experiments in water cooling a mach part 2

[sacha35]

Results of S/S unit with cheap single rad/pump and twin fan setup.

I started off with the same settings as Bazx 402a, 2.75mts of 0.31 cap tube, but different evap and suction line, the results where around what we had seen from Bazx, but this was using only a single rad and pump.

I Reduced the cap tube lenth for higher loads to 2.65mts and refilled with 402a.

all the results shown are un-insulated as i am awaiting on some insulation, so the end results should be a little bit better than shown.

The results are with two temp probes, T1 is attached to the side of the evap, and T2 is attached 6mm from the rear of the evap on the suction line.

The results shown where with the unit being run for around an hour at each test.

Pic of the S/S using single rad and know insulation.

Unloaded results:

High side 215 psi Low side 4psi

Unloaded T1 temps: -59.2c

Unloaded T2temps: -58.0c

very small amount of floodback, but this will be not be a problem as know load is on the unit, with only 50watts load applied the flooback had gone.

Loaded results:

303watt load applied.

High side 260psi, Low side 9psi

Loaded T1 temps: -33.9c

Loaded T2 tepms: -30.3c

 

[maverik-sg1]

Nice - 300w is considerably more than current SLACR quads, right?

 

[sacha35]

Nice - 300w is considerably more than current SLACR quads, right?

This is true, i just want to see what the max wattage a S/S will be able to hold with a good return of temps, anything lower than 300watts will give a higher stable load temp.

 

[4Qman]

Amazing mate.

I will look forward o see how it holds a Quad. Also R410a with he water barrel.

 

[sacha35]

Just A little bit of whats to come from this small NL11, will post up some more results latter.

-45.1 @ 259watts

 

[sacha35]

Update:Just for fun.

At 353watts it was 260psi-270psi for the high side and 11 psi for the low side

As you can see the high side was going a bit mental with this load applied.

I think this must be some kind of record for a S/S and its only an NL11

 

[sacha35]

Well some updates, got my Q6600 on the single stage and it is rock stable at 4.5gig with 1.7v, the mother board is stoping this chip clock higher due to the Vdroop, but the temp probe does not go any lower than -49c even when loaded up.

here is some results of single card run in 3D06.

http://service.futuremark.com/compare?3dm06=3374424

And crossfire stock cooled cards, but there is still a problem clocking these as ATI tool does not work.

http://service.futuremark.com/compare?3dm06=3373615

 

[maverik-sg1]

Well some updates, got my Q6600 on the single stage and it is rock stable at 4.5gig with 1.7v, the mother board is stoping this chip clock higher due to the Vdroop, but the temp probe does not go any lower than -49c even when loaded up.

here is some results of single card run in 3D06.

http://service.futuremark.com/compare?3dm06=3374424

And crossfire stock cooled cards, but there is still a problem clocking these as ATI tool does not work.

[URL="http://service.futuremark.com/compare?3dm06=3373615"]http://service.futuremark.com/compare?3dm06=3373615[/URL]

Me and blindfitter have the same issue but trail you by 100mhz, mighty impressive result that mate - we were also querying the FSB limit on the chip mobo combo of around 490mhz, looks like you cracked that one though.

Can you overclock using Riva?

Also - I think there is a simple pencil mod that can help with droop?

But I think if we decide to use better cooling our board will go to persivore for a full suite of 'proper' mods.

Great work there sachae mate, thanks for sharing.

Mav

 

[sacha35]

Thanks mav, I have not tried Riva this is booting from bios settings but i have had a look and theVdroop is all over the place, once this is done i will show what it can do then, but the most volts i can get from the M/b IS AROUND 1.63 TO 1.67V this i9s with everything maxed out, vcore at 1.7 and overdrive X63 enabled and still it flutuates around these figers shown, 4Qman is looking for the mod for me and as soon as he finds it i will mod the M/B.

The problem is with running the 2900's in crossfire though as know tools are avalible like ATI tool which works great with one card but not at all in crossfire so there is loads of room for improvment in the 3D benches running crossfire, i think with a few more tweaks i should be able to hit 27000 in 05 with one card, but am going to have a rest for a bit from the benching as i have got to get back to the phase and cascade stuff for now.

 

[sacha35]

Still holding strong, -48c to -49c with the Q6600 at 4.5gig with 1.7v and with the E6850 at 4.8 gig its holding around -54c to -55c with 1.7v, as you know though this motherboard needs the vdroo doing badly as its is all over the place and sometimes goes bellow 1.63v so it is not that stable at 4.8gig,mbut i did do a 1M and cpuz of this for your veiwing.

http://valid.x86-secret.com/show_oc.php?id=255125

 

[bazx]

hehe you are getting some mad oc for an nl11h ss

 

[sacha35]

I make you right baz, but will have to pop down yours one weekend and get this setup on your kit, i could see some W/R comeing on your cascades.

Anyway This Single stage must be getting stronger each day,30K busted in 3D05 now.

http://service.futuremark.com/compare?3dm05=3567730