Swiftech pump issue?

[FastRedPonyCar]

Split from Multi-loop Madness (http://www.wizdforums.co.uk/showthread.php?t=293) thread.
by Pug

Hmm interesting write up. My initial solution (seeming the simplest) would be to just buy a stronger pump :D

But I suppose that has already been considered by now.

I myself had flow problems (or lack there of) when I added my 2nd rad. If I had a GPU or NB block, I'm sure I'd have been in a world of hurt.

Originally, mine was running

pump -> GPU -> single 80mm rad -> CPU -> T fitting (fill and bleed) then back down to the pump. It worked for like 11 seconds till I discovered a leak at the cpu fitting that burned up the 9700 pro :mad:

So I ditched the GPU block and when I got my 2nd radiator, it wouldn't flow in the following fasion.

Pump -> 80mm rad -> CPU -> T fitting -> dual 80mm rad -> pump

And for some reason, it just could't get any flow :confused: The coolant would just sort of froth at the intake nozzle of the pump. After I re-arranged it to the following it worked though.

Pump -> 80mm rad -> CPU -> Dual 80mm rad -> T fitting -> pump

Any ideas on why it only works this way? The pump is a swiftech pump (not the newest version, the previous version). Was it just that the pump was too weak? Beacause I could blow on the tube at the end of the Tfitting and push coolant througth the whole system.

[mosoto]

The pump is a swiftech pump (not the newest version, the previous version). Was it just that the pump was too weak? Beacause I could blow on the tube at the end of the Tfitting and push coolant througth the whole system.

Have you done the Swiftech replacement impeller modification. If not then your pump may be cactus. It only took a lil' stuffin round with my loop to kill mine.
If you haven't done already, contact Swiftech and they'll send out a new head unit.

[FastRedPonyCar]

I just sent an email to them. My pump is the MCP600 and I think it's dying after only about a year of use. My temps have gone up about 5 or 6*C. And I even lowered the clock speed to try and bring it back down in the 30's. It's around 43*C idle and 48*C load :(

Back in the day, I could have it at 3.3ghz and it'd never go over 41*C under the stress test for prime 95.

So out of curiosity, is it just an improved impeller design or what? What does teh swap entail? Do I just pop open the pump, swap impellers and that's it?

Because honestly, I'm just about to the point of pulling the system and just slapping one of those 120mm zalmans in there. <_<

[Pug]

So out of curiosity, is it just an improved impeller design or what? What does teh swap entail? Do I just pop open the pump, swap impellers and that's it?

If it's anything like my Eheim ones, yes.
http://www.wizarddesigns.co.uk/imgs/13207.jpg

Btw. Welcome to WizD mosoto. B)

[FastRedPonyCar]

BTW, what does "cactus" mean?? I've never heard that expression unless referring to the plant.

[Pug]

Reading the context, I'm substituting "toast" in my mind but no, I'd not heard the turn of phrase either.

[FastRedPonyCar]

Well sounds like they've not heard of it.

Dear Drew:

Temps do not shoot up 5~10 degrees because pump is failing. You need to verify fluid level. if there is a lot of air accumulated in your system it will certainly affect performance. Particularly, if your radiator is setup upside down and at the same level as the reservoir, you could have air accumulated inside the radiator.

I recommend that you bleed your system properly before any other action is taken.

Best Regards,

Gabe

I've already done a fluid change and there's not even the tiniest air bubble. <_<

Neither resivoir are upside down and their systems dont even use radiators. I would think that they would be aware of this? :duh:

I would also think that this guy woudl realize how important the pump is. It's THE most critical part of the system. If the fluid isn't flowing fast enough, there's not going to be good heat transfer from the liquid passing through the radiators!!

Mosoto: Any other information you can provide would be awesome since it looks like Swiftech isn't going to help.

[FastRedPonyCar]

My response

I've done a full fluid flush/refill using a full bottle of HydrX and exactly one liter of de-ionized distilled water. I did the fill and bleed process letting the system run for 20 minutes to clear the system of bubbles. Before I refilled it, I sucked the air out of the system and closed it off via the T fitting and let it sit for 10 minutes. When I opened the T fitting and had my thumb over the end of the fill tube, there was still a full suction on the end of the tube telling me the system was still air tight.

It's the H2O-8600 kit so there's no resivoir for bubbles to get in the way.

I've got two radiators, both have the inlet/outlet openings oriented up.

The pump is at the bottom of the case.

[zer0]

might i be able to see a pic of your setup.

i have a swiftech water block and i dont have any issues with it and i dought you would eather.

also could there be corrosion on the inside of your water block making the heat transfer not very good.

also have you tryed re applying arctic silver on the cpu.

also one moe question. when i installed my water block this time i put hte bracket on the block the wrong way and the block was hitting the row of round silver things witch are positioned aorund my cpu. have you tryed looking at the instructions and making shure the bracket was onthe right way.

anyways good luck.

[FastRedPonyCar]

yeah here u go.

http://img.photobucket.com/albums/v212/fastredponycar/Computer/Watercooling/IMG_3579.jpg

That's how it's configured.

The plastic mounting bracket broke a while back so I've replaced it with a metal one. The temps were still fine after I replaced the bracket and there was a layer of arctic silver between the heatsink and cpu. The bracket is extremely tight so there is good contact between the cpu and the heatsink.

If there's waterblock corrosion, I can't find any debris or stuff in the coolant. It hasn't even been a year since I bought it. I also used de-ionized distilled water to help minimize that risk anyhow.

[Pug]

Hmm. I must admit, I assumed that you'd verified that the pump was failing.
Is it much trouble to drain down and test the pump out of the loop (in the sink or something)?

I must admit, I speed-read your previous post and shouldn't have, sorry.

I'd say you had air in the second rad (I can see some in the pic) which the T-line or res acts as a fluid buffer to counteract.
The rad is one of the hardest places to get all of the air out from, so to have it immediately before the pump is a recipe for disaster.

I've split this off from the thread you posted it in, 'cos I'd say it needs its own.
I doubt it's corrosion either.
Tip the case towards the motherboard side, then tip it up towards the back and see if you get an air rush.

[FastRedPonyCar]

sorry that was an old pic before I re-filled it.

I refilled it last night and let it run for about 20 minutes doing the fill and bleed process. I tipped the case every direction to get bubbles out. I also turned the radiators upside down, sideways, etc in hopes to clear out all of the bubbles.

Temps are the same :(

I got a reply back from my email asking for the pump serial number. I also did some Googling and found that there are numerous websites that test this pump and each one talks about the updated impeller so I guess we'll see what they say.

[Pug]

No prob. Never hurts to double-check. :thumb:
Hope you get a swift result (if you'll excuse the pun). :magic:

[Starbuck3733T]

good luck with the impeller - post the saga. could be useful info for other members. (Even tho a lotta folks around here don't have swifty pumps, the info is still good.)