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

Printable View

02-25-2011, 12:56 AM
Cookiesowns

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

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:
http://img821.imageshack.us/img821/6615/deadddc1.jpg

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

http://img189.imageshack.us/img189/4043/deadddc2.jpg

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.
02-25-2011, 01:09 AM
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.
02-25-2011, 01:11 AM
Cookiesowns

Quote:

Originally Posted by 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.

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 :banana::banana::banana::banana:... Hopefully, No lung cancer is involved ROFL
02-25-2011, 01:24 AM
kroks

I've seen the exact same damage here
http://forum.hardware.fr/hfr/Overclo...6.htm#t2092963
:(
02-25-2011, 01:30 AM
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.
02-25-2011, 01:41 AM
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 :D

http://www.koolance.com/water-coolin...roduct_id=1058
02-25-2011, 03:00 AM
IanM

Quote:

Originally Posted by Cookiesowns

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.
02-25-2011, 03:04 AM
TJ TRICHEESE

some ddc's just randomly die, thats why people usually use two in series for redundancy.
02-25-2011, 03:08 AM
Cookiesowns

Quote:

Originally Posted by IanM

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.

Quote:

Originally Posted by TJ TRICHEESE

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
02-25-2011, 03:40 AM
IanM

Quote:

Originally Posted by Cookiesowns

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

Quote:

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
02-25-2011, 08:51 AM
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.
02-25-2011, 10:54 AM
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.
02-25-2011, 12:45 PM
drwheeler

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

Crater
02-25-2011, 12:50 PM
DexNfX

Is it really the XSPC res to blame?
02-25-2011, 12:52 PM
kroks

Quote:

Originally Posted by kroks

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

EK top here...
02-25-2011, 01:54 PM
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.
02-25-2011, 02:14 PM
Erklat

Quote:

Originally Posted by Waterlogged

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

:yepp:

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.

Quote:

Originally Posted by Cookiesowns

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 :D
02-26-2011, 01:03 AM
Cookiesowns

Quote:

Originally Posted by 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.

Quote:

Originally Posted by drwheeler

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

Crater

Quote:

Originally Posted by 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.

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..
02-26-2011, 08:30 AM
Waterlogged

Quote:

Originally Posted by DexNfX

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).
02-26-2011, 12:49 PM
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
02-26-2011, 01:11 PM
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.
02-26-2011, 04:33 PM
thegcpu

Quote:

Originally Posted by 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

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
02-27-2011, 09:02 AM
Cookiesowns

Quote:

Originally Posted by thegcpu

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.
02-27-2011, 11:13 AM
bmaverick

Quote:

Originally Posted by Waterlogged

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.
02-28-2011, 08:10 PM
Cookiesowns

Quote:

Originally Posted by bmaverick

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.
02-28-2011, 08:27 PM
stangracin3

Quote:

Originally Posted by matari

Is top to blame? official no, but un-officially NO. 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.

Fixed, you said it yourself in your own post
the user improperly installs the O-ring

my DDC w/ xpsc restop is still going strong after ~2years

according to you that makes me smarter than most people with a ddc. thanks
02-28-2011, 08:38 PM
bmaverick

Quote:

Originally Posted by Cookiesowns

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

Laing isn't in Asia. They are located in Hungry in Europe. The pumps work just fine in their stock configurations. Heat is a general killer of most electronics. The DDC series needs to have the base cool enough in a stock setting.

Now, adding a performance top to increase the flow will increase the noise, the heat and load effort for the little pump. The life span becomes unstable to predict.
02-28-2011, 08:40 PM
Cookiesowns

Quote:

Originally Posted by bmaverick

Laing isn't in Asia. They are located in Hungry in Europe. The pumps work just fine in their stock configurations. Heat is a general killer of most electronics. The DDC series needs to have the base cool enough in a stock setting.

Now, adding a performance top to increase the flow will increase the noise, the heat and load effort for the little pump. The life span becomes unstable to predict.

Yeah, I guess thats the problem.. Well, its kay then I guess..

Anyone have some figures of the MCP35x?
03-01-2011, 03:23 AM
Erklat

Quote:

Originally Posted by Cookiesowns

Anyone have some figures of the MCP35x?

Figures of flow or figures of death rate?

MCP350 indeed seems less affected by this, my still going strong after 2yrs.
I think I'll stick with them, imho they pack enough flow when doubled in a single loop, and between somewhat lower performance(?) and reliability I'll always opt for second. :up:
03-01-2011, 05:57 AM
Tackleberry

Quote:

Originally Posted by bmaverick

Laing isn't in Asia. They are located in Hungry in Europe. The pumps work just fine in their stock configurations. Heat is a general killer of most electronics. The DDC series needs to have the base cool enough in a stock setting.

Now, adding a performance top to increase the flow will increase the noise, the heat and load effort for the little pump. The life span becomes unstable to predict.

It is HungAry:rolleyes:
And they are from Germany.

about life span----any tests?
03-01-2011, 11:19 PM
Cookiesowns

Quote:

Originally Posted by Erklat

Figures of flow or figures of death rate?

MCP350 indeed seems less affected by this, my still going strong after 2yrs.
I think I'll stick with them, imho they pack enough flow when doubled in a single loop, and between somewhat lower performance(?) and reliability I'll always opt for second. :up:

Death rate
03-02-2011, 11:14 AM
zytrahus

I blame your xspc top. not the first time nor the last.
http://www.xtremesystems.org/forums/...d.php?t=258626
03-02-2011, 07:57 PM
stangracin3

Quote:

Originally Posted by zytrahus

I blame your xspc top. not the first time nor the last.
http://www.xtremesystems.org/forums/...d.php?t=258626

i blame the inability to check the o-ring seal

edit:

also read this post from the thread you linked

Quote:

Originally Posted by churchy

Are you shure that pumps aren't overtightened when mounting to top? I vaguely recall reading that by that you can introduce friction because of pump's impeller touching a bit top due to small clearance in some top cases. My memory might be wrong though. Also .. check how hot their PCBs get when used on that bay reservoir. In some builds that place is in airflow deadspot .. add to that 3.25 being the most powerful (and probably hottest) DDCs, and you might see some overheating/burning out PCBs issues.
03-02-2011, 10:54 PM
azcrazy

Quote:

Originally Posted by Cookiesowns

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 :banana::banana::banana::banana:... Hopefully, No lung cancer is involved ROFL

yea there is in the power section under hardwear temps, u havre to lower it is in 10's , 100-90-80
03-03-2011, 11:47 AM
NaeKuh

Im suprised no one told the op the cpu has whats called a thermal shutdown protocal, which was implimented on all intel cpus since p4 netburst days.

What that means, in the event of a cooling fail, the cpu will dow throttle itself until the processor hits a temp of 110c. At that point it will shut down the system.

Then when u try to start the system up again witha failed cooling, you will get a halt in bios, which says cpu has overheated.
03-03-2011, 01:33 PM
zeroibis

Quote:

Originally Posted by kroks

EK top here...

The top that was used does not matter as there are people who are going to say it was caused by XSPC regardless of what actually occurred. There seems to be a trend that creates the following condition:

public pump_fail(any:pump_used)
{
if(MCP355_pump==pump_used)
{

if(custom_top)
{
return("Caused by XSPC top");
}
else
{
return("Caused by XSPC top");
}
}
else
{
return("Caused by pump failure");
}
}

Now ignore the above code and note the important result. There appears to be a trend in the output of pump_fail(). Apparently pump failures are the result of 1 of 2 things, an XSPC top or a pump failure. Now there is another interesting trend where we find that that the only time that a pump can fail from its own failure is if a MCP335 is not involved. If a MCP335 is involved then the failure is caused by a XSPC top regardless of what top is used. This is why your statement about it being an EK top has no relevance to the cause of this problem. The only relevant information here is that the pump was a MCP335, this is the only thing that appears to matter when it comes to pump failures. :shakes:
03-03-2011, 02:30 PM
JoeBar

Quote:

Originally Posted by NaeKuh

Im suprised no one told the op the cpu has whats called a thermal shutdown protocal, which was implimented on all intel cpus since p4 netburst days.

What that means, in the event of a cooling fail, the cpu will dow throttle itself until the processor hits a temp of 110c. At that point it will shut down the system.

Then when u try to start the system up again witha failed cooling, you will get a halt in bios, which says cpu has overheated.

Very detailed description...

I sense a user experience... :lol:
03-03-2011, 02:59 PM
NaeKuh

Quote:

Originally Posted by JoeBar

I sense a user experience... :lol:

i plead the 5th.

But trust me, ive done just about everything stupid you can do with water and then some, to see how far i can get away with it.
03-05-2011, 10:41 PM
Cookiesowns

Quote:

Originally Posted by NaeKuh

Im suprised no one told the op the cpu has whats called a thermal shutdown protocal, which was implimented on all intel cpus since p4 netburst days.

What that means, in the event of a cooling fail, the cpu will dow throttle itself until the processor hits a temp of 110c. At that point it will shut down the system.

Then when u try to start the system up again witha failed cooling, you will get a halt in bios, which says cpu has overheated.

Many have, and I know about it, I'm just not a fan of running above 90c at 4.4ghz with 1.4125v :O I used to only get 65c max after 30 minutes of linux with 78F of ambient

All times are GMT -8. The time now is 04:38 PM.

XtremeSystems