I'm a Watercooling Noob!

[WindowsGeek]

Title related. I am hunting for a high-pressure water pump. GPH does not matter too much, I just need pressure. My system is 3/8in ID (10mm ID). I have a Koolance waterblock that has very low resistance, however my radiator has very high resistance... it's a '62 Chevy transmission fluid cooler. No, I am not going to buy a different radiator. I am wondering if anyone has any good recommendations pump-wise for high resistance systems. Thanks!

 

[clifford_cooley]

Check out these pumps:

1. Laing D5 Revision 2 Variable Speed Water Pump
2. Swiftech MCP655 12 VDC Pump Liquid Cooling System
3. Koolance PMP-450 12V Variable Speed Pump

Edit:
Another thought would be to configure two pumps in series. pumps work better when they don't have to pull as hard on the inlet side. The first pump would allow the second pump to work at its full potential. After all this is the same idea behind dual-stage on an air compressor or turbo on a car.

 

[WindowsGeek]

Thanks Cliff those are certainly better than the fountain pump I have now. However, the barbs are too big. I have a 3/8in system... now to find 1/2 to 3/8in adapters

 

[Kougar]

Check out these pumps:

1. Laing D5 Revision 2 Variable Speed Water Pump
2. Swiftech MCP655 12 VDC Pump Liquid Cooling System
3. Koolance PMP-450 12V Variable Speed Pump

Okay, firstly the Laing D5 is the exact same pump as the MCP655, Swiftech rebadges them. Second, this pump is best for high GPH and low pressure... It's been awhile but I do recall that the Swiftech MCP355 was better suited for high-pressure pumping, but it has a lower GPH rating than the larger 655.

I'm not familiar with the Koolance pump listed, nor many others for that matter... there are some custom fabbed tops for the MCP355 (and 655) that will increase flow or pressure ratings too, as yet another option. The most extreme users will recommend Iwaki pumps, these are industrial grade units and can deliver any sort of pressure rating you're looking for, but they come with a matching price.

After a certain point it'd become cheaper just to get a standardized computer radiator though, so I'm not sure how far you want to go with it.

 

[clifford_cooley]

Okay, firstly the Laing D5 is the exact same pump as the MCP655, Swiftech rebadges them.

I knew they looked familiar.

 

[Nibiru2012]

Cliff - how is your new water cooler setup working? I am curious and would you post the temperatures, if possible?

Thanks,
~Nibs

 

[clifford_cooley]

how is your new water cooler setup working?

It's not a WC setup, I purchased the same cooler as yodap. The ZALMAN CNPS10X Performa is the cooler I decided on.

 

[WindowsGeek]

I found that one out after long hours and much lost sleep looking at water pumps that is a joke, but yes they are all restamped Laing D5 and D5-2 pumps. I ordered a Laing DDC pump (rebranded by Swiftech as the MCP355) because it had the highest PSI. They were also cheap, $40 got me one. It seems to have enough pressure for my Chevy radiator, so I am happy. I'm about to install my waterblock now

 

[WindowsGeek]

Okay. I have the waterblock installed. However, I am wondering why my CPU is running at 45*C idle and 55+ under 10 minutes of load... this does not seem right, my old heatsink (with 4? heatpipes) ran cooler. It's not massive like the one linked earlier in the thread, it is more like an AMD Phenom II stock. Well, it was one. Is this because the thermal paste needs to burn in? If yes, I have never seen performance this bad coming out of un-baked thermal paste.

My wc setup is Tank->Pump->CPU->Radiator->Tank and the coolant I am using is antifreeze. I am not using specialized coolant for a CPU because that stuff is expensive. Instead I am using AutoZone's car coolant. Is this a bad idea?


-EDIT- I peaked at 68 load and then brought my system back to idle. 68C was about 15 minutes of playing Minecraft.

 

[Nibiru2012]

IF you are using pure antifreeze instead of a mixture of 50-50 water to antifreeze it won't cool as good. The water is needed when you use an antifreeze because antifreeze is much more viscous than water and holds the heat. The use of antifreeze is done purely for it's anti-corrosion properties.

How smooth is the mating surface of the CPU waterblock? If it has some machining marks on it then it won't mate as well as one that has a mirror-smooth finish.

If you used the Arctic Silver products, then they do require a break-in period; usually about 100 hours or so. The newer nanotherm compounds are better in that they don't require a break-in period and are usually superior to AS products.

If you use just straight distilled water you'll probably get the best cooling performance, plus there's no minerals to cause any issues with the waterblock or radiator.

 

[WindowsGeek]

Thanks Nibs! I was using 50/50 antifreeze and water, and still am. The block had a mirror finish. I also cleaned the CPU's thermal dissipator, applying fresh heat paste. The paste I used was Koolance's "brand" that came with the block. Probably not as good as Arctic Silver, but I was out of that and I didn't want to go to Wilsonville and get more

I was also about to use just straight water, but I went with the antifreeze coolant mixture because I was worried about corrosion in the radiator, as it's intended use is not watercooling a PC. I am not even sure what it is made of, probably some aluminum alloy. I have also heard about some metals interacting with each other in water systems, causing faster corrosion or "thermal barrier crusts" on your waterblock.

Anyways I left the computer on last night, hoping to burn in the thermal paste. I am also using a different program for measuring temperatures, but there is a notable difference.

NEW TEMPS​

With a room temp of 22*C, idle temp is 25-27*C and load temp is 38*C. Not bad! I am now getting much better temps than I could have dreamed of on the old air cooler. The only fans I have running on the radiator are two 80mm case fans.

--TIP-- Don't use SpeedFan for monitoring temps. It is inaccurate in reading temps, at least it was for me. Instead I used CoreTemp, it seems to be much more accurate in reading temps. I judged accuracy by rebooting and looking at the BIOS's "PC Health Status" section. The radiator is also completely passive at the moment, all fans are stopped. SILENCE IS BEAUTIFUL.

 

[Kougar]

Alright, sounds like ya got things well in hand. Just be careful to factor your room temp into your readings, it will significantly change them.

Antifreeze is generally not recommended or even used by watercoolers... as Nibiru said it's only purpose is for anti-corrosion and this isn't an issue unless you're mixing metals in your loop. But I tend to agree with you, the heater core is probably aluminum.. Most WC equipment use either copper or nickel plated blocks / rads for this reason. Aluminum and steel will both cause galvanic reactions pretty easily when mixed with these metals. There are some studies on XtremeSystems.org where any anti-freeze in a loop decreased performance... the benchmark standard is still distilled water and a few drops of algaecide to prevent growth.

Even if the paste needed to "cure", you wouldn't gain anything more than a 5c reduction in temps. Paste is still paste... the only thing I can suggest is made sure you use as little as possible, because paste is in of itself is still insulator... too much of it will drastically decrease the heat exchange process. I'm going to guess that it took awhile for that pump to work most of the air pockets & bubbles out of that radiator and that was behind the performance more than anything.

When ya look at the BIOS's temp readings, keep in mind the BIOS doesn't support all of the powersaving features for the CPU... the CPU should give you lower temps within Windows than in the BIOS. Not all BIOS's use the on-die temp sensors either, most default to the sensor underneath the socket. CoreTemp and RealTemp will only use the on-die sensors, while (for probably the majority of boards) SpeedFan will give both.

38C is a much better temp, sounds sweet. Hope ya enjoy your new setup!