• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

can anyone beat my temps

turbolude

turbolude

Gawd
Joined
Jan 23, 2005
Messages
790
just installed my maze 4 gpu block and swiftec ram sinks onto my BFG 7800gt.

im oced to 510/1.23

for some reason i can oc mem higher then ,my core on this card comepared to my last evga. wich hit 515-520/1.20 but this card runs extremely cooler now with this block. Ram sinks dont even feel warm they are chill. i think the core is limited to volts. im going to try to up the volts to 1.5 threw bios when i get a chance and see if it helps. OH also im using my D4 pump and evo 360 rad with trip 130cfm deltas on high. almost forgot my temps are weird. i run idle 31c and under load 31c. thats correct its no typo 31C. but if i turn off 2 fans the temps rise fast up to 50c. idk it works for me i guess lol
 
turbolude said:
i run idle 31c and under load 31c. thats correct its no typo 31C. but if i turn off 2 fans the temps rise fast up to 50c. idk it works for me i guess lol
Click to expand...

Seems a little fishy that your temps don't raise any
 
very

my neptune gets between 45 and 49C when OCed from standard of 430MHz to out-of-box OC of 490MHz then more OCed by me to 525MHz. RAM is like 1350.
 
turbolude said:
just installed my maze 4 gpu block and swiftec ram sinks onto my BFG 7800gt.

im oced to 510/1.23

for some reason i can oc mem higher then ,my core on this card comepared to my last evga. wich hit 515-520/1.20 but this card runs extremely cooler now with this block. Ram sinks dont even feel warm they are chill. i think the core is limited to volts. im going to try to up the volts to 1.5 threw bios when i get a chance and see if it helps. OH also im using my D4 pump and evo 360 rad with trip 130cfm deltas on high. almost forgot my temps are weird. i run idle 31c and under load 31c. thats correct its no typo 31C. but if i turn off 2 fans the temps rise fast up to 50c. idk it works for me i guess lol
Click to expand...

I've got you beat...

I'm running the system in my signature... but its at 2.70GHz on the processor now, I dropped the 2-2-2-5 timings... I'm idling at 29*C on the processor, 36*C load on the processor, 31*C idle and 31*C load on the 7800GTX... (Thats right, no change either... hell, when I had my air cooler modification, I was idling at 33*C and loading in the low 40's)

GTX > GT

I'm using an RBX on the CPU,
Maze4, OCZ ramsinks, Stasis PWM block on the 7800GTX
DD Black Ice Extreme II w/ two 100CFM 120mm fans pushing
and D5 pump.
 
it's sounding to me like those 7800s have some pretty "l33t" temperature sensors....... :|
 
I think the temp sensors are just fine... I just don't think people realize how shitty of a stock cooler came on the 7800's for the first few months they were out... It was just like one big solid hunk of aluminum with a few fins... so you would see temps idling in the high 40's and 50's on stock cooling, and load in the 70's and 80's...

Hell, I saw my temps drop bigtime just by lapping the PWM heatsink and using some copper thermal paste on it, prior to getting the stasis PWM... that dropped temps almost 10*C... Those power regs got BURNING hot...

The back of the card also got hot with the stock cooler... The back of my GTX feels ambient, just like any other part of the PCB, motherboard, or anything else in the case...

So I don't think its a defunct temp sensor.
 
unless there is something wrong with the sensor, then there should be SOME increase in temps between idle and load. it doesn't have to be as dramatic as you see with the stock cooler, but it should still be there.

i have the same waterblock, half again as much coolant flow, twice your radiator surface area, and close to twice your fan power. i idle around 31 and load at 39 on a card with a lower heat output. it also doesn't feel warm on the back of the card: the thermal conductivity of the PCB and traces are enough to distribute such a minor temp differential pretty evenly across a large area of the card.
 
DFI Daishi said:
unless there is something wrong with the sensor, then there should be SOME increase in temps between idle and load. it doesn't have to be as dramatic as you see with the stock cooler, but it should still be there.

i have the same waterblock, half again as much coolant flow, twice your radiator surface area, and close to twice your fan power. i idle around 31 and load at 39 on a card with a lower heat output. it also doesn't feel warm on the back of the card: the thermal conductivity of the PCB and traces are enough to distribute such a minor temp differential pretty evenly across a large area of the card.
Click to expand...


There doesn't have to be any fluctuation... Simple physics can prove that... So long as your cooling is capable of pulling every ounce of extra heat off of your card as it is generated... And it could be just that in our case (I dont have time to look into the specifics because I am at work).

And two things from your setup...

Half again as much as my coolant flow... the context of this leads me to believe that you are running a pump that has half the flow of mine? If thats the case, I'd say thats something right there...

Twice my radiator surface area... a quadruple 120mm radiator?

and twice the fan power? On the radiator... so four fans pushing or pulling?

From your post... it sounds like you have very low flow throughout your system, which would explain the 8*C temp difference you are seeing on your card...


Now if you meant 150% of my flow, and along with the 4x 120mm radiator... I would simply say overkill for a pelt thats only keeping your 3000 Venice at 18*C (sounds like a 52W)... And if your GPU is in series with your CPU pelt, that could also be where you are gaining heat.
 
yes my temp sensor is fine. when i unplug the pump the temps go up. but this trip rad is lEET i can have all the fans off and it still doesnt change temp. maybe after awhile it might. i also think 7800gtx run cooler then the gt's
 
turbolude said:
i also think 7800gtx run cooler then the gt's
Click to expand...

Don't think so... Its the same core... gt is simply less 4 pipes and slower clocks...

Same core + 4 more pipes + greater clockspeed would = higher temp.... (on stock cooling)

But since they are the same core, this is why we experience the same temperatures.
 
im running a single rad for the cpu loop it maxes at 44-45 @ 3ghz mem is running 2,2,2,5@250 thinking about getting a 2x120 rad but maybe i just need a 200cfm fan to drp temps a few deg.
 
so you are running two separate loops... and a triple 120 for just your vid card?

And good luck finding a 200CFM 120mm fan that runs on a molex... It sounds to me that your water set ups are outside of your case.
 
BLiTzKRiEG said:
There doesn't have to be any fluctuation... Simple physics can prove that... So long as your cooling is capable of pulling every ounce of extra heat off of your card as it is generated... And it could be just that in our case (I dont have time to look into the specifics because I am at work).
Click to expand...
yeah, well, i'm on spring break so i have very little time pressure on me just now, and unfortuately nothing fun to do now that my skiing plans have been shut down.......

i might see if i can dig up the numbers later this afternoon, but here's the situation in words.

any thermal solution that works pulls every gram (no ounces!) of heat out of the part being cooled. if they don't remove all the heat, the temperature just keeps on increasing until the core burns up and shuts down. if there is a difference in the heat being put in and the heat being taken out: the temperature of the object changes. that's the law!

first law: heat will only flow from high temperature region to lower temperature region.

this is true unless you are doing work on the system with a heat pump, which is to say that it's true for a normal water cooling loop.

there will always be a temperature differential across the copper of the waterblock, then another temperature difference from the wet side of the block to the coolant, then another difference from the water to the wet side of the rad, then another from the wet side to the dry side of the rad then still another from the metal of the rad to the air. the temperature differential will increase up until the time that the rate of heat flow in and out equal each other. that's the law!

within the normal opperating temperature range of these parts, the magnitude of the temperature differential from hot side to coolant can be calculated using the part's thermal resistance. thermal resistance is expressed in degrees C per watt of heat load. thermal resistance varies based on temperature differential from hot to cold and coolant flow rate. within a given high-flow loop, flow rate is fixed and coolant temp varies by less than 1C. this isn't the law, but it's how parts are tested and rated.

going from idle to load, the watts of heat that the block has to deal with increase, and so the temperature differential across the block also must increase by an amount equal to the blocks thermal resistance multiplied by the difference in heat load.

dangerden doesn't publish numbers on their products, so from the competition: http://www.swiftnets.com/products/mcw55.asp

i might look up the nVidia white papers later, in order to get correct numbers, but supposing an incread in thermal load of 30 watts from idle to load, and a coolant flow rate of approximately 1.5 GPM (your loop should be in the 1.0 - 1.5 GPM range) that would be a temperature increase of 1.5 C over and above any increase in coolant temps, wtih a very similar swiftech block.

BLiTzKRiEG said:
Half again as much as my coolant flow... the context of this leads me to believe that you are running a pump that has half the flow of mine? If thats the case, I'd say thats something right there...
Click to expand...
you should read more semi-old books. half again, 1.5 times, 50 percent more, 150% is what i'm saying. i have a much higher flowing pump with just a touch less head. supposedly 800 GPH max flow and 14 ft. max head.

BLiTzKRiEG said:
Twice my radiator surface area... a quadrupile 120mm radiator?
Click to expand...
that would actually be two BIX2 rads, rather than a single quad.

BLiTzKRiEG said:
and twice the fan power? On the radiator... so four fans pushing or pulling?
Click to expand...
two fans pulling on each rad. two U1A panaflos on one and two L1A panaflo's on the other.

BLiTzKRiEG said:
From your post... it sounds like you have very low flow throughout your system, which would explain the 8*C temp difference you are seeing on your card...
Click to expand...
already addressed this. it's a VERY high flowing loop. one of the few things i could do to push it higher would be to stick a killer iwaki pump in there.

BLiTzKRiEG said:
Now if you meant 150% of my flow, and along with the 4x 120mm radiator... I would simply say overkill for a pelt thats only keeping your 3000 Venice at 18*C... And if your GPU is in series with your CPU pelt, that could also be where you are gaining heat.
Click to expand...
trust me, this isn't overkill: pelts are just that inefficient when you start to put them under heat load, which is why phase is so much better if you have the money for it. if i knock things back to stock speeds, with this 226 watt pelt, i load at -14 C as a result of the reduced heat load on the pelt. coolant temps are about the same.

i fold when i'm not gaming. there is minimal difference in heat load from my proc/pelt going from not gaming to gaming: any difference in GPU temps are the result of increased heat load from the GPU, and coolant temps are once again about the same.
 
BLiTzKRiEG said:
so you are running two separate loops... and a triple 120 for just your vid card?

And good luck finding a 200CFM 120mm fan that runs on a molex... It sounds to me that your water set ups are outside of your case.
Click to expand...
they're out there if you can deal with the noise: http://www.sidewindercomputers.com/de12gf.html

there are others as well, one of which moves even more, is thinner, but a whole lot louder.
 
ah, now I see the mental pic of your setup...

I am still confused on your stock load temps vs your OC load temps with that 226W pelt...

a 32*C jump from a .14V increase seems like alot... is the pelt really that inefficient?
 
THATS THE JET ENGINE DELTA im looking at the 220cfm normal size 120x38mm but sold out at sidewinder. my pc is pretty noisy right now as it is, 5 120mm fans 4 of wich are 130cfm running on high untill i get myself a fan controler
 
BLiTzKRiEG said:
There doesn't have to be any fluctuation... Simple physics can prove that... So long as your cooling is capable of pulling every ounce of extra heat off of your card as it is generated... And it could be just that in our case (I dont have time to look into the specifics because I am at work).
Click to expand...

Actually according to simple physics this is completely wrong.
 
best [486] said:
i have a 80MM one of them and it only cost me 5$ at ax-man
Click to expand...
good for you......that's very nice.......you didn't buy the same fan, now did you?

delta makes MANY fans, at many performance levels, suited for a variety of uses. tonnes of companies use OEM delta fans in their products. you can probably get two or three delta fans for free by stripping apart a dead computer. none of those fans can put out 220 CFM, and neither can the one that you picked up for $5.

i have plenty of delta fans myself.......from time to time i have thought of making myself a hovercraft with them. cheap, little delta AFB fans aren't much to talk about.
 
My 4200 X-2 runs 33 to 36c depending on ambient while running 2 instances of folding at home 24/7.

I’m using a Koolance Exos at half speed. The Cores are OC’d to 2.7 per side. Not too shabby.

My 3800 X-2 at 2.5 per side runs 48c in the same room with an XP-90 with a ThermalTake CAT-90. Again, running two instances of folding at home 24/7.
 
You must log in or register to reply here.
Back
Top