Reusing water-damaged computers

[DrBeaker]

Okay folks,

 

Long story short - We had a flood from a room above one of our IT suites. The computers in the IT suite were affected by this and had water damage/covered under water over the weekend.

 

There were decent PCs (Dell SFF, 4th gen i5, 8GB RAM) and were all in working order before this happened. The PCs were taken away and "tested" by some company involved with the insurance. They've had them for a few months and just returned them to us after the "testing" they've done. A lot have come back as "passed" from whoever this company was and they haven't been replaced under insurance for whatever reason. I've no idea who this "company" are, what they did to test them etc.

 

I've expressed my concerns about reliability of using this kit now due to this water damage, but I've been told to set these computers that were returned up and use them.

 

What steps or advice would people give?

Anyone else been in the same situation?

Anything to test them out and verify they work?

 

Any thoughts or ideas from the community welcome.

[Ditto]

Wow, that's a tricky one. You would hope any reputable insurance company wouldn't be taking any risk with returning electrical equipment, but I guess the testing conducted (no pun intended!) is as good as the company they used. My inclination would be to take a couple of the PCs to be independently assessed by a trusted and qualified third party.

[bald_pig]

The insurance would have more to lose if they gave them back and they caused a fire, so I reckon you're probably alright.

[DrBeaker]

The insurance would have more to lose if they gave them back and they caused a fire, so I reckon you're probably alright.

 

It was all handled by the SBM despite me asking for updates on it. I'm cynical (maybe unfairly) and suspect they just sat in a warehouse somewhere for weeks on end. We had all of them returned and our PAT tester failed about half of them when he tested them, but the rest passed.

 

Just want to make sure that due diligence has been done for Health and Safety etc

 

 

I've expressed from my viewpoint that for each of the units, their reliability and safety is now in question when put back into a production environment and personally I'm still not convinced, they should have been written off. Especially if some were returned as "passed" but our PAT tester failed them.

Edited September 11, 2019 by DrBeaker

[FN-GM]

With 4th Gen processors these are pretty old computers. Are they not on their way out anyway?

[3s-gtech]

Old?! Dear sweet innocent one, in a Welsh school these would be new-fangled enough to be viewed with suspicion and fear, like a caveman with fire.

 

They’ll most likely be fine. I’d do a visual check of the mobo and PSU to set my mind at ease, but water damage is not always terminal - especially if it was just the case getting damp. I’m surprised that the insurance didn’t bin the lot of them though.

[FishCustard]

I would ask to see the risk assessment the insurance company undertook for returning the PCs, and judge based on that. I agree that water damage isn't necessarily terminal.

[Tefters]

Have had this before...

I said I would only implement them with an email confirming it was their decision to re-instate them and that my recommendation of not using them for fire/child safety had been noted and considered.

 

Had to be on email format and signed off so I then have proof should something go wrong later down the line.

Since said incident 3 computers and 2 monitors out of 30 of each have fizzled out.

[noelmm]

If half of the returned machines failed your PAT testing process I would ask to see the documentation from the companies testing. I would also require an email from SLT confirming that you are required to use these machines when you are uncertain if they are safe or not, without this email I wouldn't install them on the grounds of health and safety.

 

Who's your H&S manager? Speak to them and ask them for clarification, once again in writing.

[Ditto]

With the PAT test details, I'd also be playing things a bit more hardball. Get this info to go back to the SBM to pass on to the insurer.

[Ditto]

P.S. Not 'cynical'. You are doing the right thing to query this and for the right reasons.

[hallb15]

Rust/Corrosion are likely causes of the failures. If not dried out properly, even the 'good' units could fail later due to delayed corrosion. I would definitely want to see a copy of the report the testing company did. Your SBM should be contacting the insurance company as they need to know whoever they've contracted to test the units has passed potentially dangerous equipment. Your PAT testing company will have issued a report-use that as leverage.

 

Send one of the PAT failed ones to bigclive.com hopefully he'll dissect it and then we can all see what happened and be better informed for future water damage incidents!

[DrBeaker]

With 4th Gen processors these are pretty old computers. Are they not on their way out anyway?

 

Ha! These if working, will be upgrades for the admin computers which are still using Core 2 Duo RM PCs!

 

Rust/Corrosion are likely causes of the failures. If not dried out properly, even the 'good' units could fail later due to delayed corrosion. I would definitely want to see a copy of the report the testing company did. Your SBM should be contacting the insurance company as they need to know whoever they've contracted to test the units has passed potentially dangerous equipment. Your PAT testing company will have issued a report-use that as leverage.

 

Send one of the PAT failed ones to bigclive.com hopefully he'll dissect it and then we can all see what happened and be better informed for future water damage incidents!

 

I've said from an IT viewpoint that their reliability is now in question now going into a "production" environment (staff actually using them) with no response back. I've said I'm concerned from an electrical safety point of view for staff and students (with no response back), but if the PAT test has 'passed' them I assume that would cover me in that area and push it back to them if there's any issues or comeback.

 

We got recently got rid of most of those affected/failed Dells with a e-recycling company.

 

I would want to at the least:

- check it over for damage/dirt/corrosion inside with a visual inspection.

- Do a memory, HDD or other stresstest of the unit

[HTCPCP]

I used to (when I had time and money) do quite a bit of water cooling, and, with that, have had water pouring out of my case on numerous times. Did it always break stuff? No, far from it. However, there were certain things that I always made sure I did after a leak or water cooling failure

Change the power supply if water got near it - Unless it was a top mounted PSU, or I was 100% confident water went nowhere near it, then it got replaced

everything was unplugged, and dried seperately - not always necessary, but ensures that there is nothing hiding anywhere.

Stress test everything - ideally comparing it to pre water incident performance - Run one of the many benchmarking tests on that machine, and if you have one, a machine that hasn't been water damaged

[DrBeaker]

I used to (when I had time and money) do quite a bit of water cooling, and, with that, have had water pouring out of my case on numerous times. Did it always break stuff? No, far from it. However, there were certain things that I always made sure I did after a leak or water cooling failure

Change the power supply if water got near it - Unless it was a top mounted PSU, or I was 100% confident water went nowhere near it, then it got replaced

everything was unplugged, and dried seperately - not always necessary, but ensures that there is nothing hiding anywhere.

Stress test everything - ideally comparing it to pre water incident performance - Run one of the many benchmarking tests on that machine, and if you have one, a machine that hasn't been water damaged

 

Any particular stress tests you'd recommend? Never really had to do something like this before, we'd have just replaced the lot as it was cheaper in the long run. I've got in mind DBAN to wipe and test the HDD, Memtest for a RAM test and the UBCD for a strestest, keeping it on for 24x hours for stability etc. The PAT-tester has passed them, but it's the same one in Easter who did a visual inspection and literally just stuck a sticker on stuff as I watched him in another room.

[hallb15]

For testing the CPU I've used https://www.cpuburnin.com/ for years. You need to run it for each CPU core on the machine - e.g. run4 instances simultaneously on a quad core CPU.

It certainly gets things nice and hot!

[HTCPCP]

I tend to use either Aida64 or SiSoft Sandra (neither are free), however on a basic level, something like

Memtest

CPUburnin - but keep an eye on how long it takes to throttle

Passmark - gives you an overall and easily readable figure on CPU performance!