Another dead MCP355 DDC3.2 almost took my 980x with it.

[Cookiesowns]

Well so today while I was working with a few partners for a business document, I smelt a STRONG plastic PCB burning smell, So while I was checking where it was coming from, my pc bsod'd, so then as soon as I noticed it, I cut the power to everything, upon opening up my system, right around where my water pump is I noticed a strong burning smell... So I figured its either the pump or my fan controller. I took everything down and re-checked, seemed fine. So i did a dry test, with only power to the pump. GPU loop pump seems perfectly fine... then When i looked at the cpu pump... wasn't running. So i shut down my pc, stripped everything.

Took out my res, and voila, HUGE scorched mark on the cpu DDC pump...

I'm typing on the 980x as we speak, the logs showed that the cpu hit about 90c before bsoding, at 4.4ghz with 1.41v.... Luckily it's running fine on stock clocks, and stress testing as we speak on air.

Stress testing seems fine........

Let me upload pics:


See the melted plastic? This isn't even the bad part




Ouch, wonder where the damage is from?

NOTE, THERE WAS NO WATER LEAK, that loop was running fine for the past 3 months or so, I checked, O-Ring was intact, no water leaks whatsoever... The pump just spontaneously combusted.. Luckily, it didn't take my precious 980x with it.

[C-N]

Looks bad but if this is a regular occurrence it would be wise to set max temp shutdown a bit lower.... would be a real shame to loose a nice chip like that for a poxy pump.

[Cookiesowns]

Looks bad but if this is a regular occurrence it would be wise to set max temp shutdown a bit lower.... would be a real shame to loose a nice chip like that for a poxy pump.

I'm not sure theres a shutdown temp on the R3E is there?

I'm not sure what to do, shall I contact GABE @ Swiftech? I don't have the box for the pump anymore..

Laing Should really re-design these pumps, this is not only a huge health risk, but also pose a serious issue, If the pump burns, it could potentially cause a fire, or just take down whatever is with it. With that being said, Luckily no harm was done, except a dead 60$ pump.

My Living room now smells like serious ... Hopefully, No lung cancer is involved ROFL

[kroks]

I've seen the exact same damage here
http://forum.hardware.fr/hfr/Overclo...6.htm#t2092963

[C-N]

Sorry can't help with the R3E I've never owned one but would be very surprises if that feature is missing.

I would if I were u contact Swiftec, even if they do nothing for you the fault will be logged & if they see enough of them something may get done about it.

[phantomferrari]

are you running the pump at the full 12v? id say maybe get the ddc 3.25 . i believe martin did a review and stated that it consumes less power than the ddc 3.2 so this also hopefully means less heat

ive heard that these things get hot and i thought this product sounded a little silly. looks like you could use it

http://www.koolance.com/water-coolin...roduct_id=1058

[IanM]

I'm typing on the 980x as we speak, the logs showed that the cpu hit about 90c before bsoding,

No need to worry about the CPU, it will throttle down and then shut off at 90°C so your log is just showing that Intel's CPU saving emergency shutdown features are working as exactly as expected.

[TJ TRICHEESE]

some ddc's just randomly die, thats why people usually use two in series for redundancy.

[Cookiesowns]

No need to worry about the CPU, it will throttle down and then shut off at 90°C so your log is just showing that Intel's CPU saving emergency shutdown features are working as exactly as it expected.

Hmm... Well I'm not soo sure, as I might have disabled intel TM, and speedstep LOL, cpu seems fine, I'll do some more testing tommorow.

some ddc's just randomly die, thats why people usually use two in series for redundancy.

Well, if it only died, I would be less pissed or more pissed, not sure, as the smell did get me alert, but the fact that the way it melted down, scared me quite a bit....

if there was anything flammable near it, a wire, or paper, even the foamy anti-vibration pad, I'm pretty sure there would be a possibility of a fire.

I'll be contacting swiftech tommorow morning to see what I can do about it.


Whats your opinion? I'm thinking about switching to a single loop setup as I put my 580s back under water, bigger mess of the tubing and all, but I guess it would be worth it.

I would possibly need a new resevoir correct? Unless you people somehow know how to rig it so that I can run the pumps in series

I currently have the XSPC dual bay dual DDC res for dual loop setups :P

[IanM]

Hmm... Well I'm not soo sure, as I might have disabled intel TM, and speedstep LOL, cpu seems fine, I'll do some more testing tommorow.

I think the safety feature is still active, I don't think it can be disabled. In the Volume 1 datasheet here: http://www.intel.com/design/corei7ee/documentation.htm

6.2.3
THERMTRIP# Signal
Regardless of whether or not Adaptive Thermal Monitor is enabled, in the event of a catastrophic cooling failure, the processor will automatically shut down when the silicon has reached an elevated temperature (refer to the THERMTRIP# definition in Table 5-1). THERMTRIP# activation is independent of processor activity. The temperature at which THERMTRIP# asserts is not user configurable and is not software visible

[Waterlogged]

XSPC res kills yet another DDC.

Warranty voided if stock top is removed.

Only pump that retains warranty with an after market top is the MCP35X but, that's only because it comes with a different stock top. I'm guessing the warranty is voided on that one too if removed.

[matari]

If the pump and O-ring are not installed correctly, water can slowly trickle inside. Really hard to notice. Those reservoir are made by hand. I am wondering if the pump and o ring did seal correctly.

[drwheeler]

Looks like explosive FET damage. Mine was caused by water leaking past the O-ring.

Crater

[DexNfX]

Is it really the XSPC res to blame?

[kroks]

I've seen the exact same damage here
http://forum.hardware.fr/hfr/Overclo...6.htm#t2092963

EK top here...

[matari]

Is top to blame? official no, but un-officially yes. If the O-ring is not installed correctly, the water will slowly leak in. And remember that reservoir is made by hand, there may be some imperfections that caused the problem.

[Erklat]

I'm guessing the warranty is voided on that one too if removed.

That is what Gabe told me when I said I ain't paying $20 for a top I do not want in that pump's thread.

Hmm... Well I'm not soo sure, as I might have disabled intel TM, and speedstep LOL

You can't disable shutdown. Don't you worry, it is almost impossible to put CPU on fire. Well, if you are using s1156 or s1155 it is somewhat easier but hard nevertheless

[Cookiesowns]

If the pump and O-ring are not installed correctly, water can slowly trickle inside. Really hard to notice. Those reservoir are made by hand. I am wondering if the pump and o ring did seal correctly.

Looks like explosive FET damage. Mine was caused by water leaking past the O-ring.

Crater

Is top to blame? official no, but un-officially yes. If the O-ring is not installed correctly, the water will slowly leak in. And remember that reservoir is made by hand, there may be some imperfections that caused the problem.

As I've once said before, the O-Ring seal was perfect, NO WATER LEAKS AT ALL, if it were to leak, It would've done it much sooner, Seem to just be caused by a defective FET then I believe, these things do get hot..

[Waterlogged]

Is it really the XSPC res to blame?

Nearly 85-90% of all current DDC deaths can be informally linked to a XSPC res top. The rest of the loop does play a part but, the res top is the one commonality of the dead pump loops. I tried to do further research into the matter but met resistance from the community. Lots of ppl were willing to fill out the poll but, very few were willing to give further loop info which was the most important part of the survey. I had the topic deleted so it couldn't be taken out of context (in case you go looking for it).

[86 5.0L]

your 980x is perfectly safe, 100*C is when thermal protection kicks in(aka clock modulation), if the chip keeps getting hotter(around 135 °C, I believe), it will then trip the THERMTRIP# Signal. These cannot be turned off or messed with by the user.

my i7 920 has cruised at 100*C many times before, its still kicking @ 4.2GHz

Your PC BSOD'ed because it become unstable, not because of any safety measures, it would just lock up if THERMTRIP was activated

[ScottALot]

Yeah I had an XSPC single pump res top and there was some sort of glue keeping the reservoir and res-to-pump interface together that decided to fail and water just started leaking... no more XSPC tops for me.

[thegcpu]

your 980x is perfectly safe, 100*C is when thermal protection kicks in(aka clock modulation), if the chip keeps getting hotter(around 135 °C, I believe), it will then trip the THERMTRIP# Signal. These cannot be turned off or messed with by the user.

my i7 920 has cruised at 100*C many times before, its still kicking @ 4.2GHz

Your PC BSOD'ed because it become unstable, not because of any safety measures, it would just lock up if THERMTRIP was activated

I concur that @ 90c your 980x is fine. Here is a video of my 875K running @ 99c for 10 minutes. One question did any water get into the pump? My 355 burned like that, but I had water go down the power lines into the pump. Go to 2:20 in the video


http://www.youtube.com/user/thegcpu?...23/Hx1QwCUZSKg

[Cookiesowns]

I concur that @ 90c your 980x is fine. Here is a video of my 875K running @ 99c for 10 minutes. One question did any water get into the pump? My 355 burned like that, but I had water go down the power lines into the pump. Go to 2:20 in the video


http://www.youtube.com/user/thegcpu?...23/Hx1QwCUZSKg

Yeah, seems fine. Doing 3.8ghz stable on stock air cooler , as to water with the pump. Let me repeat myself, I had the same loop running, for about 3 months, I don't touch anything, and the O-Ring was perfectly intact. If it were to leak, I believe due to the confined spaces both pump would've fried. But no, only one pump did, and it would be the cpu.

I still think the cause is by faulty worksmanship.

[bmaverick]

Nearly 85-90% of all current DDC deaths can be informally linked to a XSPC res top. The rest of the loop does play a part but, the res top is the one commonality of the dead pump loops. I tried to do further research into the matter but met resistance from the community. Lots of ppl were willing to fill out the poll but, very few were willing to give further loop info which was the most important part of the survey. I had the topic deleted so it couldn't be taken out of context (in case you go looking for it).

This discussion was popular a few weeks back on OCN. The original poster was blaming that ALL versions of the DDC pumps were victims to this problem without much proof beyond the 3.2 PCB versions. There have been no reported problems of PCB failures of the MCP350 (DDC-1) black impeller types using the Phillips IC. The pre-3.2 PCB pumps running at 12VDC@10W has a totally different PCB.

[Cookiesowns]

This discussion was popular a few weeks back on OCN. The original poster was blaming that ALL versions of the DDC pumps were victims to this problem without much proof beyond the 3.2 PCB versions. There have been no reported problems of PCB failures of the MCP350 (DDC-1) black impeller types using the Phillips IC. The pre-3.2 PCB pumps running at 12VDC@10W has a totally different PCB.

Interesting, I mean, its just the fact that these pumps fail nearly the same way, I think theres a flaw somewhere. I'm not going to blame laing, there a taiwanese company correct? Maybe I can talk some sense into them, haha.