• Please review our updated Terms and Rules here

Working 5150 Died, can I test Motherboard caps in situ with ESR?

trickydee

New Member
Joined
Sep 18, 2011
Messages
9
Location
UK
Hi All,

Over the last year I have been working on building a nice clean 5150. And had just got to the point where I had the machine working reliably when it stopped working ...

The machine was my second 5150 which I was also using to debug a fault on another 5150 (which has a constant noise on power up see my other threads). During this time I have managed to build a nice stash of working CPU's / RAM, video/floppy cards, PSU's etc.

Prior to the machines death, I had a fully working machine that would boot to DOS (original with a VGA card). Landmark digs ran clean (with a CGA card in place).
I noticed just before full death had 'hung' - with an DOS image still on the screen (from VGA). I rebooted the hung machine and now it does not post (blank screen/PSU fan spinning/ no beeps).

I have tried swapping (for know good):

IBM BIOS
Landmark BIOS
PSU
Video Card
CPU

I am now assuming that the issue is a Capacitor somewhere?. My soldering skills are bad, so I don't really want to start unsoldering caps to test.
My question is it possible to use an ESR tester to test the capacitors in situ?.

Am I missing something obvious (I did run through Modem7's excellent debugging list)

Any help appreciated.

Kind Regards

Rich
 
Someone may have another opinion, but in my experience with IBM motherboards, the caps (mostly all tantalums) either work perfectly or go short circuit. So I've never felt the need to do an ESR test - that's something I do on electryolytic capacitors and usually in monitors or power supplies.
 
First thing would be to measure the voltage at the board, is the PSU shutting down ? ( i believe some still spin the fan when shutdown) which would indicate a bad tantalum or you could measure the resistance between each supply and ground with the PSU unplugged to see if you have a shorted tant.
 
In general, no. The bypass capacitors are all in parallel with each other between Vcc and Ground, so if you try to measure one in circuit, you are measuring them all at the same time. You must desolder one leg.
 
In general, no. The bypass capacitors are all in parallel with each other between Vcc and Ground, so if you try to measure one in circuit, you are measuring them all at the same time. You must desolder one leg.

yes, but you can at least see if any are shorted, then you have to go one by one if none have any discolouration.
 
Wow thanks everyone for the responses.

I'll will try testing all the capacitors for shorts. I previously did this but got inconsistent results.
Am I right in thinking a multimeter set to ohm measurement (10k above) connected to both sides of the capacitor should show 0 with a short? and an increasing number if working?. My electronics skills are novice :)

I will also re test the voltages on the board as per Gary's comments, although I have used another known good PSU and got the same result - so I think it's more the board.
 
Yes. 0 or a very low ohms value indicates a short. When you first put the meter on it, if the cap is working right (or at least not fully shorted or open), it will show a low resistance, which will then go up as the capacitor charges (or a high resistance or open circuit if the capacitor is already charged). This is easier to see with an oldskool analog meter with a needle.

If you're seeing 0 ohms, then you probably have a shorted tantalum capacitor. Do this with everything including the PSU disconnected from the mobo and all the expansion cards pulled. From there I'd do a visual inspection, and desolder one leg of the most gnarly looking ones, one at a time and checking between each, until the short goes away. Then I would check each one that's been one-leg-desoldered and replace any that are bad. Or just replace all the desoldered ones if there aren't too many, since you already have one leg off anyway.

Just keep in mind that all those bypass caps are wired in parallel. When you're testing the capacitor in-circuit, you're testing them all at the same time, so one shorted bypass capacitor on the board will make all the rest show shorted too. That's why you need to desolder the worst-looking ones one at a time until the problem goes away.
 
Last edited:
I am now assuming that the issue is a Capacitor somewhere?
Based on your, "I noticed just before full death had 'hung' - with an DOS image still on the screen (from VGA). I rebooted the hung machine and now it does not post (blank screen/PSU fan spinning/ no beeps).", I do not think so. If a capacitor had gone short circuit whilst the 5150 was running, the power supply would have shut down at that instant, not kept running.

I will also re test the voltages on the board as per Gary's comments, although I have used another known good PSU and got the same result - so I think it's more the board.
Yes, if the voltages and POWER GOOD signal from the power supply are good, then the problem is not a short-circuit capacitor (overloading the power supply).

(I did run through Modem7's excellent debugging list)
A 'list' does not sound like the diagnostic procedure at [here]. That procedure, in the case of 'PSU fan spinning' would have had you measure the voltages and POWER GOOD signal.
 
Back
Top