Switching Power Supply Repair

[bryan]

Hello:

Have one of those common brick type switching power supplies. Typical 12vdc 3A output. Is still working fine, but I have narrowed this supply down to putting noise back onto the AC line which is affecting the control of some X10 equipment. The issue recently started so i suspect one of the internal components is failing, anything in particular I should look for, the electro caps physically appear fine, but will pull them tonight and check their ESR.?

 

[Nigel Goodwin]

Almost certainly electrolytic's, and could even be the large reservoir one.

If you've got an ESR meter you don't normally need to remove them to test them.

 

[bryan]

Thanks Nigel. Just ordered a ESR meter off EBAY. Seems to have favorable reviews on various sites. Did have a DIY ESR I made but broke the analog meter and can't be bothered to fix.


**broken link removed**

 

[KeepItSimpleStupid]

There is a relatively expensive filter for that. I had to place one on a Carrier furnace so the blower would not mess up the X-10.

If it worked before, it's probably the EMI filter components.

 

[bryan]

Turns out 1 dead 33uf 400v cap on the board. Easy find with my new ESR meter that arrived from Ebay the other day. Going to replace it along with all the others.

 

[Nigel Goodwin]

Turns out 1 dead 33uf 400v cap on the board. Easy find with my new ESR meter that arrived from Ebay the other day. Going to replace it along with all the others.

Presumably that's the reservoir capacitor? - amazing how some SMPSU's still work with that completely O/C, and yet some 'explode' if it is

It would explain the huge increase in RFI, that's a common symptom of it's failure.

However, one point to bear in mind, such capacitors don't test well on ESR meters - even good ones tend to have a far higher ESR than you would expect - but in this case it's likely to be so bad that it's obviously duff.

 

[tvtech]

Presumably that's the reservoir capacitor? - amazing how some SMPSU's still work with that completely O/C, and yet some 'explode' if it is

It would explain the huge increase in RFI, that's a common symptom of it's failure.

However, one point to bear in mind, such capacitors don't test well on ESR meters - even good ones tend to have a far higher ESR than you would expect - but in this case it's likely to be so bad that it's obviously duff.

Hi guys

Even easier with Main Smoothing Caps.....don't bother even taking it out to test it...

If you are running on 220VAC....you need around 300VDC to around 320VDC on that Cap for it to prove it is doing it's job...

Take your Meter and simply read what Voltage is on the Cap with Mains on.

If it significantly less than 300VDC....the Cap is stuffed.

Promise you

Regards,
tvtech

 

[Nigel Goodwin]

If it significantly less than 300VDC....the Cap is stuffed.

Promise you

That would have worked better if the OP wasn't in Canada

Same principle of course, and you are right - the voltage on the reservoir will be considerably less than it should be.

 

[tvtech]

That would have worked better if the OP wasn't in Canada

Same principle of course, and you are right - the voltage on the reservoir will be considerably less than it should be.

LOL Nigel

I am talking strictly 220VAC...like here. My bad

On 110VAC.. the same Principle applies. Just less DC Voltage across the MSC.

Regards,
tvtech

 

[ChrisP58]

Another way to find bad DC filter caps in power supplies is to set your meter on AC to measure the ripple voltage when the supply is running under load.

There will always be some ripple, but not too much. Unfortunately there is industry standard number for how much is too much, but I'll throw out a target of < 10% of the DC value.

I'll start measuring those values on good and bad supplies to gather some statistics.

 

[bryan]

Here is the output on a scope (AC coupling) with the bad cap 100mv per div



Here is a back up power supply with the same specs. I suspect the output will be similar when the defective one is fixed. 20mv per divison

 

[Nigel Goodwin]

Another way to find bad DC filter caps in power supplies is to set your meter on AC to measure the ripple voltage when the supply is running under load.

As always, only some meters would do that, for most you would have to add an external blocking capacitor.

But why bother?, simply measuring the DC voltage across it (as tvtech suggested) would clearly show the fault.