D5/MCP-655 users: significance of pump speed? (SURVEY)

If you using a D5/MCP-655 pump, could you please test the actual difference between the 5 pump speed settings. Please report both idle and load CPU temperatures (in Celcius, if possible) for all 5 settings. Let the temperatures settle completely before reporting them! If you want you can also tell the main components of your WC rig and your CPU type.

I just like to know how much influence does the flow rate have on temperatures in real world situation. Thanks for your help in advance!


[Setting: idle/load]

1:
2:
3:
4:
5:

CPU:
WC:

When I get mine ill do it, I doubt anyone will do it though, too lazy

-aDaM^ said:

When I get mine ill do it, I doubt anyone will do it though, too lazy

Click to expand...

have faith in ye fellow overclockers...
i'll do some tests when i get home.
think 20 minutes in between each test is sufficient time for things to settle out?

Haste266 said:

have faith in ye fellow overclockers...
i'll do some tests when i get home.
think 20 minutes in between each test is sufficient time for things to settle out?

Click to expand...

yup 20min is plenty. I think this test will show up like a sore thumb when using a Strom cause its restrictive. Should be interesting.

I'll try once I get my rig in the loop...just finished making the loop at 3AM

well I havent tested all the settings but just two.
For the first several days of my loop (after bleeding and leak testing for 3 days) I ran at setting 5. After about 4 days I switched it to setting 3.
I havent seen any difference at all in temps between settings 3 and settings 5

I'm using
BIP II
Maze 4 gpu
maze 4 chipset
6002-775
typhoon res

Stratus_ss said:

well I havent tested all the settings but just two.
For the first several days of my loop (after bleeding and leak testing for 3 days) I ran at setting 5. After about 4 days I switched it to setting 3.
I havent seen any difference at all in temps between settings 3 and settings 5

I'm using
BIP II
Maze 4 gpu
maze 4 chipset
6002-775
typhoon res

Click to expand...

find out the difference between 1 and 5, im sure there will be a difference.

I ran my pump at Speeds 1, 4 and 5, and never noticed any difference in my load temps, which I found surprising.

Voodoo Rufus said:

I ran my pump at Speeds 1, 4 and 5, and never noticed any difference in my load temps, which I found surprising.

Click to expand...

is there even a noticable change in sound or flow?

Level 1 was very quiet, but still audible to me. But I run my computer as close as I can to silent. Level 4 was alright, but 5 had too much whine for me.

Flow difference between levels was very significant. 1 was a snails pace, 5 was haulin'

Slightly off topic question, can you change the pump speed while running? (I'd assume so, but I'm interested)

Burdman27911 said:

Slightly off topic question, can you change the pump speed while running? (I'd assume so, but I'm interested)

Click to expand...

yes you can.

#1 - Idle 38c Ambient 34c | Load 43c ambient 34c
#2 - Idle 37c Ambient 34c | Load 43c ambient 34c
#3 - Idle 36c Ambient 34c | Load 42c ambient 34c
#4 - Idle 37c Ambient 34c | Load 42c ambient 34c
#5 - Idle 36c Ambient 33c | Load 42c ambient 34c

Ambient is internal case temp(motherboard temp, whatever u wanna call it)
Opteron [email protected] 1.475v
Swiftech Apogee on cpu, MCR-220 rad
I also have a Stasis Neptune Block in the loop but currently isnt attached to a card due to a very small leak(waiting for replacement)

Here is how i did it. Turned my computer on when i got home from work, went grocery shopping...then started the tests when i got back home so everything was definitely at normal idle temps.

I started out at #5 setting and decreased as I went. I ran each stress test for 20 minutes long each with Stress Prime 2004 for each core. After 20 minutes, temp was recorded. I then would turn the dial 1 setting lower and wait 20 minutes. Then recorded the # for the idle temp. Rinse, repeat.

Very surprising actually. I was expecting a higher temp increase...
at the #1 setting i can BARELY even tell the pump is running. its hard to tell its running even when putting my hand on it. #3 seemed to be the loudest setting for me as it seemed to put out the most vibration(pump is on a pad that came with it)

hope this helps.

Haste266 said:

#1 - Idle 38c Ambient 34c | Load 43c ambient 34c
#2 - Idle 37c Ambient 34c | Load 43c ambient 34c
#3 - Idle 36c Ambient 34c | Load 42c ambient 34c
#4 - Idle 37c Ambient 34c | Load 42c ambient 34c
#5 - Idle 36c Ambient 33c | Load 42c ambient 34c

Ambient is internal case temp(motherboard temp, whatever u wanna call it)
Opteron [email protected] 1.475v
Swiftech Apogee on cpu, MCR-220 rad
I also have a Stasis Neptune Block in the loop but currently isnt attached to a card due to a very small leak(waiting for replacement)

Here is how i did it. Turned my computer on when i got home from work, went grocery shopping...then started the tests when i got back home so everything was definitely at normal idle temps.

I started out at #5 setting and decreased as I went. I ran each stress test for 20 minutes long each with Stress Prime 2004 for each core. After 20 minutes, temp was recorded. I then would turn the dial 1 setting lower and wait 20 minutes. Then recorded the # for the idle temp. Rinse, repeat.

Very surprising actually. I was expecting a higher temp increase...
at the #1 setting i can BARELY even tell the pump is running. its hard to tell its running even when putting my hand on it. #3 seemed to be the loudest setting for me as it seemed to put out the most vibration(pump is on a pad that came with it)

hope this helps.

Click to expand...

Great post Haste, thanks for the info.

I don't have time to do this test but I can tell you one thing for certain and that is my Storm block is a lot more sensitive to flow rate than my 6002 block was.

Still, I notice no difference really unless I set it to position 1, then I do see about a 2c rise at idle almost immediately.

I only tested #5 & #3

#5 - Idle 26C - Load 36C
#3 - Idle 27C - Load 36 ~ 38C

Apogee & BIP II

Hmm...glad I can run this pump slower than full blast now, causes my whole case to vibrate.

something i forgot to add was that i ran the last test at the #1 setting for 45 minutes to make sure my 20 min. time was enough to heat up the system.
i had the same reading at both 20 and 45 minutes.

Very interesting results, thanks - you can still post them. So what early conclusions can be made?

1. Flow rate has a very small meaning in computer water cooling systems. The flow rate difference between setting 5 and 1 is a lot as can be seen here:

http://www.systemcooling.com/images/reviews/LiquidCooling/Swiftech_MCP655/image16big.gif

Still, there were only a 0-2 degrees difference in CPU temps and in some cases larger flow rate even increased the temps. In some cases these are probably within the error margin/resolution of the meter so the actual temperatures might've been the same.


2. DDC/MCP-350 has a similar flowrate than D5/MCP-655 at setting 3-5 as can be seen here:

http://coolercasesuk.co.uk/showthread.php?t=15
http://www.systemcooling.com/images/reviews/LiquidCooling/Swiftech_MCP655/image17big.gif

The next graphs suggests even better flow rate for DDC in real systems:

http://www.cooling-masters.com/articles-33-9.html
http://www.cooling-masters.com/articles-33-11.html

The flow rate of DDC can be very close to D5 in MORE restricting systems! (Translate here - it's French )


3. DDC heats the water less than D5. Here you can see how much of the energy is going into the water. They measured probably how long it takes for the water temperature to increase a certain amount of degrees:

http://www.cooling-masters.com/articles-33-13.html

Again, DDC shines. D5@5 heats the water 0,3 degrees in 200 seconds. DDC heats the same amount in about 450 seconds! This is because DDC uses less energy and it is also made of plastic which doesn't conduct heat as efficiently. This is a good thing in pumps. Here you can see that DDC actually outperforms almost all pumps. Only AquaXtreme 50Z outperforms it in this particular simulation. It also seems that using multiple pumps is a bad idea. See the lower graph:

http://www.cooling-masters.com/articles-33-16.html


I think this is a very important issue as you can have a cooler/equally cool system with:

• less noise
• smaller power consumption
  • less heat to the water
  • less power draw from the PSU -> less heat to the system
• lower cost
• smaller size

I'm not saying that D5/MCP-655 is a bad pump. Most likely it WILL outperform DDC in certain kind of systems - (what kind of?). It just seems that flow rate is not everything. I'm very new to water cooling so please, let's continue this discussion (if anyone is interested). I might've missed some important things.

I have never tested ANY system, with any combination of components, in which the D5 (large Liang pump of any variety) would yield better results than using the DDC (the 10W version).

For all the obvious reasons.
Big high flow pumps are like 1/2" tubing.

This is an interesting topic. I assumed running at the highest setting would give the best performance, but I forgot about the heat dump from the motor going into the water. Whether that is more important than the change in flow rate I am not sure. Was getting about 40-42 cpu load with setting 5. (I have a gfx card and cpu in the loop). I just turned it to 2. If it changes in the next hour i'll let you know.