Guys Check your SATA power plugs!!!

S

staticlag

[H]ard|Gawd
Joined
Mar 26, 2010
Messages
1,679
Pics http://imgur.com/a/TzOUB

Yeah so Friday night I notice a burning smell in my house.

I run into my computer room and instantly notice that it is off.

I notice that the light on the PSU is still on. So I hit the power off button on my UPS.

Intense burning plastic smell coming from my computer case. I pull off the case and check for flames. None. I let it air out and later come back with a flashlight.

My first thought is that the PSU blew a Cap. It's about 7 years old. But I wonder why the power light was still on in the back.

I'm kinda nervous about how much I might have lost. I have backups for all my data so that I'm not really that worried about loosing data; but I have quite a bit of hardware in my Coolermaster Haf X case.

Gtx680,
i7 2600k
16GB RAM
4 velociraptors
2 samsung evo 120gbs
2 7200rpm laptop drives
2 - 2TB drives
1 - 1 TB drive


First thing I sniff test the PSU. No smell.

I then use the flashlight to check the traces and the caps on the MB. All Good.

Then the Video card. Good.

Okay. So this has me worried... the only other spot it could be is in the drive cage.

So I fear that I may have lost at least 1, if not a couple drives since they are packed pretty tightly not only into the case's bay, but into 2 hotswap bays I installed.

I peer in. and see a burn mark on the top of my drive bay. Closer inspection reveals that a Molex -> SATA adapter spontaneously caught fire.

Luckly.... I disconnected it from my main SSD about 3 months ago when I got a new thermaltake 2.5" hotswap bay.

It was just hanging there. About .5" from a 2TB drive, and .5" below the case's firewall to the 5.25" bays. The 12V had shorted to the neutral next to it. Black dust everywhere.

I see the remainder of a word on the cable. (-henol)

A quick google gave me horror stories like this:

http://www.panthercomputers.com/index.php/faqs/38-company/177-immediate-advisory-product-recall

Yeah... so RUN. and check all your HTPCs, NAS'es, and gaming rigs to make sure you don't have any of this Amphenol junk in your machines.

I was so lucky that my adapter was in just the spot it was. hanging harmlessly in a drive bay with nothing above it but metal.

I powered up my PC after double checking everything, and like an A10 warthog... works perfectly, thank God!!!

I was lucky that I moved my SSD off of it and it wasn't connected to anything. Others are not so lucky and loose at least 1 drive to this.
 
Last edited:
What side appears to be shorting, the SATA side? The Molex side is a pretty stout design.
 
if your +12 shorted to ground the power supply should have gone into protection... I've seen power supplies with unprotected +5 and +3.3v rails but never run across one with an unprotected +12v rail

also 99% of the cheapy hot-swap trays I've seen are absolute garbage... use either a server from a large OEM or supermicro etc... everything is I have had too many issues with to trust...

also if you see henol, thats probably Amphenol, as far as I know they are a well respected company that makes a ton of different connectors amongst other things
 
I second that too, about a year ago I was trouble shooting a hard drive and the sata power connector shorted and caught on fire. Burned the shit out of the connector on the hard drive. Had to send it off to a guy in canada that does PCB work on hard drives.
 
One thing I notice with SATA power cables is it takes a lot less effort to work its way loose compared to molex. If it comes loose at an angle it could potentially short power and ground.
 
mi7chy said:
One thing I notice with SATA power cables is it takes a lot less effort to work its way loose compared to molex. If it comes loose at an angle it could potentially short power and ground.
Click to expand...

The OP said his was not connected to anything when it happened.
 
interesting, guess those adapters are bad news!!
 
Happened to me too.

I built a PC for a friend over 3 years ago.
Black smoke started to come out from it so quickly shut it down.
The SATA power connector to the hard drive was burned and misshapen.
Luckily the drive survived it and has been good since, so it was a problem only inside the power connector.
 
Engineering student, just got an email about AmphenolRF, their radio frequency hardware division i suspect, looking for interns. Maybe I should go clean up the place.
 
Are only Amphenol adapters affected? I have one of these (not in service) but it has no printing anywhere.
 
jwcalla said:
Are only Amphenol adapters affected? I have one of these (not in service) but it has no printing anywhere.
Click to expand...

It seems so.

I read somewhere that the plastic they used was contaminated and the contamination made the plastic conductive. Failure is only a matter of time for Amphenol adapters.
 
The only power cables I use are those supplied with the PSU (Corsair or Seasonic). Hopefully, none of those have issues...
 
don't use a adapter =)

With larger cases you need to buy Adapters for some stuff which is one of the reasons I'm sporting a mid tower...
 
If the problem was only with an adapter I would agree.
It was a connector on the PSU that set on fire in my friends PC.
 
I know this thread is a few months old, but I just got this one today:

10675563_10154880778025246_5106538125314501633_n.jpg


That adapter has been in that computer since 2008 when it was built from what I can tell.
 
I'm only going to say a little about this for various reasons...

1) I've seen this problem first hand.

2) We believe it's caused by a carbon "bridge" between +/- that's gradual enough for the PSU to consider this a resistive load vs. an actual short. (not the best explanation, but it'll do) This was actually verified by x-ray imaging. As such the PSU thinks it needs to pump out more amps, which it happily does until...

3) Things get hot enough and they smolder or flat out catch on fire

4) Now have custom Molex>SATA adapters with inline fuses to avoid prevent this from happening

5) There's speculation that the problem lies more with the clamshell design vs. molded design; or was it how the wires are actually terminated inside. I honestly can't remember which one is the problem... if I do I'll reply later.

BTW, I've NEVER liked the SATA power connector for this very reason. Yes, they're smaller and more idiot proof vs. the Molex conenctor; but all those pins in close proximity has bugged me since I first saw them in the late 90's... or whenever HDDs started using them.
 
Somehow I feel there should be a lawsuit about this. Stuff gets seriously damaged by these things.
 
A part of me agrees, but good luck with that! There are millions of these adapters in the world and a very small handful of failures. Granted one of our failures was catastrophic in MANY ways; but who knows who made the damn adapter and/or if they're even still in business.
 
fastgeek said:
2) We believe it's caused by a carbon "bridge" between +/- that's gradual enough for the PSU to consider this a resistive load vs. an actual short. (not the best explanation, but it'll do) This was actually verified by x-ray imaging. As such the PSU thinks it needs to pump out more amps, which it happily does until...
Click to expand...

Surely not, otherwise almost everyone would suffer from this issue and there would be a class action suit.

I think the design/material is so bad that it allows cross contact of earth pins to adjacent pins on the opposing connector when the molding isnt good enough, but not using the material it is made from.
 
Nenu said:
Surely not, otherwise almost everyone would suffer from this issue and there would be a class action suit.

I think the design/material is so bad that it allows cross contact of earth pins to adjacent pins on the opposing connector when the molding isnt good enough, but not using the material it is made from.
Click to expand...

Maybe you misunderstood me, or maybe I didn't word this properly. Overtime resistance is built up between the +/- pins that the PSU sees as a load, not as a short. We've had failed adapters analyzed by a very respected 3rd party FA lab and can assure you that, yes, surely so. Remember the bit about having these parts x-rayed? I've actually seen precisely what I'm attempting to describe.

As mentioned before, there was also some serious concern about how the three prong "trident" connector was attached to the wiring itself. If memory serves, the "strip and crimp" method was problematic as quality control in regards to the strip length and/or excess exposed wire beyond the crimp left a lot to be desired.

None the less, we designed and created a fused adapter and retrofitted hundreds of these things worldwide. Since then, we haven't had any problems.

If you knew what kind of financial nightmare this $1 adapter cost some very large companies, you would understand that I'm not bullshitting. :) And, no, I won't discuss those particulars.
 
fastgeek said:
Maybe you misunderstood me, or maybe I didn't word this properly. Overtime resistance is built up between the +/- pins that the PSU sees as a load, not as a short. We've had failed adapters analyzed by a very respected 3rd party FA lab and can assure you that, yes, surely so. Remember the bit about having these parts x-rayed? I've actually seen precisely what I'm attempting to describe.

As mentioned before, there was also some serious concern about how the three prong "trident" connector was attached to the wiring itself. If memory serves, the "strip and crimp" method was problematic as quality control in regards to the strip length and/or excess exposed wire beyond the crimp left a lot to be desired.

None the less, we designed and created a fused adapter and retrofitted hundreds of these things worldwide. Since then, we haven't had any problems.

If you knew what kind of financial nightmare this $1 adapter cost some very large companies, you would understand that I'm not bullshitting. :) And, no, I won't discuss those particulars.
Click to expand...

If resistance is building up, that implies it was a short circuit to begin with.
Build up of resistance is a reduction in conductivity.

If you mean that it becomes more conductive over time, I doubt that very much.
There would need to be a process that causes a material to change over time and that material would have to be able to become more conductive. Not that common a process.
If the process is caused by heat, then there must have been a problem connection to begin with, either a short circuit or a bad connection on a high power circuit.
As the condition you describe causes a short circuit after some time, then it does not start as a short circuit.
A hard drive connection is not a high power connection so a bad connection wouldnt produce enough heat.


The only thing that can produce enough heat to cause the level of destruction I saw is a short circuit of a power rail to another line.
This must be the direct problem.
Therefore, the connector is made badly or the connector material is too pliable, meaning a short circuit is possible.
 
For what it's worth, at least one company did the right thing and issued a recall.
http://www.panthercomputers.com/index.php/acs-west/38-company/177-immediate-advisory-product-recall

Nenu,

You're welcome to your opinion! I'm not an EE, so I might get some of my terms wrong and a lot of this happened well over a year ago. Simply recall being told that the operating theory, when all was said and done, that the PSU didn't view the problem as a short circuit, but a demand load and it just kept pumping power out until things went to hell. (Pyrolysis was mentioned) These PSUs put out over 50A on the 12V rail and 30A on the 5V, and there are three of them per system. Oh, and they were ruled out as being the source of the problem.
 
A PC power supply sees resistance, whether in the form of reactive or direct.
For a PSU to give out more current, there must be less resistance.
If there is high resistance, the current flow is low.

PSUs do not put out anywhere near 50A on the 12V or 30A on 5V rail unless they are presented with a very low resistance, like multiple gfx cards and a CPU running full tilt.
Not possible on a hard drive connector unless there is a short circuit.

The PSU cannot be the source of the problem unless it is outputting too high a voltage, even then it wouldnt destroy the connector, it would destroy the hard drive..
The problem is the connector, we already know that.

I dont trust the tests or information you have been given.
The results do not stand up to basic scrutiny.
 
As said, you're welcome to your opinion. Come back to me when you own your own well established FA lab and have x-rayed and analyzed all the failed components. :) BTW, you're the one who was questioning the PSUs, not me; there's no question that the problem was within the connector. Since the items in question were highly damaged / melted there's a lot of conjecture as to what happened.
 
You dont own a well established FA lab and dont have the skills to perform your own diagnosis.
You've been led astray and it sounds like you paid a lot of money for the privilege.

I didnt question the PSU internals because I havent ever thought they are the problem.
I was pointing out how easy it is to determine that they are not the problem.
You dont need to own or pay through the nose for a well established FA lab to work it out :p
 
Last edited:
As said, I'm not an EE; but there were multiple EEs involved from various companies. All I'm doing is trying to share some info. I wish the report wasn't under NDA as I'd gladly share it. (By the way, this didn't cost me a cent personally) If that's not good enough for you, then that's fine, honestly no skin off my nose. But until you have some first hand experience with the matter, and not simply being another internet know-it-all... (bites tongue)
 
I have first hand experience.
I built a PC and a SATA connector went up in smoke.
The hard drive survived so the problem occurred within the connector.

The same PSU is still working nearly 4 years later and the PC is rock stable.
A SATA connector on that PSU is still in use (the owner plugged another drive in) and hasnt degraded in any way.
Nothing is degrading or becoming conductive.

Its simply a short circuit that occurs due to the connector either being malformed or the connector material is soft enough such that when not fitted straight, 2 conductors can touch that otherwise should not be able to.
 
You must log in or register to reply here.
Back
Top