-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
S939 A64 IHS Removal
- Thread starter Jokester
- Start date
Soldato
Ok my temps have dropped.
but Im getting one of the core very rearly jump to 206c
43 37
43 37
42 38
42 37
43 38
42 39
42 38
43 38
43 206
44 38
43 38
43 37
43 38
44 38
40 35
39 35
39 34
Now this is under S&M testing. I have tightened the screws up as much as I can by hand.
but Im getting one of the core very rearly jump to 206c
43 37
43 37
42 38
42 37
43 38
42 39
42 38
43 38
43 206
44 38
43 38
43 37
43 38
44 38
40 35
39 35
39 34
Now this is under S&M testing. I have tightened the screws up as much as I can by hand.
Soldato
dont worry about thatClairvoyant said:
what version of coretemp are you using ?
Soldato
Final8y said:dont worry about that
what version of coretemp are you using ?
0.94
Soldato
Efficiency drops the hotter the chip gets, so you ether up the volts to make up for it at a given speed or find a away to get rid of more of the heat but you can end up in a catch22 Vcore add yet more heat as most know.
At low speed you can get away with higher temps & still be stable because the chip does not need to be as Efficient so can afford to waste some as heat.
I was doing a high Mhz Orthos run at low Vcore & failed with in 30min temps only peaked at 48c, i got the desk 12" fan out & aimed it at cooling towers & got peak temps down to 42 & passed Orthos at same settings.
At low speed you can get away with higher temps & still be stable because the chip does not need to be as Efficient so can afford to waste some as heat.
I was doing a high Mhz Orthos run at low Vcore & failed with in 30min temps only peaked at 48c, i got the desk 12" fan out & aimed it at cooling towers & got peak temps down to 42 & passed Orthos at same settings.
Last edited:
Soldato
I just cant get passed 2.7
El Jimben said:Maybe I need to improve my cooling then as my Opty just won't play ball above 2.8GHz despite ridiculous vcore, RAM not being stressed etc
It might help a little to remove IHS. On my x2 3800+ I can now run at 2.88ghz and about 1.57-1.58V according to software readings. Earlier I could run it at 2.8ghz and 1.55V, and even with high end water cooling core temp went up to 60c! (sixty!)
Everybody with proper cooling and 939 cpu should pop it, even if it only brings down the temps 
Soldato
Did anyone with a high temp difference between cores@load on an X2 notice that they were closer together after removing the IHS ?
I'm seriously considering doing this as my 4200+ core temps are 11-13c apart in Orthos@load - core 1 is 40-42c, core 2 is 53-55 - when I'm only running at 230x11 / 1.375v. This is with a Scythe Mine / 120mm 80cfm Evercool fan. Both cores idle around 31-34c, the 2-3secs after I start Orthos, core 2 jumps 20c to 53c ( core 1 goes to about 38c then slowly gets to 40'ish )
Only concern, apart from killing my cpu, is that the Scythe Mine will need tweaked to make good contact with the cpu, as it uses clips to fix to the cpu socket
I'm seriously considering doing this as my 4200+ core temps are 11-13c apart in Orthos@load - core 1 is 40-42c, core 2 is 53-55 - when I'm only running at 230x11 / 1.375v. This is with a Scythe Mine / 120mm 80cfm Evercool fan. Both cores idle around 31-34c, the 2-3secs after I start Orthos, core 2 jumps 20c to 53c ( core 1 goes to about 38c then slowly gets to 40'ish )
Only concern, apart from killing my cpu, is that the Scythe Mine will need tweaked to make good contact with the cpu, as it uses clips to fix to the cpu socket
Associate
- Joined
- 11 Jan 2006
- Posts
- 186
WHY
This might sound like a dumb question, but why would you want to remove the metal thing of the top of the CPU? i mean what is there to gain?
This might sound like a dumb question, but why would you want to remove the metal thing of the top of the CPU? i mean what is there to gain?
Caporegime
Pandorasghost said:
Better heatsink contact, lower temps, higher overclocks
The IHS can have poor contact, which means no matter how good your IHS-Heatsink contact is, you still may get rubbish temps. It also is a barrier to heat transfer.
MoodyB said:Did anyone with a high temp difference between cores@load on an X2 notice that they were closer together after removing the IHS ?
I'm seriously considering doing this as my 4200+ core temps are 11-13c apart in Orthos@load - core 1 is 40-42c, core 2 is 53-55 - when I'm only running at 230x11 / 1.375v. This is with a Scythe Mine / 120mm 80cfm Evercool fan. Both cores idle around 31-34c, the 2-3secs after I start Orthos, core 2 jumps 20c to 53c ( core 1 goes to about 38c then slowly gets to 40'ish )
Only concern, apart from killing my cpu, is that the Scythe Mine will need tweaked to make good contact with the cpu, as it uses clips to fix to the cpu socket
Yeah, closer together but not quite exact same temps. Could be because the actual core isn't 100% flat as well... Funny thing though, with my last x2 3800+ after removing IHS the hottest core was now the coolest and visa versa
Final8y said:
I have this arriving soon, if customs pull their finger out.... (this is the actual CPU that's on it's way to me, so same stepping as you).
It'll be going under one of these -
Anyone prepared to say "Yes, that will still sit nicely on the core" or "OMG noes! The ingines cannae take it cap'n"?
Also..... what speeds are you getting out of your overclock (assuming you have it overclocked) and what voltages are you needing to put through it?
Well, it arrived, and it's running hot. So I realised I have a way to safely test.
I have a spare AN8 Fatal1ty board, and now I have a spare 146, so if I took the IHS off the 146 and fitted it to the spare board, that would show me how easy/difficult it was to remove, give me some idea of how risky it was, and also show me if the core sat higher than the socket, or lower.
Initially I used a knife very much like the one Jokester used in the first post. It was fine for two of the edges, but for the other two it was really struggling. So I took a couple of pairs of pliers to a disposable razor, and got myself a paper thin blade. It worked a treat.
Turns out it's not too difficult removing the IHS, and pretty much it's easy to control how deep the blade is going in. Only part of the rubber is sealed to the CPUs PCB. But when I'm cutting it, I'm cutting through a wider part, so I found I could go round, as deep as I was comfortable with and the IHS popped off afterwards, I was miles away from the little components on it.
So I stuck the CPU into the ZIF on the AN8 and using a match I found out which was higher - the core.
I've ordered some MX-1, as well as waterblocks for the mosfet, so tomorrow I'll be stripping everything down, taking the IHS off, I might even lap the CPU waterblock a bit more.
Edit : Also the IHS being referred to as a cap is a bit misleading. It's a thick lump of aluminium. I was expecting it to be hollow, but no, it's only slightly slimmer in the middle than on the outsides. It really is a heatsink.
Edit : At stock volts, speed etc the 185 is running at 37 on one core and 30 on the other. I'll see tomorrow how much that changes.
I have a spare AN8 Fatal1ty board, and now I have a spare 146, so if I took the IHS off the 146 and fitted it to the spare board, that would show me how easy/difficult it was to remove, give me some idea of how risky it was, and also show me if the core sat higher than the socket, or lower.
Initially I used a knife very much like the one Jokester used in the first post. It was fine for two of the edges, but for the other two it was really struggling. So I took a couple of pairs of pliers to a disposable razor, and got myself a paper thin blade. It worked a treat.
Turns out it's not too difficult removing the IHS, and pretty much it's easy to control how deep the blade is going in. Only part of the rubber is sealed to the CPUs PCB. But when I'm cutting it, I'm cutting through a wider part, so I found I could go round, as deep as I was comfortable with and the IHS popped off afterwards, I was miles away from the little components on it.
So I stuck the CPU into the ZIF on the AN8 and using a match I found out which was higher - the core.
I've ordered some MX-1, as well as waterblocks for the mosfet, so tomorrow I'll be stripping everything down, taking the IHS off, I might even lap the CPU waterblock a bit more.
Edit : Also the IHS being referred to as a cap is a bit misleading. It's a thick lump of aluminium. I was expecting it to be hollow, but no, it's only slightly slimmer in the middle than on the outsides. It really is a heatsink.
Edit : At stock volts, speed etc the 185 is running at 37 on one core and 30 on the other. I'll see tomorrow how much that changes.
Last edited:
sablabra said:How's progress going? Guess you ruined your spanking new CPU?
It's not a lump of aluminium, it's a lump of copper
Nah, place I get my watercooling blocks from was 1 block short for the mosfets. I'm not stripping everything down twice, so I'll wait on the block arriving.
And yeah, it's a lump of copper - but the point is that it's a great big lump. not just a thin cap.
Well, I've been somewhat successful.
First of all I'm still waiting on a second mosfet block. But I decided to remove the IHS without stripping everything down.
Undid the nuts on the two bolts, pulled and pulled and pulled the waterblock up on the pillars - and it ripped the CPU out of the ZIF, not a nice surprise for me there. Turns out it was just thermal grease that was doing it, and it was just suction. Thankfully no harm done, no pins belt, a couple of them look shiny, but no damage done.
The IHS came off much easier this time. Definately the thing to use is a blade from a disposable razor. Being a bit more paranoid about doing it on a 185 than a 146 I taped up the razor, to give myself some sort of guide as to how far it could go in. It worked a treat, I went round it once, and then the second time I stuck it in as far as the tape, and dragged it all the way around, IHS fell off.
No damage to the core, no damage to the components mounted around the core, and no pins bent.
Problem 1 happened when I tried to reseat the waterblock. The waterblock has 2 bolts, which form pillars, there's a spring on each pillar and a nut which tightens up. Previously with an IHS I couldn't go wrong, I just tightened the bolts and I was away. Now with just the naked core it's possible for the block to sit at an angle, and very difficult to tell if it is or not.
On my first boot I had a 15 degree difference between cores, I slackened off both nuts and gave it a bit of a shoogle, and it's now even under load, and 5 or so difference at idle. Seems that the springs are enough pressure to get it even, I just need to make sure I don't overtighten them, or they can't do their work.
I've used MX-1 for the first time. Using pillars I was not entirely convinced their application method would work - you throw a dod on in the middle, and sit the heatsink down on it, and then shoogle it and it's done. Since I wouldn't be able to twist the block at all, I spread the grease over the core a little. Supposedly as it "sets" over the next few days I'll see the temperatures go down.
The end result as far as temperatures go was moderate at idle, about 8 degrees. But at load it was a massive decrease, about 20-25 degrees on each core.
The problem I have now is that at 1.525 volts and 3Ghz speed the VRM motherboard sensor hit 80 degrees. That's with the "OTES" fans (noisy, noisy noisy) on at full pelt. I'm still waiting on the other mosfet block, so hopefully when that arrives I can get the VRM temps down, and try overclocking it properly.
Edit : Finally done.
At idle - CPU 22 degrees. PVM 30 degrees. System 28 degrees.
First of all I'm still waiting on a second mosfet block. But I decided to remove the IHS without stripping everything down.
Undid the nuts on the two bolts, pulled and pulled and pulled the waterblock up on the pillars - and it ripped the CPU out of the ZIF, not a nice surprise for me there. Turns out it was just thermal grease that was doing it, and it was just suction. Thankfully no harm done, no pins belt, a couple of them look shiny, but no damage done.
The IHS came off much easier this time. Definately the thing to use is a blade from a disposable razor. Being a bit more paranoid about doing it on a 185 than a 146 I taped up the razor, to give myself some sort of guide as to how far it could go in. It worked a treat, I went round it once, and then the second time I stuck it in as far as the tape, and dragged it all the way around, IHS fell off.
No damage to the core, no damage to the components mounted around the core, and no pins bent.
Problem 1 happened when I tried to reseat the waterblock. The waterblock has 2 bolts, which form pillars, there's a spring on each pillar and a nut which tightens up. Previously with an IHS I couldn't go wrong, I just tightened the bolts and I was away. Now with just the naked core it's possible for the block to sit at an angle, and very difficult to tell if it is or not.
On my first boot I had a 15 degree difference between cores, I slackened off both nuts and gave it a bit of a shoogle, and it's now even under load, and 5 or so difference at idle. Seems that the springs are enough pressure to get it even, I just need to make sure I don't overtighten them, or they can't do their work.
I've used MX-1 for the first time. Using pillars I was not entirely convinced their application method would work - you throw a dod on in the middle, and sit the heatsink down on it, and then shoogle it and it's done. Since I wouldn't be able to twist the block at all, I spread the grease over the core a little. Supposedly as it "sets" over the next few days I'll see the temperatures go down.
The end result as far as temperatures go was moderate at idle, about 8 degrees. But at load it was a massive decrease, about 20-25 degrees on each core.
The problem I have now is that at 1.525 volts and 3Ghz speed the VRM motherboard sensor hit 80 degrees. That's with the "OTES" fans (noisy, noisy noisy) on at full pelt. I'm still waiting on the other mosfet block, so hopefully when that arrives I can get the VRM temps down, and try overclocking it properly.
Edit : Finally done.
At idle - CPU 22 degrees. PVM 30 degrees. System 28 degrees.
Last edited:
just removed mine, taken loaded temps from almost 70 (both cores) to about 55 both cores. temps were equal for cores +/- 2 degrees since lapping (but not really any cooler)
orthos stable at 1.525v gone from 2920 to 3030ish
110 mhz - cpu is 939 4800+ lcb9e
orthos stable at 1.525v gone from 2920 to 3030ish
110 mhz - cpu is 939 4800+ lcb9e
Jokester, just a thought. With both AMD and Intel now producing their CPUs with soldered/welded/glued/whatever it is IHS, which are not removable in the normal way. You might want to put something in the very first post here, in case some poor blighter reads it and gets a knife out and tears the core from the CPU.
The Halk said:
I changed the title to S939 A64 IHS Removal a few months ago so that's about a prominant as it'll get short of saying S939 in giant letters
.Jokester