Dell GX620

[PCBONEZ]

Originally Posted by speg_head

Would it be safe to say that 0.00 is a good reading if the capacitance reading on those caps are accurate?

That's a good sign but not enough to call it good.
The reason is the test signal from those is at 100kHz.
A standard Multi-meter uses a DC test signal.
You need to know what the cap is doing at DC.
-
You DON'T need a fancy or high accuracy meter for this.
You aren't looking for a specific number, just 'high' or 'low' as like a go/no go check.
A cheap $5 [or $2 at some places] would do fine.

An old style analog meter [one with a needle] is much easier to use for this check because the needle swings to one end or the other quickly and you don't have to wait for 5+ second on each cap for the digital numbers to stop changing. - No so much of a bother when you are only checking 3 or 4 caps but it gets OLD quick if you have more caps to do.

Note: When you check can't know the caps initial charge vs discharge condition. Even if you short the cap first there is some residual charge. Polarity of the leads doesn't matter but you do have to wait for charge-discharge of the cap to steady out if you have leads reversed from the caps actual condition. The DC test signal in the meter will charge the cap in the direction of the lead's polarity.
- If the cap [assuming a good one] is partially charged in the direction of the leads polarity then it will go high and stop [or slow WAY down]. That point is where the meter should read way high.
- If the cap [assuming a good one] needs to discharge then charge to match the leads then with an analog meter you will see it go low, might pause, then it will go way high and stop [or slow WAY down]. Once all the excitement is over you have your meter reading.
- With a digital meter all you will see is numbers flashing and you won't know where it is in the cycle [or if it's done] so you end up waiting longer for a steady reading. With an analog meter when you see the reading way high and the needle is slowing to a stop you can call it good without waiting for it to stop completely.
.

[PCBONEZ]

You -are- checking with the caps out of circuit - right?
Checking caps that are in parallel doesn't work.....

[edspegeddd]

Yes I am testing all of the caps out of circuit. Im a bit confused with the multimeter test though. I have a digital one, not sure what setting to use to do the test. There is DCA, DCV, and a section labeled with an ohm sign. In each section are numbers ranging from 200m to 2000.

I tried setting it to 2000 in the ohm section and what the 2200uf capacitor shows are numbers that keep increasing to about 1900 , at that point it stops and goes back to default value of 1.

When doing a DCA test at the 2000u setting i get a 1 or 0 reading on the same cap.

If you can advise which test i should be doing would help,

Eddie

[PCBONEZ]

You don't test amps, you test resistance.
A resistance meter sends current through the device under test and compares it to the current through a known [internal to the meter] resistance.

You are looking for a high resistance.
You are using the 2k-ohm [2,000 ohm scale] and it is over-ranging.
Mine flash all digits when over ranged but meters are different.
Try the 20k-ohm scale. [20,000 ohms] That works on most meters.

.

[Daltex]

Just posting a pic of my usff gx620 mobo. Had zero knowledge of caps before stubling across this forum. Was getting erratic usb mouse behavior and other complants. Finally the bsod occured with memory error noted. Replace with known good ram and got same bsod. Read this forum and help topic etc. and decided to pop the cover and found 6 bad 2200/6.3's. Ordered from here and then upon further reading noticed another culpret- 820/6.3 grrrrr.

Anyway I have notice the prior pics showed three different caps that seemed to be bad are now poly's as of 2007. Just thought I'd post a pic for anyone that cares.

Thanks all for helping us newbs not trash an expensive item instead of a quick fix (hopefully).