Does pump speed matter?

If you have an unused port on your reservoir that comes in contact with the coolant, you can use this:
1619884655804.png

Alphacool 17364 Eiszapfen Temperature Sensor Plug​


This is what I am using to get my coolant readings and you won't need any other adapters to get it to work. And since the reservoir is your starting point of your coolant after it runs through all the radiators, that might be the best place to get your reading.
 
If you have an unused port on your reservoir that comes in contact with the coolant, you can use this:
453934_1619884655804.png

Alphacool 17364 Eiszapfen Temperature​

I already have one of those alphacool, it's installed as you mentioned in your post.
 
Because i would like to know/monitor the water temp going into the rad
The sensor can be fastened towards the radiator though. I have that sensor in black and have mounted it between the fitting and the radiator. Saves you two fittings and you avoid having 150 or so grams suspended on the tubing.
 
Too much work just for a little extra gain
20210502_130114.jpg
20210502_131329.jpg

Anyway, i have noticed that the water in the reservoir move way faster than before and so i have some doubts that maybe the Flow meter isn't giving the right numbers or maybe could be faulty.
 
This probably shows that your radiators are probably your biggest restrictor, but you did see improvements. Also, keep in mind, as cool as you are running already, you are reaching a point of diminishing returns. You still managed to get your coolant temp to drop .8C, your flow rate has improved by .3 L/min, and your flow flywheel (I presume) has sped up by 6 rotations a second.

Is your pump running at the same speed as before? Is it at max speed or slower?

I'd be curious to see how much the flow rate on the meter changes with different pump speeds to maybe help determine if the meter is displaying accurate numbers or not.

Lastly, while it probably doesn't create a huge impact, flow meters do restrict flow rate as well
 
Thinking to put it back as it was
It's ultimately your choice.

Me, personally, I think it looks better and less cluttered not using the distribution block. With the way it was laid out, I fail to see what benefit the distribution block was giving you other than more potential leak points, slightly more coolant capacity, a more cluttered look, a lower flow rate, and more work for you.

They say beauty is in the eye of the beholder, so maybe you like the way it looks or feel like its a waste having it if you don't use it. But...is there any benefit to adding it back into the loop when it comes to your current setup?
 
is there any benefit to adding it back into the loop when it comes to your current setup?
Not at all.
The only little gain as you mentioned slightly more coolant.
At the end i leave it as it's.
Have a question for you.
Is it worth to add a second pump?

Edit: you forgot to give me your opinion about this: I'd be curious to see how much the flow rate on the meter changes with different pump speeds to maybe help determine if the meter is displaying accurate numbers or not.

I posted some pictures for you.
 
Last edited:
Not at all.
The only little gain as you mentioned slightly more coolant.
At the end i leave it as it's.
Have a question for you.
Is it worth to add a second pump?

Edit: you forgot to give me your opinion about this: I'd be curious to see how much the flow rate on the meter changes with different pump speeds to maybe help determine if the meter is displaying accurate numbers or not.

I posted some pictures for you.
Your flow rate is fine. Go add another pump if you want to waste more money chasing diminishing returns.
 
Your flow rate is fine. Go add another pump if you want to waste more money chasing diminishing returns.
Was just a question, curiosity :D
I am not willing to pay 130€ for another pump+res combo.
I am done messing with this Build, only thing i will do is to re-apply the thermal paste, then i am done for good
 
Was just a question, curiosity :D
I am not willing to pay 130€ for another pump+res combo.
I am done messing with this Build, only thing i will do is to re-apply the thermal paste, then i am done for good
I know the feeling trying to get everything right. I did the same crap with my first wc loop. It took a lot of wasted money before I realized wtf am I doing. I spend more time dicking around with the loop then enjoying the system.
 
It took a lot of wasted money before I realized wtf am I doing. I spend more time dicking around with the loop then enjoying the system.
Exactly what i have been doing lately.
I have built lot of Watercooling system and never had issues but this time things didn't go as planned:banghead: We're not perfect and mistake happens.
This build is just 5 months old.
I have been thinking for future Watercooling system I'll use tubing with bigger diameter.
 
Not at all.
The only little gain as you mentioned slightly more coolant.
At the end i leave it as it's.
Have a question for you.
Is it worth to add a second pump?

Edit: you forgot to give me your opinion about this: I'd be curious to see how much the flow rate on the meter changes with different pump speeds to maybe help determine if the meter is displaying accurate numbers or not.

I posted some pictures for you.
I can't say whether or not the flow rates are accurate, but we are seeing a consistent change in flow rate based on the revolutions per second it is recording, which is roughly between .043-.045 liters/min for every 1 revolution per second. Your coolant temps, which I assume showing are at idle, don't seem to change much unless the pump is running at max speed. Which even then it only drops by .7-.8c, not a huge difference. Where it will be interesting to see is what your max coolant temps (as well as peak CPU and GPU temps) end up at varying speeds when under an extended load, since that will be more important. That's what you should be chasing. If you find the difference between the lower flow rates and the higher flow rates is minimal, then you may be best to set your pump at a slower speed instead of eating up more power, creating more noise, creating more heat (from the pump), and shortening the life of your pump.

What we can confirm is that you do have consistent movement of coolant. Increasing pump speed usually doesn't provide as much of a cooling benefit than increasing fans speeds or adding more fans to the radiators.
 
Just a reminder that with modern blocks higher pressure helps since water channels are very small. Water flow and water pressure are not the same thing. I'm running three pumps: 2x D5 Serial and a little giant aquarium pump. Each setup alone provides more than 3 lpm of flow. Together they only reach 4.9 lpm. However, the little giant provides MUCH better cooling than the D5s as they primarily serve to increase the flow through my high-flow radiators after the pressure drop of my high-restriction waterblocks. Generally speaking the location of various components doesn't make any difference however the one thing that does change in a closed loop is the pressure head through restrictions. My setup that drops my water temp the quickest and keeps the components cooler is reservoir -> LG pump -> optimus gpu block -> optimus cpu block -> RAM water block -> serial D5s -> 2x 480 monstas -> EK monoblock (second system) -> MORA3 420 -> (back to reservoir). Obviously I have long runs of tubing and many more blocks and quick disconnects than most but it was interesting to see how different pump combos and locations ultimately changed the performance of my loop.
 
Just a reminder that with modern blocks higher pressure helps since water channels are very small. Water flow and water pressure are not the same thing. I'm running three pumps: 2x D5 Serial and a little giant aquarium pump. Each setup alone provides more than 3 lpm of flow. Together they only reach 4.9 lpm. However, the little giant provides MUCH better cooling than the D5s as they primarily serve to increase the flow through my high-flow radiators after the pressure drop of my high-restriction waterblocks. Generally speaking the location of various components doesn't make any difference however the one thing that does change in a closed loop is the pressure head through restrictions. My setup that drops my water temp the quickest and keeps the components cooler is reservoir -> LG pump -> optimus gpu block -> optimus cpu block -> RAM water block -> serial D5s -> 2x 480 monstas -> EK monoblock (second system) -> MORA3 420 -> (back to reservoir). Obviously I have long runs of tubing and many more blocks and quick disconnects than most but it was interesting to see how different pump combos and locations ultimately changed the performance of my loop.
For me, my loop is pretty simple. For me I have my D5 pump -> Optimus CPU Block (~20" tube length) -> Corsair XR7 240mm Radiator (~9" tube length) -> back to reservoir (~9-10" tube length). The tubing used is Primochill 3/8" ID 1/2" OD. While the case is a mATX cube, so pretty small, it's a two chamber design. My pump/reservoir and radiator are on the back side, hence the longer tubing needed than would be expected. But it allows me to do smooth, wide curves with the tubing so as to not kink it nor the need for using angled adapters midway through the run.

As a curiosity, I decided to measure coolant temp, CPU temp, and CPU package temp while running FurMark CPU Burner across all threads at varying pump speeds to see what the difference was. Here are my observations:

I started my test at my current configuration, 50% (~2,450-~2,500 RPMs) and let CPU Burner run until coolant temp and CPU temps leveled out. I then increased it in 10% increments while CPU Burner was running, waited for temps to level out then recorded them. I continued to do this until I got up to 90%. I then backed it back down to 50% then ramped it down to it's minimum speed (~32%, ~1,500 RPMs).

  • 50%, 1,450-1,500 RPMs (Silent):
    • Coolant Temp: 33.5C
    • CPU Temp: 59C
    • CPU Package Temp: 70C
  • 60%, 3,000 RPMs (Hint of pump noise):
    • Coolant Temp: 33.8C
    • CPU Temp: 59C
    • CPU Package Temp: 70C
  • 70%, 3,550-3,600 RPMs (Noisy):
    • Coolant Temp: 33.8C
    • CPU Temp: 59C
    • CPU Package Temp: 70C
  • 80%, 4,100 RPMs (Silent):
    • Coolant Temp: 33.9C
    • CPU Temp: 59C
    • CPU Package Temp: 70C
  • 90%, 4,650-4,700 RPMs (Hint of pump noise, vibrations):
    • Coolant Temp: 34.1C
    • CPU Temp: 59C
    • CPU Package Temp: 70C
  • 50%, 1,450-1,500 RPMs (Silent):
    • Coolant Temp: 33.7C
    • CPU Temp: 59C (brief jump to 60C but settled back down to 59C)
    • CPU Package Temp: 70C (brief jump to 71C but settled back down to 70C)
  • 40%, 1,900 RPMs (Hint of pump noise):
    • Coolant Temp: 33.2C
    • CPU Temp: 59C (1-2 minute jump to 60C but settled back down to 59C)
    • CPU Package Temp: 70C (1-2 minute jump to 71C but settled back down to 70C)
  • 32%, 1,500 RPMs (Silent):
    • Coolant Temp: 33.1C
    • CPU Temp: 60C
    • CPU Package Temp: 71C
  • 50%, 1,450-1,500 RPMs (Silent):
    • Coolant Temp: 33.5C
    • CPU Temp: 59C
    • CPU Package Temp: 70C
Interesting observation is that the temp difference between CPU and CPU Package when under load is around 10-11C. At idle, coolant temp will sit between 28C and 30.5C depending on the time of day and if there is any background tasks. CPU and CPU Package temps will typically be in low to upper 30s, again depending on time of day. The temp difference between CPU and CPU package at idle is much smaller than when under load, varying between 3-5C.

Based on these observations, I don't see any advantage with my setup to run the pump above 50% in most circumstances. It didn't seem to make any difference in CPU temps. All I noticed was an increase in coolant temps and more noise/vibration from the pump. And same goes in the opposite spectrum, I saw no benefit going below 50% pump speed. Coolant temps will drop a little, but CPU temps wouldn't really change and any brief spikes (albeit only 1C noticeable) takes a little longer to settle back down to the same temp achievable at higher pump speeds at the expense of more noise. When dropping to the bare minimum pump speed, that was the only time I noticed temps increase (albeit just 1C) and not go back down until pump speeds were increased.

When it comes to a balance of noise, vibration, and cooling capability, 50% pump speed looks to really be the sweet spot for me.
 
Last edited:
My system behaves similar, except for the vibrations as I use the soft rubber isolation VirtualMirage's system (takes away pretty much all vibrations). Up to 50% the pump is almost silent, at 70% percent there is low hum and at 100% it is fairly audible and increase in noise is linear due to lack of vibrations in the case. I probably have slightly higher resistance loop due to larger radiator and more tubing. With an ek-velocity block in an AMD 5900x the difference between 50% pump and 100% pump speed is less than 1 degree. The large difference are from 20% to 50% where the difference in temp is around 1 degree less per 10% pump speed increase. E.g. going from 20% to 30% will improve temps by slightly above 1 degree and going from 40 to 50 is close to 1 degree lower temp. After 50 % there are very small gains. I don't have a flow meter, but there is a slight amount of movement in the res around 35% and quite noticable movement at 50%.
 
My system behaves similar, except for the vibrations as I use the soft rubber isolation VirtualMirage's system (takes away pretty much all vibrations). Up to 50% the pump is almost silent, at 70% percent there is low hum and at 100% it is fairly audible and increase in noise is linear due to lack of vibrations in the case. I probably have slightly higher resistance loop due to larger radiator and more tubing. With an ek-velocity block in an AMD 5900x the difference between 50% pump and 100% pump speed is less than 1 degree. The large difference are from 20% to 50% where the difference in temp is around 1 degree less per 10% pump speed increase. E.g. going from 20% to 30% will improve temps by slightly above 1 degree and going from 40 to 50 is close to 1 degree lower temp. After 50 % there are very small gains. I don't have a flow meter, but there is a slight amount of movement in the res around 35% and quite noticable movement at 50%.
There is only so much dampening I am able to do for mine. It just seems like at certain speeds the pump hits a resonating frequency that transmits through the case. But the way my pump is mounted probably doesn't help either. I also never go above 90% with the pump speed because it will draw from the reservoir too fast, disturbing the very small air pocket that resides in there, breaking it up into micro bubbles, and sucking them through the loop. And that's even when using the provided tube insert they recommend using for such situations.

Due to my small case, my pump/reservoir is mounted horizontally. It is hanging off the underside of my radiator and its fans via EKWB's fan mount. This is an older picture from when I was first building it out, but its layout hasn't changed:
386873_Second_Pic_In_Progress.jpg
 
There is only so much dampening I am able to do for mine. It just seems like at certain speeds the pump hits a resonating frequency that transmits through the case. But the way my pump is mounted probably doesn't help either. I also never go above 90% with the pump speed because it will draw from the reservoir too fast, disturbing the very small air pocket that resides in there, breaking it up into micro bubbles, and sucking them through the loop. And that's even when using the provided tube insert they recommend using for such situations.

Due to my small case, my pump/reservoir is mounted horizontally. It is hanging off the underside of my radiator and its fans via EKWB's fan mount. This is an older picture from when I was first building it out, but its layout hasn't changed:
View attachment 354011
We have the same mounting mechanism and I had the same problem as you when using the harder damper that came with the pump/res combo, but the soft damper works great. E.g. 50% was painfull to use with the hard damper as the entire case resonated while 40% and 60% were low noise. The soft damper, that was also supplied, can only be used when vertically mounted so it would be a no go in your case. It does make a world of difference though as there are virtually no vibrations in the radiator even though I have my pump attached to it and none of the pump speeds cause any vibrations anymore.
 
We have the same mounting mechanism and I had the same problem as you when using the harder damper that came with the pump/res combo, but the soft damper works great. E.g. 50% was painfull to use with the hard damper as the entire case resonated while 40% and 60% were low noise. The soft damper, that was also supplied, can only be used when vertically mounted so it would be a no go in your case. It does make a world of difference though as there are virtually no vibrations in the radiator even though I have my pump attached to it and none of the pump speeds cause any vibrations anymore.
It was mainly at just 70% and 90% (interesting 80% was quiet and vibration free) and maybe an ever slight hint at 40%. So it isn't a huge issue since I am usually not at those speeds. The other speeds were pretty much silent and/or vibration free.
 
I haven't noticed any vibration noise with my D5 mounted to the fans with the EK mount even at 100%. At 40% I don't hear the pump at all. I can't stand the whine at anything above 40%. I got a pretty restrictive loop with 3 rads and 2 blocks. Running it 100% only lowered my temps by 2c at most.
 
That has generally been my experience with a few DDC pumps . There may be a 1-2ish c improvement at 100% when compared to 50%.

I only run 100% for the first couple days just for bleeding purposes.
 
I haven't noticed any vibration noise with my D5 mounted to the fans with the EK mount even at 100%. At 40% I don't hear the pump at all. I can't stand the whine at anything above 40%. I got a pretty restrictive loop with 3 rads and 2 blocks. Running it 100% only lowered my temps by 2c at most.
My D5 is virtually silent up to about 55% and low noise up to about 75%. With the case vibrations the max it was almost silent at was around 35-40%.
 
My second DDC EK 3.25 died a few days ago. Was doing the normal not wanting to start on cold boots, unless you tap it, hot start it a few times. Same issue I had with the last EK branded one I had to RMA.

I am seriously thinking of switching to a D5.

Guess I'll be able to see if the head pressure difference makes a difference in my system.

I haven't had that style of pump since the OG Danger Den DD12v-D4.
 
Hello everyone.
So, i have a temperature sensor which is plugged to one of the pump ports and also have a flow meter, i have noticed that the temps on both devices it doesn't match, when the temperature sensor reads let's say 30c water temp the flow meter reads lower temps something like 29.1c/29.2c or 29.3c, now i have a question: which temps should be more accurate?
The flow meter is installed right after the pump and so it wouldn't makes any sense to think that the temperature difference is because the temperature sensor is plugged to the pump.
 
Last edited:
Hello everyone.
So, i have a temperature sensor which is plugged to one of the pump ports and also have a flow meter, i have noticed that the temps on both devices it doesn't match, when the temperature sensor reads let's say 30c water temp the flow meter reads lower temps something like 29.1c/29.2c or 29.3c, now i have a question: which temps should be more accurate?
The flow meter is installed right after the pump and so it wouldn't makes any sense to think that the temperature difference is because the temperature sensor is plugged to the pump.
The temp sensors aren't scientific grade, so they will have a fair amount of deviation in their specs. E.g. one of my temp sensors has an inaccuracy of +/- 0.3 at 28 degrees, which means that worst case you could get 0.6 degrees between a sensor showing high temps and another showing low temps in the same location. I mostly use my sensors as an indication and for controlling fan speed (radiator based on water and case fans based on exhaust temp). Temps should be marginally higher after the pump than in the res as long as the entire system is in stable state. If you read temps right after you started a thermal load and one of your sensors are at the inlet, then the outlet could be getting cooler water until the water in the res has been cycled/mixed with the hotter water. It could of course be high deviation in your sensors as well.
 
Last edited:
Actually with a 1000 lph pump the coolant wouldn't be spending enough time in the water block to absorb very much heat and also wouldn't be spending enough time in the rad's to transfer much heat. Most of the popular pumps today range anywhere from 7 lpm to 23 lpm with a head height of anywhere from 5 to 9 meters.

EDIT: Sorry, my bust on the flow rate Ya it's 7 to 23 Liters Per Minute not per hour.
that makes zero sense.
 
There is very little difference from what I have seen.

My findings. RTX 3090 while mining, 5950x idle, 2 x 360 rads.

With the pump at 30-35% (D5) GPU:~42-43c , cpu idle ~40c. coolant 33-34c.
With the pump at 100% GPU: 40-41c, cpu idle ~40c, collant 33-34c.

So shaved off at best 1-2c from the GPU with a mining load (gaming would be a better test, but I doubt it would make a huge diff)
 
Hello everyone.
So, i have a temperature sensor which is plugged to one of the pump ports and also have a flow meter, i have noticed that the temps on both devices it doesn't match, when the temperature sensor reads let's say 30c water temp the flow meter reads lower temps something like 29.1c/29.2c or 29.3c, now i have a question: which temps should be more accurate?
The flow meter is installed right after the pump and so it wouldn't makes any sense to think that the temperature difference is because the temperature sensor is plugged to the pump.
There’s a margin of error.

if its any consolation my water temp is at around 38 to 39c using the same bykski flow meter/temp monitor and a d5. Thats with a 6900xt, rx480 and 5900x mining/gaming/farming chia. Flow rate is approx 3.2lpm. Rads are a 420 and 240.
 
Too much work just for a little extra gain
View attachment 352459
View attachment 352458

Anyway, i have noticed that the water in the reservoir move way faster than before and so i have some doubts that maybe the Flow meter isn't giving the right numbers or maybe could be faulty.

I have the same flow meter and it has confused me with the flow readings. They are roughly 10% of what they should be using a D5 pump. My EK pump at 3800RPM reads a flow rate of 2.0 L/min when it should be 20 L/min with no restrictions. I have only 2 rads and a CPU block so I would think it would be at least 15 or so. This flow meter is also supposed to work or 12V but when I got mine the Molex plug is wired to the 5V side of the plug. My loop is reading 22.1C for coolant temp on the device and 23.2C on the coolant temp sensor attached to a pump port so flow seems pretty darn good, ambient temp using a thermal probe is 19C average (measured at the air intake area of the two rads).
 

Attachments

  • Photo May 13, 11 16 49 AM.jpg
    Photo May 13, 11 16 49 AM.jpg
    440.1 KB · Views: 0
  • Photo May 13, 11 18 19 AM.jpg
    Photo May 13, 11 18 19 AM.jpg
    436 KB · Views: 0
Last edited:
Hello everyone.
So, i have a temperature sensor which is plugged to one of the pump ports and also have a flow meter, i have noticed that the temps on both devices it doesn't match, when the temperature sensor reads let's say 30c water temp the flow meter reads lower temps something like 29.1c/29.2c or 29.3c, now i have a question: which temps should be more accurate?
The flow meter is installed right after the pump and so it wouldn't makes any sense to think that the temperature difference is because the temperature sensor is plugged to the pump.

My temp sensor on my flow port is about 1C higher than the flow meter. I just figure since the sensor is right on the pump that maybe the pump body is causing slight temp increase, my flow meter is after the coolant has passed through one of my 360 rads and on its way to the CPU block.
 
I have the same flow meter and it has confused me with the flow readings. They are roughly 10% of what they should be using a D5 pump. My EK pump at 3800RPM reads a flow rate of 2.0 L/min when it should be 20 L/min with no restrictions. I have only 2 rads and a CPU block so I would think it would be at least 15 or so. This flow meter is also supposed to work or 12V but when I got mine the Molex plug is wired to the 5V side of the plug. My loop is reading 22.1C for coolant temp on the device and 23.2C on the coolant temp sensor attached to a pump port so flow seems pretty darn good, ambient temp using a thermal probe is 19C average (measured at the air intake area of the two rads).

What waterblock are you using?

You're not going to get anywhere near 20L/min in an actual loop. This is with a xspc raystorm pro, bykski block on a 6900xt, maze 4 gpu block on a rx480 and two rads.

pump.jpg
 
Last edited:
Couldn't find anything regarding how restrictive this block is, how are your temps? This is a brand new loop?

Temps are great. Does not break the 70 C mark unless I am doing something like running benchmarks. Coolant sticks around 22-23C all day long. Although I have not added my 2080 Super to the loop yet and that is just the 5900X running pretty much unleashed at I have the EDC limit set to 200A and using Curve Optimizer to get peak boost on all cores. I do not normally run HWInfo64 all day so as of right now the peak temp on CCD1 has been 59C and CCD2 has been 51.8. I have not gamed at all yet today though but from the last time I ran HWInfo and gamed it may have topped out at 66C. Loop is new as of a few weeks or so ago, I changed cases about a month ago and did a temp loop with rubber tubing until I could get some more hard tubing. I did that a few weeks ago and added the flow meter. Temps are pretty much the same either way since I got the case and added the second 360 rad. They were fine with the one but I had a Corsair Crystal 570X case so the airflow to the rad, in the front, was restricted with the front glass panel on. With the Corsair 5000X RGB case it does have a front glass panel but the side one is just a filter so less restriction, I am waiting for Corsair to get the 5000 AirFlow case front panel in stock.
 
Stated flow rate for any pump, is based on literally pushing fluid though a tube. No blocks, rads etc, same with the head pressure.

You will never get the max flow rate, regardless how accurate the flow meter is, you will probably realistically never hit the lower stated number, with even a basic setup.

All you can do is minimize restriction. But even that has minimal returns.
 
Stated flow rate for any pump, is based on literally pushing fluid though a tube. No blocks, rads etc, same with the head pressure.

You will never get the max flow rate, regardless how accurate the flow meter is, you will probably realistically never hit the lower stated number, with even a basic setup.

All you can do is minimize restriction. But even that has minimal returns.

Understood. I just wanted to make sure the numbers were what others may see and I see two people on here seem to get the same numbers. I really like it for the looks and the ability to easily see what the coolant temp is. I can see my coolant is flowing in the reservoir so I know it is pumping.
 
Back
Top