Overclock.net banner

1 - 20 of 27 Posts

·
Premium Member
Joined
·
2,433 Posts
Don't understand the question. Optimal flow rate is the same for all blocks. 1GPM
 

·
Village Idiot
Joined
·
2,368 Posts
Don't understand the question. Optimal flow rate is the same for all blocks. 1GPM
I'd love to see some data to support that.
 

·
Premium Member
Joined
·
2,433 Posts

·
Premium Member
Joined
·
2,301 Posts
dont know if anyone has delta core to water vs flow for heatkiller iv, but below is graph of 20 different blocks including its predecessor heatkiller III and 2nd link martins testing. The steep part of curve is at lower flow end with flattening/diminishing returns at higher flow. 227lph = 1gpm.

some blocks respond more to flow than others, ideally find graph for heatkiller iv and decide what you think is optimum, as opinions will vary for that word choice.

On all my builds for past 20yrs I aim for ~1gpm as well...but given I prefer setting just over 4 on D5 pump (tone wise) with single cpu/gpu block and 2 rads, I end up between 1 and 1.2 gpm anyways, including with current heatkiller iv.

https://abload.de/image.php?img=tempflowall6kgp.jpg

http://martinsliquidlab.petrastech.com/MartinsFlowRateEstimator.html
 

·
The last VRM burner
Joined
·
2,590 Posts
Just get a D5 and manipulate its flowrate under stress test until you're satisfied lol.
 

·
Registered
Joined
·
7 Posts
Discussion Starter #9 (Edited)
View attachment 364912
Just get a D5 and manipulate its flowrate under stress test until you're satisfied lol.

Just trying to get an idea where to start...


Looking at the charts that were provided, if I am interpreting them correctly, it appears Δ is better with lower flow rate, but overall temperatures are lower with high flow rate.


EDIT:


Taking @BroadPwns advice, I ran a quick test.


Using LinX 0.9.5 as stress test, 35000 problem size, 3 passes (about 5 minutes). Radiator is placed next to A/C vent, that is why Water temps are lower than Ambient temps whose sensor in located inside chassis.
 

Attachments

·
Premium Member
Joined
·
2,301 Posts
If noise of pump doesnt bother you, the highest setting on a D5 will give best temps since heat transfer improves to small degree at radiator and larger degree at waterblocks with higher flow, which you can see with your own testing. Optimal flow to me is the highest D5 setting that I can tolerate the pump noise.
 

·
Premium Member
Joined
·
2,433 Posts
If noise of pump doesnt bother you, the highest setting on a D5 will give best temps since heat transfer improves to small degree at radiator and larger degree at waterblocks with higher flow, which you can see with your own testing. Optimal flow to me is the highest D5 setting that I can tolerate the pump noise.
The higher you run your D5 the more heat it dumps into the loop. So while you do want a higher flow, you also want to dump as little heat from the pump into the loop as you can.
 

·
Registered
Joined
·
7 Posts
Discussion Starter #12
If noise of pump doesnt bother you, the highest setting on a D5 will give best temps since heat transfer improves to small degree at radiator and larger degree at waterblocks with higher flow, which you can see with your own testing. Optimal flow to me is the highest D5 setting that I can tolerate the pump noise.

Pump noise is imperceptible.


Actually, all I care about is performance. Noise is the very least factor.


Setting 5 it is, 9.4 L/m.
 

·
Registered
Joined
·
7 Posts
Discussion Starter #13
The higher you run your D5 the more heat it dumps into the loop. So while you do want a higher flow, you also want to dump as little heat from the pump into the loop as you can.

Even at Setting 5, which I presume would dump the most pump heat into the loop, the load temperatures are best, which is all I care about.
 

·
Premium Member
Joined
·
2,301 Posts
The higher you run your D5 the more heat it dumps into the loop. So while you do want a higher flow, you also want to dump as little heat from the pump into the loop as you can.
once you start adding more pumps, you can run into scenarios where heat dump from pumps will exceed temp lowering from increased flow, but with 1 pump just not going to run into that issue by increasing speed, not to mention maybe 10W difference between low speed and high speed. Skinneelabs and martin both found that 1 or 2 pumps the temp lowering effect persisted in typical cpu/gpu/rad loops, only after 3rd pump did heat dump from pump exceed diminishing returns from increased flow causing slightly increased temps.

edit: @kelkel1, yes if noise isnt an issue for you, highest setting will give you best temps.
 

·
The last VRM burner
Joined
·
2,590 Posts
The higher you run your D5 the more heat it dumps into the loop. So while you do want a higher flow, you also want to dump as little heat from the pump into the loop as you can.
23W of heat at 100% (D5 from EKWB) is nothing. It won't even raise waters temperature by 1C.
 

·
Village Idiot
Joined
·
2,368 Posts
I'm not being dragged down that hole again.
So you're just going to make a baseless blanket statement, pass it off as fact, and then get angry when someone asks for data to support your claim?

Ok.
 

·
Premium Member
Joined
·
2,433 Posts
So you're just going to make a baseless blanket statement, pass it off as fact, and then get angry when someone asks for data to support your claim?

Ok.
You can use google too.
 

·
Village Idiot
Joined
·
2,368 Posts
You can use google too.
https://www.google.com/search?q=Optimal+flow+rate+for+Heatkiller+IV+pro

Hmm... I'm not seeing "1 GPM" anywhere. Maybe I'll try a different search term:

https://www.google.com/search?q=Optimal+flow+rate+for+all+water+blocks

Nope, still not seeing "1 GPM anywhere...

https://www.xtremerigs.net/2012/11/09/2012-cpu-water-block-roundup/4/

Ok now this is interesting... It certainly looks like the performance of every single block tested scales beyond 1 GPM... I thought you said 1 GPM was optimal for every block?

https://www.guru3d.com/articles_pages/choosing_the_optmal_zen2_ek_waterblock_review,8.html

Here's another interesting one. All 3 blocks (and both jet plate configurations on the quantum magnitude) scale with flow beyond 1 GPM. So how again is 1 GPM optimal?

If you know of some other testing that shows otherwise, or even better, specific flow rate vs temperature scaling for the HK IV Pro (like the OP was specifically asking about), please do share.
 

·
The last VRM burner
Joined
·
2,590 Posts
https://www.google.com/search?q=Optimal+flow+rate+for+Heatkiller+IV+pro

Hmm... I'm not seeing "1 GPM" anywhere. Maybe I'll try a different search term:

https://www.google.com/search?q=Optimal+flow+rate+for+all+water+blocks

Nope, still not seeing "1 GPM anywhere...

https://www.xtremerigs.net/2012/11/09/2012-cpu-water-block-roundup/4/

Ok now this is interesting... It certainly looks like the performance of every single block tested scales beyond 1 GPM... I thought you said 1 GPM was optimal for every block?

https://www.guru3d.com/articles_pages/choosing_the_optmal_zen2_ek_waterblock_review,8.html

Here's another interesting one. All 3 blocks (and both jet plate configurations on the quantum magnitude) scale with flow beyond 1 GPM. So how again is 1 GPM optimal?

If you know of some other testing that shows otherwise, or even better, specific flow rate vs temperature scaling for the HK IV Pro (like the OP was specifically asking about), please do share.
Probably optimal in terms of power draw + noise compared to end result.
 

·
Registered
Joined
·
226 Posts
This reminds me the old days of "big blocks" Dzen or something like that, vs low flow WB like Inovetek here in Europe....

I prefer have low flow and 4 radiators, than someone with superflow and only 2....

Hey, but what do i understand of this....
 
1 - 20 of 27 Posts
Top