pump went bad, some good came out of it

[Elganja]

I have a dual loop setup... out of nowhere today my computer was acting really funny. It would turn off out of nowhere, my first inclination was that the psu went bad. Then i went to the bios to look at the cpu temp's and so on... 85C!!!

quickly checked to see if there was flow for my cpu loop, and none. I tried changing power sources with my 2nd loop with no luck. Tried burping it to make sure there was no air in the line, no luck

So i try to make something good out of the bad... gives me a chance to move to the RP-402X2 reservoir. Had to order a new pump too to replace my bad swiftech 355, so I got a PMP-400 since I have amazon prime. It will also give me time to do some much needed fluid change maintenance that i have been putting off.

so in the end, shitty day gives me an excuse to buy new parts

I'll have a couple EK Waterblocks EK-DDC X-Top Rev 2 up for sale soon, and will RMA my bad swiftech pump and keep it around for backup.

 

[rolodomo]

I'm using a swiftech 355 that's been in heavy use for two years, so your story interests me.

Did you have the pump "rpm" leads connected to the CPU fan rpm sensor pins on the motherboard? I'm counting on this to give me a "CPU No Fan Detected" warning message during boot so I know right off the bat something is wrong, but I wonder if this is just an illusory alert feature. I have no way to test whether I would get the warning if the pump had power, but mechanically failed.

I also need to configure my temperature monitoring software to alert me at about 80 C or so before things get really out of hand and the computer shuts down.

Thanks for passing this info. along.

 

[Elganja]

I'm using a swiftech 355 that's been in heavy use for two years, so your story interests me.

Did you have the pump "rpm" leads connected to the CPU fan rpm sensor pins on the motherboard? I'm counting on this to give me a "CPU No Fan Detected" warning message during boot so I know right off the bat something is wrong, but I wonder if this is just an illusory alert feature. I have no way to test whether I would get the warning if the pump had power, but mechanically failed.

I also need to configure my temperature monitoring software to alert me at about 80 C or so before things get really out of hand and the computer shuts down.

Thanks for passing this info. along.

Didn't use the rpm lead... with the xspc reservoirs I have, I can see the water flowing from the back. The CPU loop wasn't flowing at all, while the gpu loop was flowing nicely. I then did the "feel the pump" test and could tell it wasn't working at all. I tried swapping each pumps power source just to confirm, and it made no difference. Sucks, but shit happens...

I'll tell you what though, I'm grateful for thermal protection

 

[Lightflash]

Glad that you found the problem quickly. Also glad that you had something good come from a dead pump.

 

[Interitus]

How old was the 355?

 

[Elganja]

How old was the 355?

About 9 months old.

 

[BTRY B 529th FA BN]

Nice!

Still got my loop setup - but drained. Had one for 775 and 1366 but the case I put it in ended up being to small to work in. The one thing I didn't like was the motor noise from the pump. I should break it down and put it in my 800D

 

[aigomorla]

RMA the pump... which vendor did u buy it from?

liang handles rma's though vendors... so notify the vendor.

 

[Elganja]

RMA the pump... which vendor did u buy it from?

liang handles rma's though vendors... so notify the vendor.

Jabtech... just sent them an email. I fully intend to rma it and keep a backup pump around

 

[bleucharm28]

Hmmm, just curious. Is it better to leave the computer off when I'm not at home? Or would that cause more problems for PC or the pump itself.

 

[Elganja]

RMA the pump... which vendor did u buy it from?

liang handles rma's though vendors... so notify the vendor.

here was jab-tech's response:

[FONT=Verdana, Arial, Helvetica]Hi, You will need to contact Swiftech directly for warranty service on this item.

John
JAB Computers
www.jab-tech.com [/FONT]

This is what I pretty much figured from what I read online, so I'll contact swiftech directly.

 

[Elganja]

swiftech told me they will not warranty it, since I had an EK top on it... sucks

edit: just read their full warranty and it states it plain as day. oh well... gotta pay to play i guess

 

[aigomorla]

swiftech told me they will not warranty it, since I had an EK top on it... sucks

edit: just read their full warranty and it states it plain as day. oh well... gotta pay to play i guess

You know what im gonna post a thread on XS about this.

Edit: Done:
http://www.xtremesystems.org/forums/showthread.php?t=271340

I thought they got rid of all the nonsense.

And swiftech themselves use an aftermarket top now on the 35x series.

If this is the case, i wont buy another swiftech pump again.
I'll get the laing though koolance, exactly like how you did.

 

[Elganja]

You know what im gonna post a thread on XS about this.

I thought they got rid of all the nonsense.

And swiftech themselves use an aftermarket top now on the 35x series.

If this is the case, i wont buy another swiftech pump again.
I'll get the laing though koolance.

here is what Gabe sent me

Dear XXXXXX,

We are in regret to notify you that we are declining warranty support for this item. In effect, disassembling your pump and/or use with an aftermarket top void your warranty.

For reference, even when we sell our own top, like the MCP35X Housing, it voids the warranty (see here: http://www.swiftech.com/mcp35xhousing.aspx , “important warranty information”)


Best regards,


Gabe

I can only say FAIL... it's okay though, like you I probably won't buy another swiftech pump. I'll stick with the koolance pmp400's

 

[aigomorla]

the sad thing is the 35x is the ONLY pwm ddc in its class.

The Koolance PMP-400 is the DDC3.25 not the 3.2 so its a better DDC then the original.

But koolance will RMA it unless it shows clear water damage.. ie.. leaking inside the pcb from a water leak, to a broken / snapped PCB.

I had a talk with one of there tech's about it.

 

[Elganja]

aigo I know you love the dual loop, but... after a lot of reading and having a pump fail on me... I am going to go to a single loop w/ 2 pumps. Still using the koolance reservoir (it allows such a setup w/ 2 pumps 1 loop)

Things that pushed me over to the edge:
-will have redudant pumps
-things will run cooler... by that I mean, the cpu and gpu are rarely maxed at the same time, so in the end, when the cpu needs more radiator it will get it, when the gpu needs more radiator, it will get it.
-it's dead simple to switch to a single loop

Old setup:
-Pump -> gtx360 -> gpu #1 -> gpu #2 -> res #1 -> repeat
-Pump -> gtx260 -> cpu -> northbridge/southbridge -> res #2 -> repeat

New setup:
-Pump #1 -> gtx360 -> gpu #1 -> gpu #2 -> gtx260 -> cpu -> northbridge/southbridge -> Pump #2 -> Res -> repeat

 

[aigomorla]

from koolance:
No, opening up the pump does not void the warranty. The warranty does not cover damage resulting from accident, misuse or abuse, lack or reasonable care, shipping damage, modifications, the affixing of any attachment not provided with he product, or operating components at speeds or functions other than those specified by their manufacturers.


no more swiftech pumps for me...
and i wont be recommending them either regardless of pwm if u lose warranty.

 

[Elganja]

from koolance:
No, opening up the pump does not void the warranty. The warranty does not cover damage resulting from accident, misuse or abuse, lack or reasonable care, shipping damage, modifications, the affixing of any attachment not provided with he product, or operating components at speeds or functions other than those specified by their manufacturers.


no more swiftech pumps for me...
and i wont be recommending them either regardless of pwm if u lose warranty.

ygpm about some other stuff

i wish i could just post on xtremesystems... prob could find my answers there

 

[rolodomo]

After reading this thread, here are my thoughts on the best steps to detect loop failures before CPU thermal protection kicks in. Some of this stuff I should have done a long time ago!

(1) electronic flow sensor with alarm - not only detects a pump failure quickly, but also perhaps a catastrophic leak in the loop. Con: $ and flow decrease.

(2) bios high-temp shutdown feature (I've just enabled this in my bios to 85 C). Con: not a standard bios feature and not sure whether shutdown would be graceful.

(3) auto-start Realtemp in tray during OS boot configured for graceful high temp shutdown using the windows shutdown.exe service (I've just tested this feature, works great. I have it set for 80 C currently). Con: can't think of one really - simple and free step to add a little extra protection.

(4) connect pump rpm lead to cpu fan sensor on motherboard. Con: you only get warned about a pump failure if it happens at/before boot (unless there's a program out there that shuts down the computer on a CPU fan zero rpm trigger, such as Speedfan??).

 

[Elganja]

After reading this thread, here are my thoughts on the best steps to detect loop failures before CPU thermal protection kicks in. Some of this stuff I should have done a long time ago!

(1) electronic flow sensor with alarm - not only detects a pump failure quickly, but also perhaps a catastrophic leak in the loop if you place it near the pump intake. Con: $ and flow decrease.

(2) bios high-temp shutdown feature (I've just enabled this in my bios to 85 C). Con: not a standard bios feature and not sure whether shutdown would be graceful.

(3) auto-start Realtemp in tray during OS boot configured for graceful high temp shutdown using the windows shutdown.exe service (I've just tested this feature, works great. I have it set for 80 C currently). Con: can't think of one really - simple and free step to add a little extra protection.

(4) connect pump rpm lead to cpu fan sensor on motherboard. Con: you only get warned about a pump failure if it happens at/before boot (unless there's a program out there that shuts down the computer on a CPU fan zero rpm trigger, such as Speedfan??).

damn good idea... can set that up with my gpu as well.

 

[rolodomo]

can set that up with my gpu as well.

Good point, I've got two GTX 480(s) on my loop. Not difficult to imagine a scenario where they run into thermal problems well before the CPU does!

BTW, I had to use the windows task scheduler to start Realtemp at login and then select a tickbox to run Realtemp with highest privileges. Here's a good thread with pictures: http://www.overclockers.com/forums/showthread.php?t=617710

 

[aigomorla]

1. flow sensor... not needed... unless u want to diag your loop, and u like to watch a funny little propeller spin.

2. bios high temp shutdown feature... very useful, however cpu's have thermal shut downs incase it gets 100C.
The system will reboot. And then bios will post CPU Overheating..

And i would know... im guilty many times for not remembering to plug in the stupid pump.

3. mmmm... software shutdown is at the mercy of the system. Meaning if your system hickups, and software crashes, there goes all your fault protection.

I prefer manual hard displays like temp probes everywhere and in loop.

And inside my Rad chamber:

4. Most definitely... if its a ddc and has a RPM header, i +1 this all the way.
Average RPM on a DDC is anywhere from 3800rpms -> 4500rpm.


5. Get a Sunbeam PWM Rheobus.
Plug rheobuss pwm into motherboard.
Plug fans into rheobus.

Then you can Dyno your fans via software like speedfan or straight bios
Meh... watch video:
http://s125.photobucket.com/albums/p73/aigomorla/Nadeshiko/?action=view&current=MOL001.mp4

LOL.... i felt like i was dynoing my PC system.... ahaha...
Those of u guys who live in server rooms, Yes, those are the noise of SanAce, about 6 of them ramped to 2800rpms at full throttle.

 

[Elganja]

ffs... amazon sent me a rev 1.0 res not the 1.1 so i have the wrong 180 deg connector

of course I am the one that loses out... have to send this back to amazon and order the res from performance-pc's... so now everything gets delayed 3 more days and i'm out a few more bucks

 

[Elganja]

since i had nothing better to do thanks to amazon, i pulled the failed pump apart:

what's interesting is the o-ring looked perfect. I don't see how water could have gotten past it (after 9 months no less). I'm starting to doubt water got past the o-ring...

a quick google search show'd this user with the exact same burn area:
http://www.xtremesystems.org/forums/showthread.php?t=259478

I didn't get a chance to read that thread yet...will have to do some more research tomorrow

 

[aigomorla]

ehhh ive seen much worse... lol

typical of what happens in a very restrictive loop or a very high flow loop.
The pcb heats up, and eventually leads to failure.

Did u try counter sinking the top on? basically do the opposite corners first and then tighten them all down?