Normal air to water delta ? + my system pix

[The Byter]

Hello guys !

I just got my last setup running and I have a few Q.
My system, 875k + gtx470sli.

My data:
One rx360, fans @ 100% 1800RPM, D5 @ speed 3, air in temp 23C~

oc cpu+gfx
water in: 34.1C
water out: 31.1C

oc cpu, gfx stock
water in: 33.4C
water out: 30.6C

all stock
water in: 31.2C
water out: 28.8C


I see at Skinnee rx360 review that the air to water delta is 1C~ all the way.



http://skinneelabs.com/xspc-rx360-v1/4/

Does my 3C~ delta mean:
- Low flow ?
- The rad is not enough ?
- both ?

10x

pix :

[Johnny87au]

Have you tryed running your d5 at speed 5?? 3 Delta is great dude... Your not gonna get much better then that unless your ambients are rocking!

[The Byter]

The XPSC res has one major flaw, I CAN NOT change the pump speed once installed :\

How come skinnee got 1C delta all the way ?

[Johnny87au]

You checked how much flow your getting with a flow metre? Which fans you using? Im betting martins ambients were really good..

[The Byter]

The air temp does not matter for the delta.

his air is 19-23C~

[Johnny87au]

Ambients do help alot.. 1c Delta is usually ran with a great setup and quite a bit of rad surface with good fans pushing with good static pressure.. Try adding another rad to your loop, maby a dedicated rad for the SLI cards..

[Ttocs]

In skinnees review water to water delta (water in radiator to water out of radiator) is always less than a degree Celsius, which makes our assumption of water throughout the loop always being a constant true (or at least allows us to make that assumption). Water to air delta can be anywhere from about 3 degrees Celsius (the only review I've seen which had anything close to that was a Mora and a single CPU) to an upper level of about 15 degrees Celsius. The sweet spot that Martin used to mention was about 10 degrees, xtreme people shoot for 5, I (personally) don't like anything higher than 15.

Your delta is between 8 and 11 degrees. It's fine. Only thing I might consider is next time you take apart the loop, up the pump speed to max, unless the noise irritates you. Otherwise you have nothing to worry about (unless the temperature numbers you posted are at idle, because if they are, you'll need to repost load water temperatures to give us a better idea how to help).

[skinnee]

Look at the constant across all of the test runs for that round of testing... 1.5GPM flow rate. That is why the radiator delta is ~1C.

[The Byter]

Thank you for the info

Skinnee, that is way I ask:
Does my 3C~ delta mean:
- Low flow ?
- The rad is not enough ?
- both ?

all numbers are in game load.

[skinnee]

You have lower flow than I tested the radiator at. Side note. this is one of the many reasons we are now doing variable flow rate testing on radiators, to measure flow dependence. Also, your sensors probably aren't calibrated either... so there is some variance there as well.

Turn that D5 up to S4, and you'll see the radiator delta decrease, then again at S5.

The real numbers you want to go after are air to water average though. However, it did bring a smile to my face that you are paying attention to the tests and your first bullet of lower flow was the right guess.

BTW, nice build... especially the use of quick disconnects.

[Johnny87au]

You have lower flow than I tested the radiator at. Side note. this is one of the many reasons we are now doing variable flow rate testing on radiators, to measure flow dependence. Also, your sensors probably aren't calibrated either... so there is some variance there as well.

Turn that D5 up to S4, and you'll see the radiator delta decrease, then again at S5.

The real numbers you want to go after are air to water average though. However, it did bring a smile to my face that you are paying attention to the tests and your first bullet of lower flow was the right guess.

+1 same thing i mentioned

[prznar1]

could you tell me what gets coolant first from the pump? cpu or gpu?

[Biffa]

Looks like the lower GPU from the pics

[Johnny87au]

Yeah looks like hes going up through the gpu to the cpu back to res/rad

[The Byter]

res,pump-->rad-->cpu->gfx-->gfx-->res

Thanks Skinnee

[Defoler]

Looks like the lower GPU from the pics

Its the CPU.

His loop is res+pump->rad->cpu->Gpu->gpu->res.

[Biffa]

Gotcha, I see now from this:

How the flow works with that bayres, I couldn't see from your photo which was in and which was out!

[Holst]

Unless you have calibrated your sensors they are probably only accurate +/- 2*C

[jmn2519]

The XPSC res has one major flaw, I CAN NOT change the pump speed once installed :\

This is something that needs to be addressed. I really like the new DD and Koolance pump/res combos but I don't want to have to disassemble the whole thing if I want to change the speed setting on my pump.

That's why I started the thread about controlling voltage to the pump. I'd love one of you guys to invent a product the easily do that!!!

[The Byter]

jmn2519 - you can use any voltage regulator (like a fan controller) that can handle more power (W) then the pump in order to control the pump voltage (speed).

This is the data I collected up to now:



As we can see higher pump speed (flow rate) decreased the in/out delta, BUT had no effect on air in/avg water delta...

I believe my 10C~ in air/avg water delta is ok, with another RX360 this delta will go down right?

The down side is that all of this is @ 1800RPM fans - I think I will get the GT1850.

I hope you like

[Johnny87au]

jmn2519 - you can use any voltage regulator (like a fan controller) that can handle more power (W) then the pump in order to control the pump voltage (speed).

This is the data I collected up to now:



As we can see higher pump speed (flow rate) decreased the in/out delta, BUT had no effect on air in/avg water delta...

I believe my 10C~ in air/avg water delta is ok, with another RX360 this delta will go down right?

The down side is that all of this is @ 1800RPM fans - I think I will get the GT1850.

I hope you like

Typically expected, of course more flow is gonna help your delta water in/out temps.. As yes i believe more rad surface with good static pressured fans would yield u better results , dont expect a 5C delta gain maby 2-3c..

[Conumdrum]

A high wattage CPU and two kinda hot GPUs on a 360RX? Not enuff rad. Easy enuff.

Unless your load temps running Prime 95 and furmark at the same time for 30 min are good, then your fine. Fannage won't help unless you want 3000 RPM screamers, and the rad you got doesn't scale well with them.

Get a very high FPI rad and better fans, better and louder than the AP15s. Then you can live with a single 120x3 rad.

Not enuff rad.

[StAndrew]

Looking at your temps, you rad is doing well, however can benefit from a second rad. Rule of "diminishing return" applies.

Nice rig

[The Byter]

"Rule of "diminishing return" applies." ? can you explain ?

10x !

[Church]

diminishing return is that you can enhance by adding/upgrading something only to the point. And nearer that point, the less bang for a buck you get, making at some point it unwise to keep investing.


Simple example - rads. By adding them you enhance delta of air-liquid temperature making it less and less .. but the smaller delta gets, the less you get from each next rad you add. If you think that you should put 9x120 rad for single cpu, you'd be better off switching to phase cooling
Similar is with liquid flow/pumps. You can get better cooling from more flow/more powerful pumps. But if going from 0.5-1gpm will net you 4-5C, 1.5gpm at most 1C, and above 1.5 even if you manage to get 3-5gpm it will net you decimal parts of degree.