Connecting two Komodo HD5800 blocks in parallel or serial, and why?

[dejanh]

Quick question, connecting two Komodo HD5800 blocks, should I connect them in parallel or in series. I am inclined to do series, but what would be the benefit if any of doing it in parallel?

I have two DDC 3.2 pumps in serial pumping through the loop, cooling CPU, mobo, and GPU all in one loop.

[mattkosem]

If you have other blocks in the loop, I'd go parallel to reduce the flow impact. Odds are you won't see much if any difference in temps. And, if you do, they likely won't be drastic enough to affect your GPU clocks.

--Matt

[Boulard83]

Considering you have 2 pump, id go Serial.

[individual]

Parralell, because if you go serial, any heat picked up from your first card will be transferred to your second one, but make sure you have a decent radiator, I'd have a triple at least for all that to be honest. I've got a quad for my CPU and Mobo, and a dual for my 5870.

[lowfat]

Parralell, because if you go serial, any heat picked up from your first card will be transferred to your second one, but make sure you have a decent radiator, I'd have a triple at least for all that to be honest. I've got a quad for my CPU and Mobo, and a dual for my 5870.

If he only has one loop, it isn't going to make any difference.

[Boulard83]

The temp difference in the loop is minimal. If you have decent flow and Radiator youll notice only a 1-2°c from the IN to the OUT of the radiator.

[Blauhung]

The temp difference in the loop is minimal. If you have decent flow and Radiator youll notice only a 1-2°c from the IN to the OUT of the radiator.

if not less

Also if you do run in parallel your flow rate to each card is ~1/2 what you would get if you went serial. Higher flow = better heat transfer so if you 1/2 the flow you get higher temps on each card

[Boulard83]

+1 again ! SERIAL !

[Kadett]

if not less

Also if you do run in parallel your flow rate to each card is ~1/2 what you would get if you went serial. Higher flow = better heat transfer so if you 1/2 the flow you get higher temps on each card

I followed Malik's ATCS 840 build and he went down 2/3 degrees by going to parallel from serial. And thats 1 big loop for the entire system.

I think you get a better flow overall because there's less restriction forcing the water thru the GPU blocks.

[mattkosem]

I followed Malik's ATCS 840 build and he went down 2/3 degrees by going to parallel from serial. And thats 1 big loop for the entire system.

I think you get a better flow overall because there's less restriction forcing the water thru the GPU blocks.

I had the same experience on my 4890s. The loop was only 2 rads, cpu, 2 gpu.

--Matt

[dejanh]

Ahh, heck, I will try both combinations and see if there is any difference. As it stands now the GPU temps on a heavily overclocked single 5850 with the Komodo block never really pass about 35C which is great. Second card is on its way, second Komodo is already here. I'll post back with the results after.

Having said that, my more pressing problem ATM is dealing with the VRM temps. I went straight for the Advanced Thermal Utilization step seen on page 5 here http://swiftnets.com/products/instal...ultilangue.pdf but I am getting very lousy VRM temps and I think I know why...the block does not seem to be making adequate contact with the VRMs at all, but there is no way to screw in the block any further without cutting the fasteners to shorten them. I will have to ask Gabe about this.