Temperature/Block Comparison For 8800GTX

[serialk11r]

Without actually measuring water temp that is a bold statement. Whats the efficiency of the voltage regulation? Power consumption of ram? What fans at what voltage were used?

Pa 120.2 with Nexus at 7 V would net a 2,5 Degree water temp increase for ~35 Watts additional heat dump, 1.

Now where does it say that???????? WTF are you talking about? I don't see anything that compares water/air temp difference to heat dissipated. I only see "10C coolant/air difference".
You think ram consumes lots of power? At 1GPM, it takes 250W of heat to get the water up 1C. Of course, the rad won't dissipate all of it. If the ram and stuff really was that hot, then I don't think ramsinks would do the job...

[Vapor]

At 1GPM, it takes 250W of heat to get the water up 1C.

That forumula is for coolant (well, water) temperature differences immediately before and after the heatsource. Not a sustained load coolant temp. If that were true, radiators would be useless.

Considering the 8800GTX pulls ~150W (being nice) when OC'd and the core uses at most 110W of that (closer to 100W if I had to guess), and say the EK absorbs 2/3rds of it, that's another 25-35W of heat, and on a 120x2 radiator, that could easily lead to the few C increase for coolant temps for the whole loop (witnessed in both GPU and CPU temps). Restriction is also playing a part, but the effects of restriction are probably way overrated.

[Fairydust]

Now where does it say that???????? WTF are you talking about? I don't see anything that compares water/air temp difference to heat dissipated. I only see "10C coolant/air difference".
You think ram consumes lots of power? At 1GPM, it takes 250W of heat to get the water up 1C. Of course, the rad won't dissipate all of it. If the ram and stuff really was that hot, then I don't think ramsinks would do the job...

Check the thermal resistance vs flow rate for PA120.2 chart, 0.072 (°C/W) * 35 = 2,52 °C.

[serialk11r]

That forumula is for coolant (well, water) temperature differences immediately before and after the heatsource. Not a sustained load coolant temp. If that were true, radiators would be useless.

Considering the 8800GTX pulls ~150W (being nice) when OC'd and the core uses at most 110W of that (closer to 100W if I had to guess), that's another 40-50W of heat, and on a 120x2 radiator that's already stressed, that easily leads to the few C increase for coolant temps for the whole loop (witnessed in both GPU and CPU temps).

Yes, but I'm sure there are other components that use some energy as well. You've seen the fullcover blocks' base designs right? The water doesn't flow over some components, so the thermal resistance should be pretty high. If the core runs hotter, then less heat is making it into the loop as well right? That's another thing

[serialk11r]

Check the thermal resistance vs flow rate for PA120.2 chart, 0.072 (°C/W) * 35 = 2,52 °C.

The thermal resistance is dependent on coolant/air temp difference lol.

[AuroraProject]

How much did the EK increase the cpu temps at idle?

Anyone?

[pplapeu]

"Considering the 8800GTX pulls ~150W (being nice) when OC'd and the core uses at most 110W "

I would say 150w for the core alone. It is practically an order of magnitude larger that any CPU I ever laid eyes on, and I read it has more transistors than any CPU ever.

Stock cooling people report 65C idling and 80+C loaded. My GTS showed up nearly as hot with the same chip dumbed down a tad.

[serialk11r]

"Considering the 8800GTX pulls ~150W (being nice) when OC'd and the core uses at most 110W "

I would say 150w for the core alone. It is practically an order of magnitude larger that any CPU I ever laid eyes on, and I read it has more transistors than any CPU ever.

Stock cooling people report 65C idling and 80+C loaded. My GTS showed up nearly as hot with the same chip dumbed down a tad.

What makes you think that the core will pull 150W? Just because it runs hot doesn't mean its generating that much heat. A soldering iron might only have 20W going through it but it can reach 250C or more.

[serialk11r]

Anyone?

Read: It said with the EK block temps on CPU were 30 idle, and with only the CPU in the loop, it was 24C

[Vapor]

Yes, but I'm sure there are other components that use some energy as well. You've seen the fullcover blocks' base designs right? The water doesn't flow over some components, so the thermal resistance should be pretty high.

Thermal resistance may be kinda high, but the whole point and DESIGN of the block is to draw heat away from all the hot components and into the water. Just because it's not OUTSTANDING at it doesn't mean it's not doing it. The heat is getting to the coolant. Witness the GPU ambient temps dropping from 51C (MCW60) to 40C (EK)...

If the core runs hotter, then less heat is making it into the loop as well right? That's another thing

Right, except CPU temps went up as well, which is solely indicative of either coolant or ambient temp rising....

[Vapor]

"Considering the 8800GTX pulls ~150W (being nice) when OC'd and the core uses at most 110W "

I would say 150w for the core alone. It is practically an order of magnitude larger that any CPU I ever laid eyes on, and I read it has more transistors than any CPU ever.

Stock cooling people report 65C idling and 80+C loaded. My GTS showed up nearly as hot with the same chip dumbed down a tad.

ATI pegged the R580 core drawing only a little over half the card's power at stock voltages. So I am being generous in my argument when I say 110W out of 150W.

0.072 (°C/W) * X = 5 °C, X is equal to 70W, so 105W (70 + 35 quoted earlier) may be the total card draw at these clocks. 5C delta comes from CPU temp increase. Either way, the added heat dump from the EK's full-cover is 50% higher than just the core alone based on dT, so 100W out of 150W when OC'd is probably not far fetched either.

[Krowten-Dog]

Does the fact that you did not use a washer/o-ring spacer under the blank plug on the copper side make a difference?

I noticed on my (trusty calipers) block that without the spacer under the plug, that the plug extended ~1.5 mm into the 5 mm water channel. Would that loss of 1.5 mm make any difference in waterflow?

[Whokidmo]

Question, does air cooling the mem chips on the VGA and watercooling the main chip decrease your max OC, i mean the mem chips might over heat etc, hey OP can you do some major OCing? i honestly think air cooling and WC don't mix and that even if u do push the 2 set ups to the same speeds, 700/1200, you'll eventually burn out the swiftech card cause the mem chips get so hot with you OC, thats just my 2 bits at the end of the day, WCing is better then ACing which is why were all upgrading so why use a ty hybird? sure the main chip temp is down but i honestly i'm like 99.9999% sure that the temp on the mem chips will increase by alot with the swiftech block or the mem chips will burn out, SHOW SOME OC TESTS.

btw i remind people many said that southbridges and vregs on mobos didn't need to by WCed... and now many disagree.

[Ambro]

Does the fact that you did not use a washer/o-ring spacer under the blank plug on the copper side make a difference?

I noticed on my (trusty calipers) block that without the spacer under the plug, that the plug extended ~1.5 mm into the 5 mm water channel. Would that loss of 1.5 mm make any difference in waterflow?

Good question - I didn't find anything in Eddy's instructions about the use of those two spacers.

Eddy can you shed some light on how we are supposed to hook these in?

[Snyxxx]

First, thanks to everyone for the compliments.

Second, actually I did think about the spacers on the EK block. I read somewhere on these forums that they keep the threaded area of the barb from protruding into the flow channel.

Before I screwed the barbs in, it was obvious that the barb screwing into the copper side would protrude into the water channel. However, the delrin side was a different story. The threads were not as thick as the delrin piece and figured I did not need the spacer.

In hind sight, I probably should have just put the other spacer in and been done with it.

[pplapeu]

What makes you think that the core will pull 150W? Just because it runs hot doesn't mean its generating that much heat. A soldering iron might only have 20W going through it but it can reach 250C or more.

It is practically an order of magnitude larger that any CPU I ever laid eyes on, and I read it has more transistors than any CPU ever

[pplapeu]

I earlier suggested the UGLY solution I found. Now I have posted the pics here at XS. The UGLY part really is the dual heater cores. You can see it in my link below.

As an engineer you understand the water in a water cooling solution is the heatsink.

In my solution the heatsink weighs 16.5# as it is two gallons. If you run an in case 1qt water heat sink you must rely totally on excellent rad heat exchange. I added two cores, shrouds, and 120v fans I can flip on with a lamp cord/switch. I used hardware store 4" poly 45s & tees [for two gallon res that is a dog leg with a riser to fit the round flipping coffee table enclosure-heated water enters the riser] and reduced/bushed down to the size of the cooling system, in this case 3/8". I use an I/R pump that is equivalent to an Iwaki rxl20.

http://www.xtremesystems.org/forums/...d.php?t=138100

[ranker]

Question, does air cooling the mem chips on the VGA and watercooling the main chip decrease your max OC, i mean the mem chips might over heat etc, hey OP can you do some major OCing? i honestly think air cooling and WC don't mix and that even if u do push the 2 set ups to the same speeds, 700/1200, you'll eventually burn out the swiftech card cause the mem chips get so hot with you OC, thats just my 2 bits at the end of the day, WCing is better then ACing which is why were all upgrading so why use a ty hybird? sure the main chip temp is down but i honestly i'm like 99.9999% sure that the temp on the mem chips will increase by alot with the swiftech block or the mem chips will burn out, SHOW SOME OC TESTS.

btw i remind people many said that southbridges and vregs on mobos didn't need to by WCed... and now many disagree.

Vregs appear to be more about bling. Before settling on a 3.6O GHz OC on my e6600/evga 680i combo, I was running a 3.8 Ghz OC for a while. Vregs never really caused me any concern.

The only SB that's worth WC'ing is that of the 680i. Most other mb's don't have SB's that generate as much heat.

[serialk11r]

Vregs appear to be more about bling. Before settling on a 3.6O GHz OC on my e6600/evga 680i combo, I was running a 3.8 Ghz OC for a while. Vregs never really caused me any concern.

The only SB that's worth WC'ing is that of the 680i. Most other mb's don't have SB's that generate as much heat.

Even that SB might not be worth it. A zalman thingy that costs 5 bucks will do the job fine (I mean its an SB, do you really need awesome cooling?).

[Mallen]

Even that SB might not be worth it. A zalman thingy that costs 5 bucks will do the job fine (I mean its an SB, do you really need awesome cooling?).

The southbridge on the 680i and 650i chipsets do need good cooling to get good overclocks. I am running a p5n32-e sli plus and with my water cooling down at the moment I need to run a box fan at the side of the case to keep the southbridge cool enough to hold stable.

[Whokidmo]

The southbridge on the 680i and 650i chipsets do need good cooling to get good overclocks. I am running a p5n32-e sli plus and with my water cooling down at the moment I need to run a box fan at the side of the case to keep the southbridge cool enough to hold stable.

agreed so don't you think soon if not now the mem chips on VGA boards will need to be WCed for good OCs?