Psu and pat testing

[chrbb]

Right I know this has been discussed before but I need to know what fellow technicians do in the same situation as me.

Sole technician, no onsite Pat tester - I thought I had a failing psu in a problematic computer, so I swopped one out from an idential computer to test it. Psu returned to original computer. Problematic computer solved by installing graphics card as it was onboard graphics that has failed, so original psu put back in and computer now working.

However last pat tester we had informed me that when I replace a psu in a computer it should be pat tested before returning to general use. What does everyone else, in the same situation with no onsite Pat tester, do?

[Hacksawbob]

Say nowt, It has returned to the original tested configuration and will be tested in the routine testing procedures. If its really an issue then keep a separate PSU to one side for testing purposes and/or spend the £300 for an in house tester. Although you will probably get asked to do the whole school if you go down that route if the T+Cs of your contract allow for it, could get onerous!

[powdarrmonkey]

Take formal guidance from your SMT. Testing is not a legal requirement (contrary to popular belief), but providing equipment that is safe for a competent person to use is (hence an often-blanket policy that it must be PATed, because this is cheap and convenient). It's up to your school or LEA's local policy, and if they require it to be tested they must provide the means to do so.

[Jamman960]

If we're talking about having a computer re-tested because its had its psu removed and put back in then I can't see how anyone can insist on a re-test, even if the psu has been swapped for another previously tested PSU I'd still query the need for a re-test. you'll be carrying out a visual PAT test while installing it and it'd have already been declared electronically safe. As long as your happy that it looks ok and you've not dropped or have any reason to believe that it may have been damaged then it should be safe.

If you've replaced he PSU with another untested/brand new PSU then I guess you'd need to treat it as you would any new bit of kit, either insist on pat testing before it goes out or ensure its pat tested on the next cycle.

[chrbb]

I've put the computers back into circulation as the psu was previously tested and deemed safe. But I'm unsure as to why he said a new Psu would need testing, is it the integrity of the new part or that I might muck up putting it in and trap a few wires or something?

[synaesthesia]

how does that work? PAT testing on a PC is a visual inspection only - the only part that can be tested by any means is the cable itself, not plugged into anything (if that's even possible).

[powdarrmonkey]

If we're talking about having a computer re-tested because its had its psu removed and put back in then I can't see how anyone can insist on a re-test, even if the psu has been swapped for another previously tested PSU I'd still query the need for a re-test. you'll be carrying out a visual PAT test while installing it and it'd have already been declared electronically safe.

Wrong. Part of a formal PAT is the earth bond formed between the metal case and the case of the PSU. If they are exchanged, their characteristics change.

As long as your happy that it looks ok and you've not dropped or have any reason to believe that it may have been damaged then it should be safe.

Wrong. The installation matters as much as the visual.

I've put the computers back into circulation as the psu was previously tested and deemed safe. But I'm unsure as to why he said a new Psu would need testing, is it the integrity of the new part or that I might muck up putting it in and trap a few wires or something?

See above.

how does that work? PAT testing on a PC is a visual inspection only - the only part that can be tested by any means is the cable itself, not plugged into anything (if that's even possible).

Wrong. A PC is a class 1 item, so it gets*

1. Formal visual inspection
2. Earth bond test (100mA conductivity, less than 0.10ohm resistance)
3. Fuse and polarity test
4. Insulation test (must be >10Mohm)
5. Powered run test within manufacturer's stated current requirements and acceptable leakage

*at least, these are my local requirements. Yours may differ.

The IEC lead gets a separate, aptly-named IEC test.


This is precisely why you should be asking your SMT instead of relying on folklore.

[Andrew_C]

Powered run test within manufacturer's stated current requirements and acceptable leakage

This is the only bit I'd differ with; the PC may not complete a boot cycle, and thus the run test may be skipped.

That said, I know what I'd do in the OPs situation, but I'll not say in public.

[powdarrmonkey]

This is the only bit I'd differ with; the PC may not complete a boot cycle, and thus the run test may be skipped.

I've come across testers with an extended run time, so it can be done. But if I'm going to handle the machine, I'm far more worried that it passes earth and insulation

[chrbb]

Wrong. Part of a formal PAT is the earth bond formed between the metal case and the case of the PSU. If they are exchanged, their characteristics change.

Even with the same psu reinstalled?

[powdarrmonkey]

Even with the same psu reinstalled?

Perhaps. If the screws holding it in place aren't tight, for example, the resistance between it and the case changes. That affects the earth bond.

This is academic though. You should still be getting an answer from SMT (then it's their neck on the line, not yours).

[mjs_mjs]

1. Formal visual inspection
2. Earth bond test (100mA conductivity, less than 0.10ohm resistance)
3. Fuse and polarity test
4. Insulation test (must be >10Mohm)
5. Powered run test within manufacturer's stated current requirements and acceptable leakage

This is pretty much what a PAT test is, but there are more things on our in house PAT tester, but i'm not trained to use it. Any change to a PC should in theory require a PAT test, as you have changed the make up. New equipment should also be PAT tested as it may have been damaged internally while in transport. But as with everything like this, the person doing the PAT test should decide what does need doing and when. If you don't have one (someone who's in charge of it) ask up you line management until someone gives you a definitive yes/no, or insist it gets done if your worried about it.

Most faults we will be dealing with as technicians will be ones easily identifiable by visual inspection, anything that is still suspect should go to an electrical engineer, along with yearly checks on most equipment that's in use. That's my professional opinion.

[powdarrmonkey]

This is pretty much what a PAT test is, but there are more things on our in house PAT tester, but i'm not trained to use it.

One of them is probably a flash test, which you most certainly do not do to a PC. Ever.

[mjs_mjs]

One of them is probably a flash test, which you most certainly do not do to a PC. Ever.

Yep, we have that feature. I remember the person telling me that he doesn't do that on our PC's. Or any other devices with low voltage elec's.

[powdarrmonkey]

Yep, we have that feature. I remember the person telling me that he doesn't do that on our PC's. Or any other devices with low voltage elec's.

That's because it's an application of 3,000V between the case and the live+neutral shorted together. It kills things if you don't understand it, and it's no longer recommended for in-service equipment by the code of practice, though I believe it's still done by manufacturers in controlled conditions.