Athlon XP3200 Corsair 1001

[iced98lx]

I've installed the Corsair 1001 system on my XP3200 and I"m haveing drat for luck on temps. Currently the system will only idle at 47-50.. which is rather dissapointing. Also, when did the fans switch from panaflow to the other one?? I was waaay dissapointed when I heard this fan run.

 

anyway: the system has great flow, but I don't think it's getting good contact with the cpu? I had the springs switched around (soft opposite the bump) but I've corrected that with no gains in temps. I've reapplied thermal paste, etc and am ready to give up and go back to my air cooling, it was just as loud as this, but actually worked. It sounds like there may be air in the system (slight noise from pump on start up, which quickly goes away), but I've tipped and powercycled etc and the liquid isn't frothy..

 

annnny suggestions appreciated.

 

chris

[lhyde]

Check the tubing for kinks or narrowed sections. Any restrictions to fluid flow?

[iced98lx]

flow is great, I can monitor the system in the bios until it gets to 50 degrees or so, shut it down, pull off the block and hte copper is cool to the touch, where as anything else on the board is hotter than all get out. Obviously I"m just missing something mounting wise, but I've done it 20 times if i've done it once.. I've mounted countless cooling solutions, but this one seems to be giving me trouble...

 

chris

 

EDIT: this may help- when I run the system with the fan off it idles at the same 47 in windows as with the fan on.. (radiator fan)does this point to bad flow or bad block to cpu?? has anyone with a K7 had to use a coldplate/shim?

 

 

chris

[lhyde]

My kits arrived to water cool the AMD Barton 2800's. Also a P160. I'll let you know in a day or 2 if my conversion went ok. Gonna move one system from a 24" tower to the P160. Then in a week change the last PC to WC in a Lian Li 6070A.

With my aircooled AMD I've not used a shim.

The fact that you can't feel heat on the CPU block, but the CPU gets hot; makes my wonder if you've got tight contact. The surface area of the XP being so small,(compared to a P4) it wouldn't take much of slight tilt of the block to be missing most of the chip.

[iced98lx]

My kits arrived to water cool the AMD Barton 2800's. Also a P160. I'll let you know in a day or 2 if my conversion went ok. Gonna move one system from a 24" tower to the P160. Then in a week change the last PC to WC in a Lian Li 6070A.

With my aircooled AMD I've not used a shim.

The fact that you can't feel heat on the CPU block, but the CPU gets hot; makes my wonder if you've got tight contact. The surface area of the XP being so small,(compared to a P4) it wouldn't take much of slight tilt of the block to be missing most of the chip.

my concerns exactly, but the compound seems to leave a nice uniform rectangle.. (not that this couldn't occur when I was holding it up to seat it)

 

I'm thinking a very thin copper cold plate might be in order..

 

dern it all, I purchased this sytem off the review of http://modthebox.com he ran it with an athlon xp3200 and got wonderfull results.. oh well!

 

back to troubleshooting!

 

Chris

[lhyde]

I think when I do my conversion; I will add the CPU block with the mobo out of the case. This way I can maybe see gaps, or maybe use a feeler gauge to see if there is space between CPU and Block.

[iced98lx]

I think when I do my conversion; I will add the CPU block with the mobo out of the case. This way I can maybe see gaps, or maybe use a feeler gauge to see if there is space between CPU and Block.

I'll slide my mobotray out and see if I can get any paper etc between.. just check the gap.. otherwise I"m out of options!

 

chris

[KaptCrunch]

I'll slide my mobotray out and see if I can get any paper etc between.. just check the gap.. otherwise I"m out of options!

 

chris

wear a anti-static strap an touch the paper to case first, by the way whats your room temp at?

[iced98lx]

room temps are 70-72 degrees F.

[lhyde]

Well, I've just spent the better part of a day transfering my A7N8X-E-Dx/Barton 2800 from a 24" tower that had poor airflow; to my new Antec P160. While doing this conversion, I added a CWC100-1001. With the mobo out of the case, I put the CPU block on. I could not see any gap between the block and CPU with a light behind it; nor, could I get a feeler gauge to slip into there. So I assume the block is tight to the chip. Before the change to water, I was getting 51*C under load. Now with water added, for the past hour I've watched the temp go from 39*C to 50*C under the same load. Case temp has stayed the same at 32*C. Room at 24*C

 

Iced98 ... I too am disappointed. With the P4 conversions, the temp dropped 13*C. With the AMD chip almost nil change. The only thing I can think of ... is that the P4 being 1.25" square makes a bigger contact patch on the block and therefore cools better. The AMD chip has such a small contact patch, the heat isn't being bled off as easily. I'll see how the next conversion goes; but right now I suspect the same results. It's a Barton 2800 on a A7N8X-Dx in a Lian-Li 6070A case.

[iced98lx]

well I had a revelation.. I reseated it, and suddenly.. it worked.

 

not spectacularly, but decent.

 

I idle at 44 and load to 53~ish. (48 if the room is cool, 65-70, 53 is room temp at 80) My inbox temps are pretty high (45 when cold room 50 when warm room) but that's just because I have one 80mm @ 7v and a 40mm @ 7v for my graphics card... anyway moral of the story I now have a large water loop in my system I have to dump another $100 into to make work like some reviews said it would.. I'm still going to get a shim and a coldplate.. just to see.. otherwise a new block set and some new tubing will be in order.

 

chris

[lhyde]

I completed the conversion on the last PC. A7N8X-DX/Barton 2800. Before conversion to Corsair Cool: CPU temp 50*C; MB 33*C; room 23*C. After Conversion: CPU 40*C; MB 30*C; room 23*C.

This seems to point to the maybe poor contact I have with the the other PC with a Barton 2800. It climbed to 51*C. Basically what it was on air cooling. I will try a reseating of the block and see if I can get better contact.

[KaptCrunch]

ithinks u got air locks in the system.

[lhyde]

Thanks CaptC: I'll check that out again. I did burp it pretty well and did the 45* tilt thing again, but I did notice it didn't take as much fluid as the other installs. I did check for kinks and flat spots but it is okay that way. With the small contact patch on a AMD; it wouldn't take much of a tilt on the block to make poor contact.

[iced98lx]

Thanks CaptC: I'll check that out again. I did burp it pretty well and did the 45* tilt thing again, but I did notice it didn't take as much fluid as the other installs. I did check for kinks and flat spots but it is okay that way. With the small contact patch on a AMD; it wouldn't take much of a tilt on the block to make poor contact.

 

Exactly- I moved my PC into the house after working on it in the garage and BAM idle temps are back to 46-48 instead of 44. I'm going to get a shim and maybe a small coldplate and see if that helps..

 

as for the airlocks, I tilted at 45 degree angle, and ran it for... a long time, and then turned it on and off.. and on and off... lol I can't get anymore air out of it, is there something I'm missing? My stupid incase temps are rising over 50 on full load, and up till this moment the machine has refused to stay on for more than three minets at a time without hard rebooting.. I've upped my fan voltage to 7v from 5v, hopefully this will keep the NB cooler... if not I'll add a 2nd 7v fan. suggestions welcome, as always.

 

 

Chris

[lhyde]

Surprise... Surprise. There was more air in the system. The rad was half full. Originally, when filling the system, I would lay the rad flat on the table barbs up; pump the fluid thru and then mount the rad again to the back plate. Tonight I unhooked the rad again, but instead of laying it flat I let it hang so the barbs were up at the top. Burp..Burp big air bubble. I could add another third of a bottle of water to the tank. It takes a fair amount of fluid to totally fill that rad. I'm monitoring the temps again. They're still not as low as I'd like, or as low as todays conversion. Next I will readjust the CPU block; but that won't be until after the weekend. Got to travel for the next 5 days.

Ta-ta 'til then.

[iced98lx]

well that makes sense, my rad was filled while it was mounted, barbs down.

 

*smacks head* duh. I'll have to take it off and spin it around... (good thing I left some excess tubing!) see if any air comes out... would that significantly help my temps if the coolant is cool anyway though? We'll find out.

 

also, how much did your temps go down???

 

 

Chris

[iced98lx]

well that makes sense, my rad was filled while it was mounted, barbs down.

 

*smacks head* duh. I'll have to take it off and spin it around... (good thing I left some excess tubing!) see if any air comes out... would that significantly help my temps if the coolant is cool anyway though? We'll find out.

 

also, how much did your temps go down???

 

 

Chris

 

okay I just inverted my rad completely, no air came out. (I was dissapointed)

 

here's a question- when you hook your tach signal wire up to your MB header, what RPM does it tell you? My pump is currently registering 35xx and I thought when I originally plugged it in and monitored it the reading was more like 5xxx... maybe i'm out of my mind or looking for an answer as to why this isn't working..

 

 

chris

[lhyde]

I've been away for a few days. My CPU temp dropped from 51*C to 48*C. Not as much as I had hoped. My pumps (all 4) all run about 3540rpm. They vary a little as the power supply 12v fluctuates; but never below 3500, and never above 3575.

[iced98lx]

mines gotten better with a Shim, if you don't have one, get one. Evidentally the block doesn't sit level with out it.. I'm idling right now at 44 that's without the fan on the rad on.. (hehe, I like silence) but it's also fairly chilly in here (room temp of 71~ish.) I've been loading to 48-50 depending on what I"m doing.. games are all around 46-48 prime 95 with the max-heat setting hits 50... my ambient in-case temp is 52.. I'm not convinced of the temp readings off this board though.. my bios reads 45 idle directly after turning the computer on, where as speedfan reads 35 when it first boots up...

 

 

overall still not what I was sold, but hell I"m stuck with it now, I guess I'll look for a better block and quit whinning it was cheap.

 

Chris

[lhyde]

I reseated the block today. (A7N8X-E-Dx) Not much difference ... now running at 47*C under load. So what do I need to order? A shim for Socket A or Athlon XP? The other Barton 2800 is still at 39*C, same load conditions; but a slightly different mobo(A7N8X-Dx).

[iced98lx]

I reseated the block today. (A7N8X-E-Dx) Not much difference ... now running at 47*C under load. So what do I need to order? A shim for Socket A or Athlon XP? The other Barton 2800 is still at 39*C, same load conditions; but a slightly different mobo(A7N8X-Dx).

 

 

well that's a pretty good load figure as far as I can see... The SHIM i'm refering to is just peice of copper or non-cunductive material (cardboard) you can order online.. just make sure you get one that pretty much so just has the outside ring, some of them that have cross peices down the middle interfeare with barton core, like the copper one I ordered... I just had to snip a few peices out and smash it flat again... if you don't mind doing that svc.com has the copper ones for $3.xx shipped USPS 2 days and you'll ahve it.. I can post a diagram of what I cut out.. that helped my temps a few degrees, and also adds some peice of mind as you're installing, it gets hard ot crush the core.. allthough I"ve never done it, or even seen someone who has, so I don't know what the big fuss was about.

 

anyway, are you overclocking?

do you have the latest bios in each one?

that seems really fishy, could there still be air or other issues in the one?

(I'm stilll looking for air in mine, i'm not convinced it's correct)

what are you using for load temps? A better measure might be prime95 torture test using "Blend maximum heat" which does pretty muchso what it says... then we'd be comparing apples to apples.

 

 

Chris

[lhyde]

I found an outlet for ThermalTake AMD shims. ($4.00 CDN) I ordered 2.

I'm not OC'g the Bartons; but I have 2 P4's that are OC'd at 15 percent. I run SETI@Home CLI on all PC's 24/7/365. CPU usage shows 100 %. I don't think my problem is air. The fluid in the piping is clear, not foamy looking like a "margarita". I have the up to date BIOS's in all PC's.

A7N8X-E-Dx Load = 47*C; mobo = 28*C; ambient Room = 22*C

A7N8X-DX Load = 37*C; mobo = 27*C;

P5P800 P4 530J OC'd to 3.54g Load = 39*C; mobo = 33*C

P5P800 P4 540J OC'd to 3.68g Load = 39*C; mobo = 31*C

I've never tried Prime 95; but I'll "google" it and try to find a copy to use. The SETI should be okay as it crunches work units using 100 % of the cpu cycles. I have no idea what idle temps would be unless I stop SETI for a 1/2 hour; which I may do just for comparison info.

 

Larry

 

UPDATE: CPU idle temps. Bartons don't show much change ; because,(?) they're not OC'd, or other apps running in bachround ??

E-Dx = 46*C

Dx = 36*C

530J = 32*C

540J = 30*C

[lhyde]

Well I added the shim to the Barton that runs hotter. Temps actually got hotter (55*C) under load. and are about 46* at idle. Using ASUS probe and MBM 5, temps read the same, but that is expected, because they read from the same sensors or diode. I will try to find another way of measuring temps. There are 2 sensors that come with the P160 case and I'll tape one to the CPU waterblock. Perhaps the mobo sensors are reading high. I will take the shim out again and try the reseat once more.

Updates later....

[lhyde]

I removed the shim and reseated the waterblock with a Antec P160 sensor under the ridged area of the waterblock. Using ASUS probe the CPU temp is 46*C at idle and 52*C under load (Prime95 & SETI), BUT; the Antec sensor is reading 39*C !!!, which is the same as the other Barton 2800. I think the

A7N8X-E-DX CPU sensor is not accurate. The shim showed marks around the edges, so I know the block was very tight against it.

I will not concern myself with CPU temp readings using ASUS probe on this mobo any longer. I am satisfied that the CPU with the watercooling is now in the same range as the other watercooled Barton CPU on the A7N8X-Dx.