
Actually, the fan is connected to the mains in most of those PSU so it will spin no matter how dead the power supply is.

In this case, the fan is in fact DC, and very clearly labelled as a NIDEC BETA V model TA300DC, with stated input voltage on the label as 12VDC. Therefore, the bridge rectifier is evidently doing its job :)
 
Last edited:
In this case, the fan is in fact DC, and very clearly labelled as a NIDEC BETA V model TA300DC, with stated input voltage on the label as 12VDC. Therefore, the bridge rectifier is evidently doing its job :)

In this case, it would seem so. However, almost all the units I've worked on had AC fans. Too bad though, the bridge rectifier is, at least, easy to replace :)
 
My portable 486 had a bunch of bad capacitors. The H/V caps were fine though. Some looked OK but were bad when tested.
 
Well, I couldn't find any concrete rules about necro-bumping timeframes (I admit I did not try very hard due to being too sick to really care) and this thread seems to be adequately relevant to my current repair project to bump, so...

I recently found my father's old Compac Portable III languishing upstairs outside of his office inside of its carrying bag. He says he got it right about when they came out, got the expansion unit, expanded the memory, got the 80287 math coprocessor, and got some networking or serial expansion card. Says he bought it for like $8k, which disturbs me because that's a lot even now and he basically just dumped it upstairs. I understand, as it's very obsolete, but at the same time...

Anyway, I plugged it in to see if it would boot up. I could hear it tick a relay repeatedly and it turned on its fan, so I assumed it was a power supply issue. I got the power supply out and tried it again out in the shop directly in front of my father's "smoking desk" so if I got electrocuted to death I wouldn't be in the abandoned upstairs offices. The power supply was able to run the fan and weakly trigger the relay only. I removed all of the electrolytic capacitors (except for the weird axial paper/polyester/whatever cap) from the board after labeling each by their accompanying silkscreen numbers, as I have a Fluke that can do capacitance measurements but I didn't think it can do so accurately with the components in place. Almost all of these aluminum caps tested at slightly over their labeled rated uF values, which my reading suggests does not indicate that they are actually healthy.

I honestly intended to just replace all of the electrolytic caps because I know that they have a somewhat limited lifespan and the electrolyte is not great for PCBs, but I have run into an unexpected problem: apparently these capacitors are not in physical sizes that places like Digikey and Mouser typically carry. The largest ones are so far where I'm getting stuck: there's a 6.3VDC 8200 uF that's 25mm diameter and 55 mm long, and another one that's something like 16V or 20V 1200 uF and 15mmx35mm.

I was wanting to get replacements from Nichicon, Rubycon, or TDK. TDK carries similar to the first one rated for a higher voltage and I think in a bigger can. My father (an electrical engineer) said that getting a cap rated for a higher voltage should be fine, but at the same time he doesn't want to talk about this at all and he doubts me when I say that electrolytic capacitors can fail by losing electrolyte over time due to not being hermetically sealed.

I will get all of the labeled ratings for the electrolytic capacitors, their dimensions, and also post the silkscreen labels for the wire connections (voltage and polarity mostly) as well as the wire colors for future reference or something (maybe tomorrow; I have written it all down and work is slow enough that I can also get out the digital calipers and measure lead spacing). Mostly, if anyone knows of where else I could try looking for these oddball sizes, I would really love constructive input. I would prefer that they be of high quality and not new-old-stock, since capacitors are pretty cheap as long as you're not buying hundreds or thousands of them.
 
The size doesn't really matter. As long as the voltage and value are correct and they have enough space then you can just bend the pins on the new ones to fit.
 
Uh, wow, I didn't think that post was going to get approved. Thanks for the input!

I fixed a Dell LCD monitor recently by replacing all of the electrolytic capacitors, but there was one problem: one capacitor replacement was smaller. This was also the "problem" capacitor, as it turns out, because my issue with the monitor backlight cutting out and turning back on persisted until I replaced that cap with another (slightly larger than the original). This suggests that going smaller than the original will cause problems, but I wanted to make sure that larger wouldn't also be an issue. Thankyou for clarifying there!

My father also said the voltage can be higher without issue.

The final capacitor that I've been looking at is a Sprague 430p metallized film capacitor with a 400V rating and 1.0 rating at +/- 20% tolerance. It doesn't say 1.0 what, but I think it's uF. I found a datasheet for the 430p series that does not include parts with 20% tolerance. The part number that I think is the most similar is 430P105X9400. If you look at its datasheet, which I don't seem to have on this laptop, it's about 12mm diameter by 45 mm long with axial terminations. The datasheet also specifies the rating as being DC with no mention of AC at all.

I bring this capacitor up specifically because I read somewhere that if the capacitor is involved with the AC mains, the rating has to be special or something. I can't tell if the capacitor interacts with AC that's not rectified, but anyway, I found KEMET #C4GAMUC4100AA1J seems to be close enough in size and rated a "little" higher on DC voltage. Assuming y'all think that will work out fine and that it's not going to be some special snowflake cap, I'll happily move forward with this project and we'll probably find out that the plasma display board needs new capacitors as well, or all of the tantalums will catch fire.

I did end up having to go with an 85°C 63VDC or 80VDC capacitor instead of the gigantic 105°C 6.3VDC 8200 uF monstrosity. I don't know if the conditions in the PSU will make me regret this. We'll see.

My other idea to replace the power supply was to get a modern replacement that's capable of fitting inside the housing for the old one, and get another power supply to take care of the 200VDC that the plasma display seems to call for in the user manual I found.

Specs and such for the caps etc. on the PSU are at my house on my desk, I hope I remember them at lunch.
 
In general, modern electrolytic caps are more space-efficient and higher quality than most older caps. You can usually get the same value at a higher voltage rating in the same physical space, which is always a plus. Designs that run electrolytic caps really close to their rated voltage are asking for trouble, IMO. Replacing a non-standard value with the nearest standard value is usually OK too (you can replace a 50uF with a 47).
 
Back
Top