Well, just thought I'd share some observations now that the IHS has been off for a while.
As I mentioned earlier, I dropped about 12'C at load, which is quite impressive. Unfortunatly, I've got a mild case of the 'cold bug' now as a result.
I'm very limited in how high I can OC my memory now, and I can not exceed 300MHz base HTT by much. If I push the memory beyond about 230MHz, I get nasty hard-locking when the PC is at idle. At 3.3GHz, it seems to occur no matter what I set the memory to. Seems fine as long as the temperature of ths chip is above -15'C or so. F@H helped to solve that problem, but as soon as it finished a WU and downloaded the next, I got the nasty hard-lock. Thats how quick it occurs. Cranking the vapo heaters to 100% did help, but not by much when it is at 11x303MHz. It seems ok at 3.2GHz (11x291 with a 5/6 divider for 228MHz ram). I used to have no problems at 3.3GHz prior to removing the IHS, but I used to see temps in excess of +8-9'C at load.
At idle, it drops down to -27'Cish. This is just another example proving that the cold bug stems from the integrated memory controller in the 90nm chips.
Looks like this thread is going to turn into the 'I put my IHS back on my Opteron' thread
Since I really didn't gain anything by removing the IHS (aside from better temps, and instability), it is time to figure out how to put it back on. I'm going to take advantage of the fact that it is off, and put some dielectric on the PCB before I re-cap it. I'm pretty confident that some 'cold friendly' TIM should help be beneficial under the IHS, and hopefully it'll raise my temps enough to keep the 'bug' at bay. Stay tuned for some pics..
I read somewhere that it can be reapplied with automotive gasket silicone ...high temp. What were you planning to use?
I read somewhere that it can be reapplied with automotive gasket silicone ...high temp. What were you planning to use?
I'm not too sure yet, although I'll probably want 'subzero' silicone, not the high temp stuff .. I'm actually not too concerned. Im thinking of using a clamp to apply pressure (careful not to bend the pins of course) and then secure it with a bit of silicone on the corners. There will be a significant amount of pressure on it from the evaporator, so that thing wont go anywhere.
I have to remind myself not to do too good of a job, as I need to regain at least 10'C to beat this bug dead. Never thought I'd actually find myself trying to increase the temps :shakehead
As a former mechanic... I would think dielectric on the caps and engine gasket should do the trick. What do you plan on using on the core, artic silver?
As a former mechanic... I would think dielectric on the caps and engine gasket should do the trick. What do you plan on using on the core, artic silver?
I actually have two options, Arctic ceramic and the Asetek vapochill branded interface goop. Both are non-conductive and approriate for sub zero conditions. I remember hearing that the silver varieties do not work well in cold conditions, so I have avoided using them. The vapo goop is the same stuff that I used to plug all of the socket pins, and it is a good consistency that makes it easy to apply/remove.
The last time I used arctic cermaic, I tried to remove a TDX water block, and the stuff was so sticky that it actually ripped the CPU right out of the socket with the locking lever down never again I swore
So what are the tell tale signs of the "cold bug"?
Is there a remedy?
Perhaps food for your OC article chapter 2?
Definitely will get some pics posted
The cold bug is a mysterious phenomenon, and it actually takes several different forms. In the most extreme cases, it comes as a no-post, plain and simple.. freezes at post, or doesn't even get that far. Quite a few of the Winchester based cores exhibited this. The Revision 'E' procs were greatly improved, and lots of them can operate quite well at single-stage phase change temps. One of the tell-tale signs is the CPU's refusal to operate at higher memory frequencies, or higher base HTT frequencies (like my situation). Often, you'll see people with FX series processors using very high multipliers and keeping the base HTT at close to 200MHz, and sometimes lower than that. In my case, I could boot fine at any temp, but as soon as the system started cooling down to -25 or lower, I'd get a total hard-lock. Video remained, but the system was totally non-responsive. It would prime95 for days without issue, but as soon as things cooled down, poof.
There has been a lot of discussion at XS about the bug, but there seems to be a general consensus that it is due to the integrated memory controller. Tony from OCZ believes that it is a 'drive strength' issue. As the chip cools, it tends to 'overdrive' the dimms due to the improved silicon conductivity at low temperatures. Sometimes reducing the drive strength can help, but it did nothing for me in my situation. From what I have seen, the dual core chips and E6 revision procs have the bug worse than the E4's too. It was originally believed that only 90nm processors experienced this, but it has actually been seen with clawhammers, albeit with much lower (LN2) temperatures. Intel chips do not experience this, and this is likely due to the fact that their memory controller is not on-die.
I'm really hoping that the AM2 IMC improvements will eliminate the cold bug. I'd be all over the chips at that time. I also hope that 65nm around the corner does not worsen this situation as 90nm seemed to do.
I'm already contemplating Chapter 2, so I'll keep this in mind
Thanks for quantifying the cold bug issue. I've read a few threads at XS on the subject but kept writing it off because I couldn't pinpoint a singular symptom. I was confused and thought that maybe some were qualifying the bug when it didn't pertain to them ...thinking it only had to do with sub zero cooling.
Now that I understand better ...I can say that I've experienced the same type symptoms as you. I'll fold both procs for 24-48 and shut down to allow to cool before I benchmark. Soon as the temps hit around idle I get a freeze.
One thing I thought of was the hotfix and cooln'quiet drivers installation ...perhaps there was/is an issue at some point where the processors drivers want to kick down voltages and frequencies to cool itself ...perhaps a malfunction in this area. I'm still experimenting and I did have it happen yet again yesterday.
Thanks for quantifying the cold bug issue. I've read a few threads at XS on the subject but kept writing it off because I couldn't pinpoint a singular symptom. I was confused and thought that maybe some were qualifying the bug when it didn't pertain to them ...thinking it only had to do with sub zero cooling.
Now that I understand better ...I can say that I've experienced the same type symptoms as you. I'll fold both procs for 24-48 and shut down to allow to cool before I benchmark. Soon as the temps hit around idle I get a freeze.
One thing I thought of was the hotfix and cooln'quiet drivers installation ...perhaps there was/is an issue at some point where the processors drivers want to kick down voltages and frequencies to cool itself ...perhaps a malfunction in this area. I'm still experimenting and I did have it happen yet again yesterday.
Hmm, that is very odd that you are experiencing those types of symptoms with above ambient cooling. Generally speaking, you should only see this type of behavior at temperatures well below 0'C. Although, perhaps you should try toying with drive strength a bit? could be a dram driving issue if you are having 'cold-bug like' symptoms
Hmm, that is very odd that you are experiencing those types of symptoms with above ambient cooling. Generally speaking, you should only see this type of behavior at temperatures well below 0'C. Although, perhaps you should try toying with drive strength a bit? could be a dram driving issue if you are having 'cold-bug like' symptoms
That's what I'm thinking ...but it acts the same way. It's almost like I've reached a limit of some sort ...it's a tough one to figure out.
Did you try disabling Cn'C, csimon? Many on the OCW forum claim that disabling it in the BIOS helps with OCing stability a lot since it isn't constantly screwing with its vcore.
Definitely you won't want to use a silver containing thermal compound, lemonlime. Way back when I was messing with TEC cooling, I used Arctic Silver 2 on the TEC and after a few months had the TEC smoke it's leads, When I pulled the TEC out of the waterblock/coldplate sandwich, the TEC had some silver that actually plated out on the TEC surface, which I think gave enough conductivity to burn the leads.
Yikes, thats a scary situation mudd. I was always worried about applying the silver compounds. I've never been very fond of them, and I have been using arctic cermaic for the last couple of years, and it did the trick nicely. The goop that comes with the vapo will be my TIM of choice for this reattachment project. It's designed for low-temp use, and I think it will be ideal.
I got the IHS put back on actually, the silicon is curing as we speak. Stay tuned for some pics.
On an unrelated note, I discovered some weakness in my insulation, and as I unmounted the evaporator, I was greeted to a good amount of water all down the front of the mainboard and my graphics card. Looks like this project is going to be more work than I anticipated. Thankfully I applied a ton of goop in the socket and on the chip PCB, so although it was wet, it didn't cause any damage. Going to have to let this thing dry overnight with a fan on it. Ahh the joys of tinkering with things that are working properly
I haven't ever had the plating happen from any Arctic Silver formulation since then; I just think it was a combination of the extreme cold along with the high clamping pressure on the TEC between the waterblock and cold plate that caused the foiling action (see, I remembered what Nevin of AS called it ). I ended up sending the TEC to him so he could study the foiling on it. And I had gotten some free samples from him too. This would definitely be something to worry about with your phase change though, since you run such cold temps and I imagine that the evaporator head clamps down pretty tightly too.
I've never messed with the ceramique myself. How is it compared to AS5?
I haven't ever had the plating happen from any Arctic Silver formulation since then; I just think it was a combination of the extreme cold along with the high clamping pressure on the TEC between the waterblock and cold plate that caused the foiling action (see, I remembered what Nevin of AS called it ). I ended up sending the TEC to him so he could study the foiling on it. And I had gotten some free samples from him too. This would definitely be something to worry about with your phase change though, since you run such cold temps and I imagine that the evaporator head clamps down pretty tightly too.
I've never messed with the ceramique myself. How is it compared to AS5?
Cool stuff, gotta love free samples . I think the silver varieties simply do not maintain their consistency once they reach a certain temperature.
The ceramique is good in the sense that it is non-conductive, and you don't have to be as careful, but the stuff very thick, and extremly sticky. I ripped my sempron right out of the socket with the locking lever closed trying to remove the HSF. Its especially bad when you apply it to something with a larger surface area, like an IHS, it is fine on smaller areas like GPU cores, and barton cores etc. It does do a good job though, I found that it did a little better than AS3 on my old barton rig.
Well, I got it back on, and I'm back in business. Interestingly, even with the reattachment hack-job I did, my temps are still signficantly better than 'out of the box'. AMD's TIM must really be a poor variety.
Looks successful to me. Oddly, my idle temp has also increased a fair bit. I'm at about -22'C as opposed to -27 before. That should be very helpful in combating the cold bug. Some results to come
Here are some pics. As you can see in the first one, the goop literally caked right off when I removed the evaporator. Yes, that is because it was soaking wet. A good deal of water gushed down the front of the mainboard and onto my X850 when I pulled it off :banghead: looks like the insulation weakness was on the hose entering the clamshell. I was careful to shove the insulation further into the clamshell for a better seal. I had to dry everything down overnight with a household fan.
Here is a shot of the chip cleaned up, and my first test fitting. Hard to tell by the photo, but the contact was actually very good. Goop was a little uneven there, but I got it sorted out. I decided to keep the ICS and board inside insulated just in case. That clearly saved me from my last encounter with condensation Not to mention filling the socket holes.
Here is the final product. Basically, I used some pressure to form good contact. I then applied some 'outdoor' all-purpose silicone to the four courners of the IHS. It takes about 12 hours to cure, so I put about 1/2 pound of weight balanced center on the IHS. Once it dried, it was even all around and held down well. I decided against a 'strong seal' simply because this chip will eventually be demoted to aircooling, and at that time I'll be removing the IHS again anyway. Temps turned out to be quite good to my surprise. So hopefully it will be enough to beat the bug.
That is great, lemonlime! And the pics are good too. BTW, is that dielectric grease coated across the top of your processor and proc pcb, or is that ceramique?
I hope that the higher idle temps help to alleviate your cold bug problems.
That is great, lemonlime! And the pics are good too. BTW, is that dielectric grease coated across the top of your processor and proc pcb, or is that ceramique?
I hope that the higher idle temps help to alleviate your cold bug problems.
Its a lot like ceramique actually, but it's the Asetek provided dielectric/TIM that comes with the Vapochill LS. They gave me a fair sized tube, and the stuff works great. Its not nearly as sticky as ceramique, and MUCH easier to clean up than traditional dielectric grease. Since its designed to work well at sub zero temps, I figured it was the best choice
Oh, and COLD BUG ELIMINATED!!
So many things went wrong during this little adventure, but it was worth it. My idle temps sit at around -22'C to -23'C (as opposed to -27'C prior). I was able to successfully surf around the web at 11x303MHz (With my TCCD running at 303MHz 2.5-4-3-6 1T). System didn't hitch once. Prior to this, I'd get nasty hardlocking within 10-15 seconds.
Eeep, the only thing that makes me nervous about all this is the silicon used to reattach the IHS. Are you certain it is the non-corrosive variety, lemonlime? Most standard silicons use acids to air-cure themselves, and when applied to electronics (esp. PCB) can actually eat away at the layers over time.
Most "standard out-door silicons" are of the corrosive variety. Non-corrosive silicons are harder to find and will generally say rather explicitly that they are non-corrosive.
Is it necessary for me to coat the entire wafer that way or did you do it just to raise temps?
Chris, did you get a Vapo or something? He coated it to waterproof primarily.
BTW, my naked Winnie is still alive and kicking. I coated the spot where I skinned it a little from removing the IHS with some of my daughter's nail polish and shoved it into my Dual Sata 2 board yesterday, put an SLK948-U on top and she booted right up. I guess now I'll have to find a refurb A64 board to run it in.
Eeep, the only thing that makes me nervous about all this is the silicon used to reattach the IHS. Are you certain it is the non-corrosive variety, lemonlime? Most standard silicons use acids to air-cure themselves, and when applied to electronics (esp. PCB) can actually eat away at the layers over time.
Most "standard out-door silicons" are of the corrosive variety. Non-corrosive silicons are harder to find and will generally say rather explicitly that they are non-corrosive.
Thanks for pointing that out, Khaos. I didn't check, as I didn't realize that varieties of silicon are acidic enough to cause issues on PCB. IIRC it was 'low odor, indoor/outdoor all purpose silicon adhesive', but I'm going to double check the tube when I get home. I guess its too late to worry too much about it now I'll just hope that constant contact with a -50'C evaporator will slow the corrosion processes
When you applied the silicone, did you get a strong vinegar smell? If so, then the silicone has acetic acid in it and could possibly give problems down the line. But from the looks of your pic, the only metal it's contacting is the IHS itself, which is copper, so I think you are OK. Acetic acid isn't that strong an acid anyways and I don't think it will react with the pcb coating of the proc.
When you applied the silicone, did you get a strong vinegar smell? If so, then the silicone has acetic acid in it and could possibly give problems down the line. But from the looks of your pic, the only metal it's contacting is the IHS itself, which is copper, so I think you are OK. Acetic acid isn't that strong an acid anyways and I don't think it will react with the pcb coating of the proc.
Come to think of it, it did smell a bit like vinegar. It wasn't terribly strong though. I think I'm going to rot away my proc with condensation long before that silicone causes a problem
The processor of all things should be alright. The biggest issue with using silicon to seal components for sub-ambient cooling is on the motherboard itself, where traces appear close to the surface. Acetic acid can eat away at the small layer above those traces and will then corrode the traces, destroying the motherboard.
Not being one to take chances, I use non-corrosive silicon in pretty much all computer hardware / cooling applications. I definitely recommend it if you ever, say, decide to cover the bottom of your motherboard around the socket area with a piece of neoprene that is attached to the motherboard with silicon. *cough* Not that I would ever do such a thing.
Comments
I read somewhere that it can be reapplied with automotive gasket silicone ...high temp. What were you planning to use?
I'm not too sure yet, although I'll probably want 'subzero' silicone, not the high temp stuff
I have to remind myself not to do too good of a job, as I need to regain at least 10'C to beat this bug dead. Never thought I'd actually find myself trying to increase the temps :shakehead
So what are the tell tale signs of the "cold bug"?
Is there a remedy?
Perhaps food for your OC article chapter 2?
I actually have two options, Arctic ceramic and the Asetek vapochill branded interface goop. Both are non-conductive and approriate for sub zero conditions. I remember hearing that the silver varieties do not work well in cold conditions, so I have avoided using them. The vapo goop is the same stuff that I used to plug all of the socket pins, and it is a good consistency that makes it easy to apply/remove.
The last time I used arctic cermaic, I tried to remove a TDX water block, and the stuff was so sticky that it actually ripped the CPU right out of the socket with the locking lever down
Definitely will get some pics posted
The cold bug is a mysterious phenomenon, and it actually takes several different forms. In the most extreme cases, it comes as a no-post, plain and simple.. freezes at post, or doesn't even get that far. Quite a few of the Winchester based cores exhibited this. The Revision 'E' procs were greatly improved, and lots of them can operate quite well at single-stage phase change temps. One of the tell-tale signs is the CPU's refusal to operate at higher memory frequencies, or higher base HTT frequencies (like my situation). Often, you'll see people with FX series processors using very high multipliers and keeping the base HTT at close to 200MHz, and sometimes lower than that. In my case, I could boot fine at any temp, but as soon as the system started cooling down to -25 or lower, I'd get a total hard-lock. Video remained, but the system was totally non-responsive. It would prime95 for days without issue, but as soon as things cooled down, poof.
There has been a lot of discussion at XS about the bug, but there seems to be a general consensus that it is due to the integrated memory controller. Tony from OCZ believes that it is a 'drive strength' issue. As the chip cools, it tends to 'overdrive' the dimms due to the improved silicon conductivity at low temperatures. Sometimes reducing the drive strength can help, but it did nothing for me in my situation. From what I have seen, the dual core chips and E6 revision procs have the bug worse than the E4's too. It was originally believed that only 90nm processors experienced this, but it has actually been seen with clawhammers, albeit with much lower (LN2) temperatures. Intel chips do not experience this, and this is likely due to the fact that their memory controller is not on-die.
I'm really hoping that the AM2 IMC improvements will eliminate the cold bug. I'd be all over the chips at that time. I also hope that 65nm around the corner does not worsen this situation as 90nm seemed to do.
I'm already contemplating Chapter 2, so I'll keep this in mind
Now that I understand better ...I can say that I've experienced the same type symptoms as you. I'll fold both procs for 24-48 and shut down to allow to cool before I benchmark. Soon as the temps hit around idle I get a freeze.
One thing I thought of was the hotfix and cooln'quiet drivers installation ...perhaps there was/is an issue at some point where the processors drivers want to kick down voltages and frequencies to cool itself ...perhaps a malfunction in this area. I'm still experimenting and I did have it happen yet again yesterday.
Hmm, that is very odd that you are experiencing those types of symptoms with above ambient cooling. Generally speaking, you should only see this type of behavior at temperatures well below 0'C. Although, perhaps you should try toying with drive strength a bit? could be a dram driving issue if you are having 'cold-bug like' symptoms
That's what I'm thinking ...but it acts the same way. It's almost like I've reached a limit of some sort ...it's a tough one to figure out.
I got the IHS put back on actually, the silicon is curing as we speak. Stay tuned for some pics.
On an unrelated note, I discovered some weakness in my insulation, and as I unmounted the evaporator, I was greeted to a good amount of water all down the front of the mainboard and my graphics card. Looks like this project is going to be more work than I anticipated. Thankfully I applied a ton of goop in the socket and on the chip PCB, so although it was wet, it didn't cause any damage. Going to have to let this thing dry overnight with a fan on it. Ahh the joys of tinkering with things that are working properly
I've never messed with the ceramique myself. How is it compared to AS5?
Cool stuff, gotta love free samples
The ceramique is good in the sense that it is non-conductive, and you don't have to be as careful, but the stuff very thick, and extremly sticky. I ripped my sempron right out of the socket with the locking lever closed trying to remove the HSF.
3.2GHz 1.55-1.58V
Prior to IHS removal: +3'C
IHS removed: -9'C
IHS reattached: -4'C
Looks successful to me. Oddly, my idle temp has also increased a fair bit. I'm at about -22'C as opposed to -27 before. That should be very helpful in combating the cold bug. Some results to come
Here is a shot of the chip cleaned up, and my first test fitting. Hard to tell by the photo, but the contact was actually very good. Goop was a little uneven there, but I got it sorted out. I decided to keep the ICS and board inside insulated just in case. That clearly saved me from my last encounter with condensation
Here is the final product. Basically, I used some pressure to form good contact. I then applied some 'outdoor' all-purpose silicone to the four courners of the IHS. It takes about 12 hours to cure, so I put about 1/2 pound of weight balanced center on the IHS. Once it dried, it was even all around and held down well. I decided against a 'strong seal' simply because this chip will eventually be demoted to aircooling, and at that time I'll be removing the IHS again anyway. Temps turned out to be quite good to my surprise. So hopefully it will be enough to beat the bug.
I hope that the higher idle temps help to alleviate your cold bug problems.
Its a lot like ceramique actually, but it's the Asetek provided dielectric/TIM that comes with the Vapochill LS. They gave me a fair sized tube, and the stuff works great. Its not nearly as sticky as ceramique, and MUCH easier to clean up than traditional dielectric grease. Since its designed to work well at sub zero temps, I figured it was the best choice
Oh, and COLD BUG ELIMINATED!!
So many things went wrong during this little adventure, but it was worth it. My idle temps sit at around -22'C to -23'C (as opposed to -27'C prior). I was able to successfully surf around the web at 11x303MHz (With my TCCD running at 303MHz 2.5-4-3-6 1T). System didn't hitch once. Prior to this, I'd get nasty hardlocking within 10-15 seconds.
Most "standard out-door silicons" are of the corrosive variety. Non-corrosive silicons are harder to find and will generally say rather explicitly that they are non-corrosive.
Chris, did you get a Vapo or something? He coated it to waterproof primarily.
BTW, my naked Winnie is still alive and kicking. I coated the spot where I skinned it a little from removing the IHS with some of my daughter's nail polish and shoved it into my Dual Sata 2 board yesterday, put an SLK948-U on top and she booted right up.
Thanks for pointing that out, Khaos. I didn't check, as I didn't realize that varieties of silicon are acidic enough to cause issues on PCB. IIRC it was 'low odor, indoor/outdoor all purpose silicon adhesive', but I'm going to double check the tube when I get home. I guess its too late to worry too much about it now
Yup
Come to think of it, it did smell a bit like vinegar. It wasn't terribly strong though. I think I'm going to rot away my proc with condensation long before that silicone causes a problem
Not being one to take chances, I use non-corrosive silicon in pretty much all computer hardware / cooling applications. I definitely recommend it if you ever, say, decide to cover the bottom of your motherboard around the socket area with a piece of neoprene that is attached to the motherboard with silicon. *cough* Not that I would ever do such a thing.