Pump RPMs and Pump Life Expectancy

[NFSBuff]

Greetings all,

I'm in the process of constructing my first cooling loop. At the moment, I have the loop setup outside the case for testing, sans radiator. Currently, the testing setup is like this: reservoir -> pump 1 -> pump 2 -> water block 1 -> water block 2 -> water block 3 -> reservoir.

The components are:

• Piping: All pipping and fittings are 3/8ID, 1/2OD
• Loop Turns: There no hard 90-degree turns
• Reservoir: Photon Res 270
• Pump 1 & 2: VPP655 Single Edition
• Pump Tops: Heatkiller D5-TOP
• WB 1: Heatkiller IV Pro
• WB 2: Heatkiller SB Rev3.0
• WB 3: Alphacool HF 14 Universal VGA Block - Copper Edition
• Radiator: Not present in testing loop

My question centers on the pumps speed as measured by the reported RPMs and the longevity of the pumps. To reach the magic number of 1 GPM though the loop, the pumps are currently running @ 4440 RPM each.

So my questions are:

1. Is this a normal RPM range?
2. Will running at this RPM setting continuously shorten the life expectancy of the pumps?

 

[EarthDog]

Welcome!!!

Any reason you are running two pumps in this loop?

Also, Since you are testing outside of your case with the parts not in their final locations, you will have to test all over again...next time, throw it in the final resting place so you don't have to double your efforts.

Unrestricted, that pump at full speed pumps over 6 GPM. With 3 blocks, no rads and TWO pumps, how are you only reaching 1 GPM with the pumps nearly cranked?????

To answer your questions (sorry, LOL!), HERE are the specs for your pump.

Pressure head at 12V: 3,7m
Maximum flow: 1500l/h

Step 1: 1800 rpm
Step 2: 2550 rpm
Step 3: 3300 rpm
Step 4: 4050 rpm
Step 5: 4800 rpm

I also worry about having a single radiator and cooling the CPU, GPU, and, SB (why are you cooling the SB?!!!)... List your hardware so we can tell you what size rad to get...Our general rule of thumb is 100W for every 1x120mm for optimum results.

 

[NFSBuff]

Hi EarthDog, thanks for the welcome.

Before answering your questions, a little background on this build. This is not an ordinary build for a number of reasons. First, the goal here is the quiet cooling of a 3U server. I'd like to note that the merits for and against of doing such a build are not what I'd like this post to turn into. Second, all the cooling components (except the waterblocks, obviously) are located outside the case on a specially cut Acrylic shelf that sits below the case in the rack. I should have listed the radiator I have; it's an Alphacool NexXxoS XT45 480mm, and is mounted on top of the rack. Finally, eagle eyed builders may raise exception at the reservoir to pump relation in terms of being gravity feed. This has already been addressed. The pumps sit in cavities machined out of the acrylic base, putting the inlet below that of the reservoir outlet.

The motherboard is a supermicro server board, and the SB sits at 61 C, so it is included in the loop.

EDIT: Forgot to include what it is I'm cooling. CPU, HBA and SB.

For reference on how all these parts work in concert, see the attached sketch I've hastily thrown together for this post.

 

[EarthDog]

Ahh, yes, thank you for the clarity that makes much more sense. Typically we do not see these builds and since it wasn't mentioned............


I still wonder why two pumps and how your flow rate is so low...a single one of those pumps should push 1GPM+, and likely not even cranked. Perhaps because you have several feet of tubing? No idea... but I run a similar pump, with about 6' of tubing, two blocks, and easily hit 1.5GPM with it set on 3.

 

[NFSBuff]

EarthDog, what you're wondering is why I'm here

For reference, here is the business end of the pumps, and the testing setup for the blocks.

 

[NFSBuff]

EarthDog, thanks for the comparison. I'll try different configurations of the blocks tonight, see if the it's the SB block that's causing the main restriction. (You can see from the pic it's on the smaller side) When all is said and done, I'll have close to 8 feet of tubing, plus blocks, plus rad.

To the original question, if I do end up having to run the pumps at 4600 rpm, is there any data on the pump life expectancy I should be looking forwarding to?

 

[EarthDog]

You should be able to run it full tilt for its warrantied life.

 

[NFSBuff]

So, followup time.

My proposed SB cooling block was definitely the culprit. Removing it from the test loop dropped the pumps' RPM by 1100 while maintaining 1GPM. Close inspection shows the rather horrifically constricted path water has to flow through this block: