Pump not running

[bikeracer4487]

So, I recently upgraded my motherboard and CPU, and in the process, altered my water cooling loop. For some reason I seem to be having trouble getting the pump to run and I get intermittent CPU fan error warnings on POST. I have an Asus P9X79 Deluxe mobo, and my loop consists of an XSPC dual-bay reservoid to a MCP35X pump, to a Swiftech Apogee HD CPU block to a 3x120mm radiator and back to the reservoir. The pump has a 4-pin molex connector and a 4-pin fan connector that's plugged into the motherboard. Even when I was just test running the loop, I would notice that sometimes the pump wouldn't be running and sometimes just shutting it down and starting it back up would enough to get it running. Anyways, this is really frustrating, and I'd appreciate any advice.

P.S. PSU is a Corsair AX1200 (1200W). Rest of my specs in the sig.

 

[toolmaker_03]

it sounds like the motherboard is your problem, this is one of the reasons the I do not like the pumps designed to connect to the motherboards fan header. the fan header on all motherboards are current restrictive, obviously some more than others. but still, this is your problem. if you change the power to the pump from the fan header to a Molex connector for power, this should solve your problem. you can still use the PWM signal and RPM read wires from the fan header, just change where the power is coming from.

 

[bikeracer4487]

I actually just took a look at the documentation for the pump, and it states that if the 4-pin fan head is not connected to the motherboard, than the pump will default to its maximum speed. Currently, I have both the molex and the header connected, so I will try using just the molex to see if this at least resolves the problem of the pump not running at all...

 

[toolmaker_03]

you should never connect two different powers to the same load (EX: your pump) this can seriously damage your motherboard.

 

[bikeracer4487]

you should never connect two different powers to the same load (EX: your pump) this can seriously damage your motherboard.

Wait, what?! Frick... I just assumed the pump needed more power than the mobo could give it (a la most graphics cards), so the molex would supply extra power. Are you positive about that, because if that's true, then it's a major oversight on Swiftech's part because the document is worded as such that it sounds like the molex is REQUIRED and the fan header CAN be used if you want to the mobo to be able to control fan speed.

 

[boiler1990]

The 4 pin header just provide PWM information; the pump gets all of its power via the molex connection (like all other pumps).

 

[toolmaker_03]

well that depends, are there 2 wires or 4 wires going to the fan header, if 4 wires are going to the fan header than this is bad if you connect to a Molex as well, I have seen both configurations before. if only 2 wires are going to the fan header, than yes, the pump gets its power from the Molex.

 

[bikeracer4487]

Well, there are only 2 wires, so it should only be for control. That said, I unplugged the fan header when I got home and the pump worked fine the rest of the night. However, at some point during the night or in the morning, the pump STOPPED working, and I can't seem to get it going again. I've tried piggybacking the molex connector on top of fan connectors, attaching it to it's own molex connector straight from the PSU, using a molex connector cable that didn't have any other accessories attached to it, and I've tried using it with and without the mobo fan header plugged in...all to no avail. Not sure what to do...

 

[toolmaker_03]

its either the pump is dead or the power supply is dead

 

[bikeracer4487]

Fabulous...well the power supply's fine. Everything else is receiving power properly and I actually plugged the pump into a molex connector after using a fan to confirm that that connector was working... Well frick, this thing's only a month old, and changing up the loop is SUCH a pain in my ass. Any chance it could be bad contacts, or something I could tweak with a screwdriver?

 

[toolmaker_03]

no the pump type sucks, try the MPC655 or something similar to it, I get a solid 5 years out of them, and keep them running 24/7.

 

[boiler1990]

The MCP35X is one of the best pumps out there, so I suspect it just had some faulty electronics. I'd imagine if it's only month old that it'd be under warranty.

 

[rubix_1011]

MCP35x is pretty close to an equivalent to a D5 as you can come in terms of flow rate, with the added benefit of the DDC head pressure when considering Laing pumps.

If you want better, you need to start thinking in terms of Iwaki and the like.

 

[bikeracer4487]

So, I get home and decide to give it one last go, so I plug it back in, start up my computer...and nothing...UNTIL I reach in and just push down on the pump and it suddenly starts working?! Then, I turned my computer off again and when I power back on, it's back to being dead. So I took out the pump (leaving tubes still attached) and turned on the computer while I held the pump in my hands. It felt like a car trying to turn over...as I was rotating it around a bit I could feel it try to start every 3-4 seconds or so...it would just vibrate for a brief second and then stop. Eventually it "caught" and started up, and I was able to replace it back in the case and it kept running, but now I'm afraid to turn my computer off for fear it will die again, but so long as I just stick to restarts, I'm all right... So, any ideas what's going on?

 

[rubix_1011]

You have a loose electrical connection.

It's likely that the vibration from the pump actually working is keeping the wiring in contact, but there is a good chance it might not always continue. Let's hope you have a shutdown script to execute if your pump shuts off and your CPU temps reach a threshold to trigger it.

 

[bikeracer4487]

You have a loose electrical connection.

It's likely that the vibration from the pump actually working is keeping the wiring in contact, but there is a good chance it might not always continue. Let's hope you have a shutdown script to execute if your pump shuts off and your CPU temps reach a threshold to trigger it.

Good idea. I've got Core Temp running, so I'm monitoring the temps, but I just set up a script to shut down if temps hit 60C, which should be plenty conservative since even in Prime95 I'm only hitting 40-43C (when the pump's RUNNING of course...)

 

[rubix_1011]

Just make sure it is setup to run at startup/login and you'll be set.

Just make sure it's set to trigger the script correctly- you could test with a dummy script that outputs a window text or something similar, as well as testing the shutdown script separately.

 

[seema123]

I have same problem and i have got the solution from that forum.Thanks for all the responses , seems fine now after just pushing the wires into the pump slightly (where they come out of). Been running the system for 5-6 hours and all seems to be great. Won't be photo worthy at all til I clean up all these cables.

 

[bikeracer4487]

Well I had to shut down sooner or later, right? When I did, and then turned it back on, I couldn't get the pump to "turn over" no matter how hard I tried, even though I could feel a vibration every 5 seconds or so. Anyways, Swiftech is RMA'ing them pump now (still under warranty) and I had a spare that was actually surprisingly painless to swap in (much easier than, say, a radiator or CPU block) so I'm up and running for now, until I get the MCP35X back, thanks for your help!

P.S. I'm now paranoid about pumps dying on me, so I have Core Temp running in the sys tray 24/7 and I figure that shut down script is a good thing to have regardless.

 

[rubix_1011]

I've never had a pump die on me...so this is odd.

 

[bikeracer4487]

Best answer selected by bikeracer4487.

 

[rubix_1011]

This topic has been closed by Rubix_1011