RATIONAL custom loop for 14900K and 4080S

[ThrashZone]

Hi,
Yeah I read he used one rad as intake and on the 2nd rad as exhaust
Talk about counter productive that is the best example ever thought up.

 

[BoggledBeagle]

I drained the loop and found out that it contains less than 700 ml of water, and redid the front rad the same way as the upper one, with the P14 fans pulling the air.





The card is really nicely compact, I like that, however I was surprised how heavy it is, it is possibly heavier than the large air cooled cards! I wonder if it does not need some support too?



I also found out that the CPU block has some crud in it already!



This case is not well suited for high airflow... But it looks nice!!!

 

[ThrashZone]

Hi,
Adding a filter with quick disconnects would stop a lot of clogging issues "rads are filthy" using it during filling and working the air out of the system which takes a couple days at most usually.
Of course coffee filter the fluid to.
These are awesome with soft tubing and QDC's
Spectre Clearview Premium Fuel Filter

 

[BoggledBeagle]

Hi,
Adding a filter with quick disconnects would stop a lot of clogging issues "rads are filthy" using it during filling and working the air out of the system which takes a couple days at most usually.
Of course coffee filter the fluid to.
These are awesome with soft tubing and QDC's
Spectre Clearview Premium Fuel Filter
View attachment 333913

There are FOUR disconnects in this loop, are they not too restrictive? The tubes also look very long.

 

[ThrashZone]

There are FOUR disconnects in this loop, are they not too restrictive? The tubes also look very long.

Hi,
Well if you use them correctly
Male is outlet and female is inlet so no not really that bad.
Bottom line they make life a lot easier so any loss in flow, you save ton in refilling.

Besides the filter will be more restrictive than the qdc's and it's temporary pump 100%.

 

[BoggledBeagle]

I tested temperatures with 21,5 °C ambient, all panels and filters on the case, with 70 and 50% fans and pump and front door closed and opened. With 70% fans and pump, closing the door increased temps by 3°.










How do these temps (GPU 53, hotspot 63, GPU ram 48) compare to temperatures of air cooled and water cooled 4080 GPUs?

I guess the case is bad for airflow and possibly smaller radiators must have some negative impact.

 

[BoggledBeagle]

With no external temp sensor connector on my cheap motherboard to plug the water temp sensor from the pump/res combo into, I gave up and bought the Core XT fan controller.

With the ambient temperature 20°C, 370W heat load and pump and fans curves in the picture, it stabilised the water temperature pretty quickly on 34,5 °C and the PC was pretty quiet too.

I am happy with the controller and results, unfortunately I payed a lot more for this fan controller bacause all that useless RGB stuff on it.

 

[BoggledBeagle]

I installed a temperature probe (controller came with two probes) on the input chamber of the first rad to see the temp of the water after it cooled the heat producing parts.

Then I loaded the water with cca 350W and with the pump RMP 800 the water must be flowing really slowly, because with the water at the pump at 27°C, it shot up to 43°C at the rad input. You can see the delay in water temps and how the water temp at the rad begun to drop when the pump started ramping up. With the pump up to speed, both sensors were reporting the difference of 0-0,3°C.

So with the water really flowing in the loop, temp differences when passing heat producing or cooling parts are really small. Everybody of course knows this, but seeing this myself is interesting.



You can also see an overshoot in the Water temp curve compared to the previous graph, because I configured the fans on the top rad to completely stop up to 30°C.

 

[Lei]

View attachment 333609

3) Shortest loop possible

4) Smoothest loop possible - no unnecessary bends, no right angle fittings, all changes in direction managed by hose itself, all to minimise water flow restrictions

5) Heat goes up: hoses from heat producing parts should run upwards.

Number 5: do you really think the temp difference throughout the loop is so big that direction makes sense?

Number 3: I rather take the GPU to one rad and then pour cooled down water on the CPU. Based on number 5, I guess you cool the GPU first and then run hot water on CPU.

I do it this way: go to GPU first, give it to largest rad, go to CPU, go to another rad, back to reservoir

 

[BoggledBeagle]

Number 5: do you really think the temp difference throughout the loop is so big that direction makes sense?

Number 3: I rather take the GPU to one rad and then pour cooled down water on the CPU. Based on number 5, I guess you cool the GPU first and then run hot water on CPU.

I do it this way: go to GPU first, give it to largest rad, go to CPU, go to another rad, back to reservoir

I measured a while ago that the water temp is almost the same everywhere in the loop, once the pump moves the water quickly enough...

But when the system is idle and the pump runs at minimal speed, the water gets heated with something like 30-50W by low single digit °C, So the natural convection may be helping slightly in this situation.

In the previous post I described situation, when the CPU was for a limited time before pump ramped up receiving water 8°C hotter than after that. With that a rad right after a GPU may help.

EDIT:

In idle state, when HWinfo reports CPU power draw 6W, GPU 12W, pump will be something below 10W, the real total heat load will be probably around 30W, the difference in temps of slow moving water is 3°C:



I believe that this slightly warmer water moving upwards has some positive effect on water circulation, most likely quite small.

 

[BoggledBeagle]

I was packing away the 4070 and I noticed, how HUGE is the box for the new much smaller card, it literally can be TWICE AS SMALL! This is so dumb. But it is shiny!

 

[Lei]

Is it possible doing this:
Use the big rad as exhaust, top rad as intake.
Go to GPU first, give it to front rad and blow the heat out, go to CPU and then top rad. Back to res. (Wrong, keep reading)

I would see how it goes if you use the top rad as exhaust too, while using the back fan as intake.

Hi,
Yeah I read he used one rad as intake and on the 2nd rad as exhaust
Talk about counter productive that is the best example ever thought up.

Yep, the only way closer to best counter productive would be take the loop after CPU and give it to intake.

If your intake radiator gets coolant down to 40C and heats up the inside of your case to 40C (fairly believable, ordinary values) then the exhaust radiator breathing in 40C air to cool down 40C coolant at the do nothing, because it's working on a ΔT of zero, and anything multiplied by zero is still zero.

@BoggledBeagle your loop is terrible, you're welcome

Now a better idea: use both rads as exhaust, back fan as intake. Go to GPU, rad, rad, CPU, res... Nope, your running hot water on GPU.
So go to GPU, rad, CPU, rad, res. Both rads exhaust.
Or if you want short loop, go to GPU, CPU, rad, rad but both rads exhaust.

I only have one big rad on the floor, so I don't know how air flow inside case works for water cooling. At least I'm exhausting at all times.

 

[ir_cow]

There are FOUR disconnects in this loop, are they not too restrictive? The tubes also look very long.

Quick disconnects arent that restrictive. I use 3-4 depending on my setup.

 

[Lei]

@BoggledBeagle
Your airflow is optimal for an air-cooled system, you have smooth flow and no vacuum.
But in water cooling the tubes take the heat away from components, so fans are not blowing the heat from heatsinks, but instead the rads are doing the job.
So you shouldn't blow one rad into another rad. Your big front rad (intake) is currently abusing your other rad.

You're basically using your biggest rad to blow all produced hot air back into system (including other rad, res, tubes, everything inside, mobo, ram, outerior surface of waterblocks etc )

I mean look: you're throwing water on everything and then collect it on a big birthday cake and then blowing the candles right back into what you're trying to cool.

 

[BoggledBeagle]

So you shouldn't blow one rad into another rad. Your big front rad (intake) is currently abusing your other rad.

Chrispy_ already wrote this, and on the previous page, post 48, I even provided an example calculation of the effect. Having one rad as an air intake and the second as an air exhaust led to 15°C over ambient water temperature, making both rads air intakes dropper this temp by 6°C.

I also informed in that post that I changed air direction on the top rad and both rads are now intakes!

 

[BoggledBeagle]

In idle state, when HWinfo reports CPU power draw 6W, GPU 12W, pump will be something below 10W, the real total heat load will be probably around 30W, the difference in temps of slow moving water is 3°C:

View attachment 334129

I have been thinking about this and the water must be barely moving with these low pump RPMs. Also the information about increased heat load gets to the pump water temp sensor very slowly... Perhaps I should increase minimal pump speed?

 

[Lei]

Recapitulation:

With heat load of 566W, using one rad as an air intake in the case and the second as an air exhaust would lead to coolant temperature of 35°C. Using both rads as air intakes would lead to coolant temperature 6°C lower.


Now I actually swapped the top rad to be an air intake too, the temps really improved, and now my case is swamped with warm air and I have just one 140mm fan and holes in the case to get rid of it. I do not like that too much...

Can you make them both exhaust and measure the temps?
Maybe just leave the side panel open

 

[BoggledBeagle]

Can you make them both exhaust and measure the temps?
Maybe just leave the side panel open

The performance would be similar to both as air intakes, the PC would be inside much cooler, which would be a very good thing, but also dirty, because I have no way of filtering the incoming air in this case.

 

[ThrashZone]

Yep, the only way closer to best counter productive would be take the loop after CPU and give it to intake.

lol yeah you'd hate to cool that air lets push more case hot air through it instead plus do this through a case filter as well that is well know to improve flow not obstruct it to right

 

[ThrashZone]

The performance would be similar to both as air intakes, the PC would be inside much cooler, which would be a very good thing, but also dirty, because I have no way of filtering the incoming air in this case.

Yeah back of the case or any near by cracks is where all the Unfiltered air would be coming from.

Not to worry man intake on rads is great those filters are there for a reason that escapes some people lol
all air entering the case will go out the back
If you want to improve that remove all pci-e slot covers.

 

[BoggledBeagle]

all air entering the case will go out the back If you want to improve that remove all pci-e slot covers.

They seem to be pretty permeable, are you sure that removing them would help much? The whole bottom of the case is perforated as well.

 

[ThrashZone]

Hi,
I do all intake so no they are just in the way of air exiting easy as possible.

 

[Chrispy_]

They seem to be pretty permeable, are you sure that removing them would help much? The whole bottom of the case is perforated as well.

The Define 7XL is a huge case with plenty of ventilation all over the top, bottom, and back. Removing the slot covers probably only increases the ventilated area by 2-3% which is unlikely to even cause a measurable difference of one whole degree - but hey, I could be wrong and testing it for yourself is really easy!

 

[BoggledBeagle]

I noticed that a german site reviewed the card, it is just 3-4 % quicker than the old 4080. The only big advantage is probably the price, I got this special version of the card for similar price the cheapest original 4080 were ever selling for.

https://www.hardware-mag.de/artikel/grafikkarten/inno3d_rtx_4080_super_ichill_frostbite_review/

It also seems that the updated cards are even more efficient.

 

[ThrashZone]

The Define 7XL is a huge case with plenty of ventilation all over the top, bottom, and back. Removing the slot covers probably only increases the ventilated area by 2-3% which is unlikely to even cause a measurable difference of one whole degree - but hey, I could be wrong and testing it for yourself is really easy!

Hi,
Yeah I got mid towers D450 to be exact lol