AIO/Custom Loop Variable Pump Speed question

[For Science!]

2 minutes ago, bowrilla said:

Yes, 6K on Tdie and about 2K coolant (100% fan speed). Which seems ridiculously low especially considering that the blocks are in parallel cutting flow rate roughly in half.

 

This is why I have still some doubts about the results. I might need to do it again and keep Prime95 on for longer and adding the GPU with FurMark for more heat.

Keep us posted, but I think the end conclusion that 25 l/h is still plenty flow, may be the most likely one. My loop in the 'Ncase M1 with the Eisbaer LT pump (its crap), still seems to maintain okayish temps, so I think if you have ANY flow, its pretty much good enough. The flow on the Eisbaer LT is so low that it wouldn't even start rotating my flow meter, so with my instrumentation, it has a flow of 0 l/h  .

[bowrilla]

10 minutes ago, For Science! said:

Keep us posted, but I think the end conclusion that 25 l/h is still plenty flow, may be the most likely one. My loop in the 'Ncase M1 with the Eisbaer LT pump (its crap), still seems to maintain okayish temps, so I think if you have ANY flow, its pretty much good enough. The flow on the Eisbaer LT is so low that it wouldn't even start rotating my flow meter, so with my instrumentation, it has a flow of 0 l/h  .

Well, I got similar results as der8auer. The issue with his dataset is that the flowrate is questionable. Looking at pump revs though he had about the same result. BUT he did not cut flowrate in half through the blocks.That's the thing that's bugging me. 
 

The differences with pump@100% and pump@25% throughout were (100% fan / 50% fan || diff 100% & 25% fan):

Fluid-Probe1: 2,2K / 1,3K || 0,9K

Fluid-Probe2: 2,3K / 1K || 1,3K

Tdie: 6K / 3K || 3K

Die-Avg: 5,8K / 3,4K || 2,4K

System: 1,2K / 0K || 1,2K (obviously air cooled from airflow through radiators)

GPU: 0,2K / 0K || 0,2K (it sits by the way directly in front of the bottom radiator getting lots of direct airflow on the block as well)

 

The differences between fan@100% and fan@50% are also interesting (100% pump / 25% pump || diff 100% & 25% pump):

Fluid-Probe1: 4,1K / 3,2K || 0,9K

Fluid-Probe2: 3,2K / 1,9K || 1,3K

Tdie: 4,1K / 1,6K || 2,5K

Die-Avg: 4K / 1,6K || 2,4K

System: 7,6K / 6,4K || 1,2K

GPU: 2,6K / 2,4K || 0,2K

 

[For Science!]

3 minutes ago, bowrilla said:

Well, I got similar results as der8auer. The issue with his dataset is that the flowrate is questionable. Looking at pump revs though he had about the same result. BUT he did not cut flowrate in half through the blocks.That's the thing that's bugging me. 
 

The differences with pump@100% and pump@25% throughout were (100% fan / 50% fan || diff 100% & 25% fan):

Fluid-Probe1: 2,2K / 1,3K || 0,9K

Fluid-Probe2: 2,3K / 1K || 1,3K

Tdie: 6K / 3K || 3K

Die-Avg: 5,8K / 3,4K || 2,4K

System: 1,2K / 0K || 1,2K (obviously air cooled from airflow through radiators)

GPU: 0,2K / 0K || 0,2K (it sits by the way directly in front of the bottom radiator getting lots of direct airflow on the block as well)

 

The differences between fan@100% and fan@50% are also interesting (100% pump / 25% pump || diff 100% & 25% pump):

Fluid-Probe1: 4,1K / 3,2K || 0,9K

Fluid-Probe2: 3,2K / 1,9K || 1,3K

Tdie: 4,1K / 1,6K || 2,5K

Die-Avg: 4K / 1,6K || 2,4K

System: 7,6K / 6,4K || 1,2K

GPU: 2,6K / 2,4K || 0,2K

 

I think you need to distill and present the data a bit more for it to be accessible, but I understood that as you tighten another bottleneck (radiator-to-ambient heat transfer) by lowering the fan speed, then the flow rate makes even less of a difference. Makes sense, demonstrating that the flow rate is not a very tight bottleneck in the system.

[SolarNova]

My '2 cents'.

 

1GPM is often the suggest 'target' for flow rate. You can get away with 0.5gpm, anything below it isnt ideal, but anything above 1GPM has significant diminishing returns.

 

1GPM = ~227LPH

 

I personally have had my D5 Vario running on settings 5 (max) since i built my loop in ...2013 ..i think, and its still running fine. Obviously I've rebuilt the loop a few times since for upgrades but the pump is the same and I would estimate its had at least 20,000 hours of use.

[bowrilla]

16 minutes ago, SolarNova said:

1GPM is often the suggest 'target' for flow rate. You can get away with 0.5gpm, anything below it isnt ideal, but anything above 1GPM has significant diminishing returns.

 

1GPM = ~227LPH

The question is: where does this general recommendation come from and what's the underlying data? Was it a guess? An educated estimate? Did someone just pull it out of some dark cravass of its own body? Did somebody test it beforehand?

 

Because according to my resutls and the ones from der8auer the differences seem to be marginal for daily use.

[GhostRoadieBL]

1 hour ago, bowrilla said:

This is why I have still some doubts about the results. I might need to do it again and keep Prime95 on for longer and adding the GPU with FurMark for more heat.

 

Edit: I expected to have something more in the 10-15K range on Tdie.

More heat into the loop would saturate faster but we're still talking about a single 140mm rad able to exhaust 200ish watts of heat x4 for the whole radiator setup. 

Adding a gpu worth of watts should show a shorter time to equilibrium in all the tests, if my theory is correct the equilibrium temp should be close to the same until the flowrate can't clear the mass of liquid in the blocks and you get increasing component temps. 

 

[SolarNova]

44 minutes ago, bowrilla said:

The question is: where does this general recommendation come from and what's the underlying data? Was it a guess? An educated estimate? Did someone just pull it out of some dark cravass of its own body? Did somebody test it beforehand?

 

Because according to my resutls and the ones from der8auer the differences seem to be marginal for daily use.

Ah ..good question .. im referencing tests done years back, liek when i built my loop.

 

maybe tomshardware ..ill see if i cant find them.  They were articles, tests. with results and such.

1GPM also tends to be used when testing Radiator performance.

[GhostRoadieBL]

2 hours ago, SolarNova said:

 

maybe tomshardware ..ill see if i cant find them.  They were articles, tests. with results and such.

1GPM also tends to be used when testing Radiator performance.

There is logic in a publication promoting a safe generalized setting for gpm to the average user, we've been playing in the "unique settings for each loop" area of enthusiasts eventually end up. 

 

The closest thing to what we're testing is like undervolting a cpu to increase clockspeed averages, completely breaking down what the common practice is to find a better way. 

 

[AngryBeaver]

16 hours ago, GhostRoadieBL said:

DerBauer's video was extremely close to the test I was looking to run so I can extrapolate most of the conclusion from his data. 

He was specifically looking for "does a lower pump speed reduce the temps" which isn't what I was looking for at all, there is nothing in the physical sciences of closed system cooling which supports that myth. 

However his data does support the theory of not needing 100% pump speed to maintain equilibrium in the system, the "normal" setting at 56% flow rate and "high" setting at 100% flow rate maintains the CPU temperature while greatly reducing the volume and velocity of the water passing through the blocks. A flow rate around 60% seems to be that specific system's sweet spot when at maximum watt load while reducing the wear on the components. 

 

Thanks @Natty Ice and @AngryBeaver for helping me solve this one. After some more testing on block resistance to erosion (so many different fluids) I will have to put together a "how to tune your water cooling system" tutorial to help keep people's systems from unnecessary early wear and tear. 

So rpm imo isn't the best gadget, but instead lpm or gpm. There is a point with flow rate that you start hitting diminishing returns. I have forgotten the magic number for it atm, but that should be the aim instead of a % of pump speed.

 

There are a lot of factors that go into flow rate so rpm wouldn't be a good one size fits all setting.

[GhostRoadieBL]

1 hour ago, AngryBeaver said:

So rpm imo isn't the best gadget, but instead lpm or gpm. There is a point with flow rate that you start hitting diminishing returns. I have forgotten the magic number for it atm, but that should be the aim instead of a % of pump speed.

 

There are a lot of factors that go into flow rate so rpm wouldn't be a good one size fits all setting.

Totally agree! I wish all pumps were measured in lpm instead of rpm as centrifugal pumps never move the same amount of fluid per rotation, unfortunately rpm is the most common method to track flow rate and the only variable we can actually adjust without extra less common devices added to the loop. 

luckily it is close enough for our testing purposes as the flow rate hitting diminishing returns will be the same point as (or very near to) the point where increasing pump rpm no longer provides a benefit to the loop's cooling as the fluid velocity out of the pump starts to resist the lpm of the loop and pressure increases as a result.

 

**edit - what I mean is I wish manufacturers used lpm from a flowmeter installed in the pump to control the rpm rather than leave it to the system operator to try and figure out the most optimal setting.

Edited February 4, 2021 by GhostRoadieBL

[bowrilla]

13 minutes ago, GhostRoadieBL said:

I wish all pumps were measured in lpm instead of rpm as centrifugal pumps never move the same amount of fluid per rotation,

That does not work since the actual lph/gpm depend on the restrictiveness of the loop they are in. All the pumps though have some specs including their max flowrate and headpressure.

[GhostRoadieBL]

1 hour ago, bowrilla said:

That does not work since the actual lph/gpm depend on the restrictiveness of the loop they are in. All the pumps though have some specs including their max flowrate and headpressure.

that is true, but it wouldn't be hard for the high end manufacturers to include a flowmeter after the pump.

Max flowrate works for completely unrestricted pumps but few if any manufacturer would include "flow rate with X CPU block and Y rad" which would provide real use numbers. This is the same issue manufacturers have with fan specs the numbers look great until a fan filter is placed in front of the fan and the lack of pressure is just wasting energy. 

 

**edit - what I mean is I wish manufacturers used lpm from a flowmeter installed in the pump to control the rpm rather than leave it to the system operator to try and figure out the most optimal setting.

Edited February 4, 2021 by GhostRoadieBL