Can you test a SATA controller?

[ohio_grad_06]

Run the drive hooked up as a secondary drive to another pc, and then run gsmart control on it to get a quick diagnosis. It would be interesting to see that on all 3 drives. If all 3 tested good, then I'd suspect board. For it to keep cooking drives, either dell is sending out drives that keep failing, hard to believe with 3 WD black drives in a row though, or something with the board is cooking drives. I think I would advise the client if the unit is not in warranty that given the price of the repair, you would recommend a new unit with an extended warranty in case of further issues instead of replacing the board, and inform them the price of the board plus 2-3 hours of your time would approach new machine cost anyway. Just my opinion.

 

[mr m]

I once had a Dell desktop corrupt HDD's. It turned out to be a faulty SATA cable. I swapped in a used HDD and started the onboard drive diags. I started to flex the SATA cable while the drive was being tested and found the culprit.

Dell has onboard and/or online diags to test their hardware. Although PC diags have never impressed me for their accuracy. Especially after using diags designed for mainframes and mini's.

 

[Diggs]

Back to your original question. There is no method available to us that will directly test the SATA chipset per se, that I am aware of. But Linux has a cli command, hdparm. This allows one to directly access some of the HD firmware by bypassing the overlying OS. You should be able to script the commands to test the SATA and HD together, looping commands.

Personally I've only seen a couple of times where I suspected the motherboard for something like this, multiple HD failures. In both cases I just ordered new MB's and never saw the customers again. Of course I have no way of knowing what the real problem was.

Same here. Over the years this is probably the third time I suspected a controller issue and instead of blowing by it I thought I would ask the question.

hdparm, pcils and a few others just identify and basically show the controller present. I figured someone had scripted something that might be able to do more but no luck so far.

Dell had sent 4 Seagate Momentous drives to the client. Warranty ended in June which is why they called me. The last one was from me the WD Black. No issues on the Win10 install. I never checked the SATA cable and don't have easy access to one in that configuration. Very good point if the client has been pulling hard drive in and out himself.

Client is long time customer and great to work with. He covered the cost of the drive and some of my time and is off to buy a new computer.

 

[Diggs]

Ding! Round two-

Same customer bought a new mid-range HP laptop for his son as a replacement. Seven weeks later that hard drive fails!!! They swear it has never left the desktop in seven weeks and hasn't even had the lid closed. They say they treated it like a desktop after the last go-around. I have known this customer for a long time and have no reason to doubt. (Hard drive tested bad in HP boot diagnostics. I didn't see any accelerometer app in diagnostics even though almost every new laptop has them now.) The kid really doesn't do anything with it other than a bit of web browsing and a huge amount of Roblox game playing.

(I put down the laptop and backed away slowly - told him to work the warranty.)

@ edit - 7 weeks out and the manager at the local Best Buy replaced it. (Best Buy has a 30-day policy. After that it's supposed to go to HP.)

 

[buzzcut]

Maybe they should scrap the laptops and stick to desktops.
Oh ya! strap and/or bolt it down to the desk.

 

[NviGate Systems]

I would have loaded HDD Sentinel and checked the GShock attribute. Laptop drives have an accelerometer and will record impacts above a certain threshold. OEM can and will deny warranty if the value is too high. Worth looking into.

Based on my experience, they are far too trusting of kid or it's a conspiracy.

 

[labtech]

Seagate Momentuses are some of the most sensitive drives to suspect handling. Looking at the g-sensor SMART attribute should help.
Best route is SSD to rule out the behavior.