Simple procedure for replacing Cap's on PSU...is this right?

[ran]

I just want to ensure I am doing things ok here - so I apologize if these questions seem infantile..

Please let me know if I'm missing anything majour....

I especially want to know if I'm failing to take any safety precautions here (e.g. do I need to "discharge" anything before opening up the PSU?)

My current procedure for replacing cap's in a PSU:
1. Unplug and remove the PSU
2. Lay the PSU down on a tray for soldering work
3. Wear grounding strap on my wrist
4. Open up the PSU
5. Desolder capacitors
6. Clean soldering areas with a little 90% Isopropyal alcohol and use small-gauge needles to make sure holes are clean (e.g. 27G/30G)
7. Re-solder new capacitors
8. Close the PSU
9. Replug the PSU (not installed in computer)
10. Attach a couple of 120mm/80mm fans via molex connectors (should have a load on the PSU, right?)
11. Jump the PSU with a small wire to the Green(Power on) line and a black ground line on the MB ATX connector
12. Verify stable +3.3v, +5v, and +12v lines on the PSU
13. Remove jumper to power off PSU and reinstall in computer.

Done.

Does that sound reasonable?

[gonzo0815]

Re: Simple procedure for replacing Cap's on PSU...is this right?

I would pull the plug during operation. This will discharge the HV capazitors and reduce the risk of lethal shok to probably near zero. I personally have never used antistatic wrist wraps or grounding strap during PSU rework. Imho it isn`t necesary, unless you are desolering some switching mosfet or tuff like that. But as long as you discharge it and won`t watch a open psu running it should not harm.

[ran]

Re: Simple procedure for replacing Cap's on PSU...is this right?

I would pull the plug during operation. This will discharge the HV capazitors and reduce the risk of lethal shok to probably near zero. I personally have never used antistatic wrist wraps or grounding strap during PSU rework. Imho it isn`t necesary, unless you are desolering some switching mosfet or tuff like that. But as long as you discharge it and won`t watch a open psu running it should not harm.

Thanks very much for the response. Unfortunately, it's already unplugged, opened and has several capacitors removed.

Should I try to do something to discharge the remaining capacitors (like these huge 200V 820uF Cap's?)

Ran

[linuxguru]

Re: Simple procedure for replacing Cap's on PSU...is this right?

In most cases there's a large resistor (~1M) that discharges the primary-side caps with an RC time constant of a few minutes. It's unlikely that there's much charge left in the primary capacitors now.

However, removed primary capacitors can develop a small residual charge due to hysteresis and dielectric absorbtion. That charge is enough to display sparks when shorted by a screwdriver or similar, even after several weeks out-of-circuit - they're safer in-circuit than out-of-circuit. Funny, huh?

[willawake]

Re: Simple procedure for replacing Cap's on PSU...is this right?

are you sure pulling the plug during operation is ok? i have seen a few reports of people causing damage when doing that to shutdown a stuck computer.

all the psus i have worked on have discharged properly. i think the problem is when working on damaged psus. i usually also use a 100 ohm resistor to discharge the primary caps and also finally tap around with an electricians screwdriver (usual thing i do when working on anything dangerous)

when recapping psus i think it is important to pay attention to the circuits on the rear of the board. the reason is that they are often not obvious and some circuits joined by solder which you might remove during the recapping. so for that reason i do one cap at a time.

i usually use a solder sucker half depressed to remove the solder. then i solder back using equal good amount of solder, fixing any tracks around the cap hole which became thinner during the recapping.

most important is to deal with the glue inside the psu by using a razor knife to cut around the cap or using micro clippers carefully. if the cap is not coming out it is probably still stuck and possibility to disturb some adjacent component.

the fans you are using to put a load on the psu wont put much load at all. maybe if it is a strong 120mm fan it will put 5w.

[gonzo0815]

Re: Simple procedure for replacing Cap's on PSU...is this right?

A good psu never minds a shut of during operation. Why it should? Even those crappy Dear and Codekon psu´S have no problems with that. If anything is getting dammaged, it is probalby a faulth wich are alreadey in the circuit before. I am not an expert of SMPS psu`S, but i am very familiar with the generall desing and circuits and how it works. From my point of view, there is no reason, why there should go something wrong. I know, that some auto voltage range switching psu`s have problems with certain voltage ranges, but if you unplug it or hit the switch, there is not mutch time for those problems to occure (only the hold up time, usually not more than 17 to 20ms).
But i agree, if someone feels uncomfortable in doing this, laying it discharging or even avoid touching the hot area is ok too.

[ran]

Re: Simple procedure for replacing Cap's on PSU...is this right?

the fans you are using to put a load on the psu wont put much load at all. maybe if it is a srong 120mm fan it will put 5w.

Well, I was "told" that turning on a PSU like without any load was actually bad for the PSU, so that's why I thought I'd hook up a couple of fans - one is a 120mm high-cfm fan (it can blow itself along my bench...)

Is this not true? If it is true, is there a better way to put a load on it without actually plugging it back into the MB?

[Per Hansson]

Re: Simple procedure for replacing Cap's on PSU...is this right?

The person that told you this is correct, the thing is that a fan is so little load it almost counts as nothing... A couple bad harddrives or something would be better...

Also if the powersupply has an on/off switch I flick that and then try to start the computer a few times, that clearly discharges the PSU because you can see the fans try to spin...