SATA power plug fire hazards...

[BLinux]

so one my projects currently requires a 4-pin molex to SATA power adapter cable. as I was searching for some, I came across a mention that these cables pose a fire hazard. pictures paint a thousand words:

sata power fire - Google Search

And there's more discussion all over the internet too. In fact, I think there's a thread on STH about someone's server that caught fire due to the SATA power plug of a DVD drive.

some youtuber has some videos about it too:

his claim is that this problem is primarily with the "molded" SATA power plug side of the molex->SATA cables.

However, I don't fully buy that just yet. First, the google image search shows a few "crimped" SATA power connectors that caught fire. And secondly, if this is SATA power plug manufacturing defect that results in fire, wouldn't we see this also with other types of cables with SATA power and not just the molex->SATA power cables?

Anyway, anyone here have 1st hand experience? And what do you guys all think? I'm a little paranoid now about buying the cable i need ; don't want to introduce a fire in my rack or house.

 

[ttabbal]

I have some in use for years without issues. I don't doubt that it can happen, but I don't think it's particularly common. I also don't buy the cheapest ones on ebay though...

 

[BLinux]

I have some in use for years without issues. I don't doubt that it can happen, but I don't think it's particularly common. I also don't buy the cheapest ones on ebay though...

yeah, i think i must have 1 or 2 some where in some system that has been in use for years. however, that's such a small sampling, i don't think it would mean much. or, do you mean you have a much larger sampling and haven't ever run into this issue?

tbh, this is the first i've heard of it. but when I did a search, show many things popped up it got me concerned.

 

[ttabbal]

I doubt my setup could be considered statistically significant really. A total of 10 or so such adapters, 5400 and 7200 RPM SATA drives. Perhaps 6 years for most of them.

I have seen problems caused by people trying to run large arrays using a bunch of splitters. Just tripped the overcurrent protection on the PSU though. I could see something like that causing a problem with cheap PSUs though. People around here wouldn't do that, would they?

 

[T_Minus]

The bad ones (IMHO) have very thin wires and/or very brittle on the SATA side. You can tell the material (plastic?) used is of lower quality, and plugging and unplugging just a few times will often deform or crack them.

If this was widespread the reviews on amazon, monoprice, enwegg, etc, would reflect this... and many shitty quality adapters do reflect their quality in reviews. I think it's Cablematters? that I've had good luck with all their cables/adapters.

 

[BLinux]

i don't think this is a matter of overloading the rail, or thin wiring.

check out this video:

 

[PCBONEZ]

The contaminant inclusions shown in some of the parts in the video appear to be flux left over from acid/rosin core solder.
Hence the related corrosion issues.
Proper cleaning post soldering would eliminate the problem.
Acid Free 'No Clean' flux would do it too.

I realize that only helps if you build your own.
Small 'nameless' CN/TW/etc.. manufacturers aren't going to sacrifice production numbers for quality any time soon.
.
.
Related:
"Stacking" Y-Splitters may contribute to this if you aren't paying attention.
Most modern drives pull about 2 amps on 12v during spinup, a (rare) few up to 2.25a.
So, the pins in SATA connectors (1.5 amp/pin) are 'safe' for 2 drives max.
The pins in HD4P (usually inappropriately called "Molex connectors") are safe for 4 drives max.
.. Typically the HD4P have 9 amp pins. 11a and 13a pins are available but there is no way to know if the higher amp pins were used or not. The safe bet is to assume they are 9 amp rated.
.

 

[Patriot]

To um update this thread... with first hand experience... Amphenol adapter failed on a customers build, was not loose or askew to create a short but shorted internally due to thermal expansion allowed by soft molded plastic. There is no pcb damage on the DVD drive side as to suggest it was the cause.

Fire was thankfully very localized... no soot on motherboard and no nearby cables show any heat damage... aside from the adapter of course.
Even the sata data connectors attached to the optical drives are fine. OCP shut down the rig immediately and it was hefted outside.

Some minor scarring on the paint but otherwise fine.

 

[AndrewX192]

It's a shame that with SAS and NVMe SSDs if you don't use a backplane, you don't even have a choice to use crimped power connectors. Have you contacted Amphenol or the vendor you bought that adapter from?

 

[T_Minus]

"shorted internally due to thermal expansion allowed by soft molded plastic"

Clearly it's failed, and went up in smoke but do you know 100% it was due to the "expansion allowed by soft molded plastic" or maybe it could have been a manufacturing defect internally with the plastic from the get go?

I only question the cause of the failure because of how many of these are in use in SC216 chassis w\rear 2.5" hot-swap and if it was simply a "bad" plastic that allows expansion we'd see thousands of these fail, and we don't. That makes me think it's more manufacturing problem than the material actually used.

clearly amphenol isn't "safe" from this issue like the other brands that do use cheap plastic as we'd hoped tho

@Patriot thoughts?

 

[Patriot]

"shorted internally due to thermal expansion allowed by soft molded plastic"

Clearly it's failed, and went up in smoke but do you know 100% it was due to the "expansion allowed by soft molded plastic" or maybe it could have been a manufacturing defect internally with the plastic from the get go?

I only question the cause of the failure because of how many of these are in use in SC216 chassis w\rear 2.5" hot-swap and if it was simply a "bad" plastic that allows expansion we'd see thousands of these fail, and we don't. That makes me think it's more manufacturing problem than the material actually used.

clearly amphenol isn't "safe" from this issue like the other brands that do use cheap plastic as we'd hoped tho

@Patriot thoughts?

No but I was planning on shaving down to the half that isnt melted and getting the stereoscope out for a closer look at the soldering to see if I can root cause it any better than a guess...
As someone on IRC pointed out, despite my confidence that it is a genuine amphenol cable... it could have been one that got recalled or a knockoff.