Xigmatek HDT-S1283 & SD964 "heatpipe direct-touch&q

[Mats]

While most CPUs with a lid have some sort of TIM, some of them are soldered (or glued) onto the CPU, including the X4 and most 775 CPUs.
Solder is of course a kind of TIM, all I'm trying to say is that it's not that easy to remove soldered lids.
Just in case someone didn't know this and wanted to remove it.

[cmthomson]

While most CPUs with a lid have some sort of TIM, some of them are soldered (or glued) onto the CPU, including the X4 and most 775 CPUs.
Solder is of course a kind of TIM, all I'm trying to say is that it's not that easy to remove soldered lids.
Just in case someone didn't know this and wanted to remove it.

The metal part of the IHS may be soldered to the rest of the case, but the TIM I'm referring to is between the CPU die and the IHS, and is definitely not solder. Here's a diagram from the Intel data sheet:

[mattthemuppet]

wether it's TIM or solder it has a similar melting point to what ever's used to stick the core to the substrate - I remember some OCing site trying to remove the heatspreader from a P4 extreme and he had to heat it up so high to remove it that he pulled the core off with it. I think he was inspired by the fact that you can do the same to AMD chips with IHS' just by a bit of cutting and twisting.

Either way, TIM or solder, it's purpose is the same - to facillate more efficient heat transfer between core and IHS.

[cmthomson]

Look at this: http://www.cooling-masters.com/news-741 ... oses-.html

Wow, that's quite an imprint.

I had a look at our 1263 sample. It has some slight marks but nothing like the above. Perhaps the example above is of sample that was handled roughly during installation? Or maybe just left on the CPU/mobo for a long time. That HS does seem to apply a fair bit of pressure. Mind you, just about every base develops some kind of blemish at the perimeter of the CPU heatspreader.

None of those images imply "crushing" to me, merely indentation.

For a heat pipe to be "crushed" it would have to flattened so much that the internal wick/fluid structure could no longer function. This would amount to the pipe being pretty much flattened, which is clearly not the case in the photos above.

My interpretation of these slight dents in the pipes is quite opposite of the posters above: this is a good thing, that improves the heat transfer from the IHS to the outside of the heat pipe.

[cmthomson]

whether it's TIM or solder it has a similar melting point to whatever's used to stick the core to the substrate - I remember some OCing site trying to remove the heatspreader from a P4 extreme and he had to heat it up so high to remove it that he pulled the core off with it.

Well that makes perfect sense. If the IHS is soldered to the carrier, and you're trying to remove the IHS, you'd have to heat up the edges of the IHS to the melting point of the solder. So think about what's happening to the TIM and the core: they are being heated up too, and because the TIM is efficient, the core is being heated to solder-melting temperatures, and of course separates from the flip-chip substrate.

In general, I've found the methods of these extreme overclock sites to be pretty suspect. In this particular case, the guy was trying to take apart a CPU package with the equivalent of a blowtorch instead of a more appropriate cutting tool. Had he cut it apart, he may well have discovered a greasy TIM between the core and the IHS. (Or not: unless someone conducts a suitable destructive experiment, or an Intel engineer pipes up, we aren't about to find out).

[Mats]

The metal part of the IHS may be soldered to the rest of the case, but the TIM I'm referring to is between the CPU die and the IHS, and is definitely not solder.

Soldering the lid to the die surface is very common these days.
The first time I heard about it (I think) was in 2005, probably a PD.

Examples:
9600 X4 with lid removed. The CPU die got ripped off of the package PCB. Thread.
More pics. Dunno if the latter is a PD or a C2Q.
Article about removing a soldered lid, with horrible results.

[cmthomson]

The metal part of the IHS may be soldered to the rest of the case, but the TIM I'm referring to is between the CPU die and the IHS, and is definitely not solder.

Soldering the lid to the die surface is very common these days.
The first time I heard about it (I think) was in 2005, probably a PD.

Examples:
9600 X4 with lid removed. The CPU die got ripped off of the package PCB. Thread.
More pics. Dunno if the latter is a PD or a C2Q.
Article about removing a soldered lid, with horrible results.

Well that last link is the most informative and revealing. It documents quite convincingly that the "solder" softens at about 85C. Let me assure you, there is no "solder" that melts below the boiling point of water. In fact modern ROHS (non-lead) solders melt at 265C+, so whatever this TIM is, it is not "solder".

[Mats]

In fact modern ROHS (non-lead) solders melt at 265C+, so whatever this TIM is, it is not "solder".

LOL, well, who said it was standard solder??
Of course they can''t use anything with a 265° C melting point, quite obvious. You don't have to tell me that.

For instance, SnIn52-based solder melts at 118° C (Sn = Tin, In = Indium). I'm not saying that they're using that, though.

It doesn't matter to us customers if they're soldered or glued, all we need to know is that the lids are really hard to remove.

[colm]

Why couldn't they have included a S478 adapter as well? This would be great in my T7200 desktop system. <sigh>

-D

I agree, given the results....
I recently added a crazy agp card and the "old" p4 2.8e is hitting 65C on tight reverse duct and an xp-90 firmly in place on arctic silver.. and I am sure there is other stuff i paid too much attention to 2 years ago upon mounting it...but...Why doubt it now i guess. 12000 hours, still cranking out the top number for the 2.8 at distributed.net and error free.

there is that top 5 tdw number in the history of the cpu, and the 2.8e has one of those numbers...bah. it is good to see other coolers and see what is evolving. I am no doubt 775 socket next system anyway..
Thank my lucky stars for heatpipe cooling, and means to verify reviews at spcr...the reason I got an xp90 2 years ago with not a single regret.

the tests are good here in other words...the 21c ambient is not my 85F and the cpu is smarter than the past (I hope) I bet the coolers are as good as it appears here. great article.

[rpc180]

I've been running one of these for about a week with a bolt through kit, nexus PWM fan and AS3 (yes, 3). Its been pretty good. My E8400 will hit 44C when gaming, folding, and monitoring temps. But when its just folding it never breaks 40C. The nexus PWM fan runs between 1100-1350rpm bios controlled. Though I'm thinking of dropping a spare S-flex 800 on it instead, I ran the nexus at 350rpm overnight while folding to see how it would do - temp change was minimal (2-3 degrees).

[b0bber]

I ordered one online, and installed it to replace my scythe ninja, and the least to say is that I'm not impressed...

Although the fan performed exceptionnaly well in all the benchmarks I read, I really was deceived by its performance in my system.

I'm running a q6600 (P5K-E) at 3.2ghz (8x mult, 400*4fsb) at 1.22v under load and the temperatures aren't impressive at all.
The fan is inside a p182 case, with 5 noctuas NF-S12-1200, three at full speed (one inside the upper front intake attached with the metal clips), one on the heatsink, and one on the back of the case, the upper fan is running with the ULNA, and the PSU fan too.
There's no fan on my 8800GT wich is using the Accelero S1 (which really is an outstanding heatsink).
I'm running 40 celcius idle and 59 to 60 celcius under load, with an room temperature of about 21 degrees celcius, compared to 38 idle and 54 under load with the scythe ninja.

The heatsink is installed using thermalrigt's LGA775 bolt-thru kit (I had to cut out the push-pins of the original mounting system with a pair of pliers), and some arctic silver MX-2 thermal paste, which I've tried to put in many different ways and quantities (single vertical srtipe on both dies, spread, using a razor etc..) with almost the same results.

Oh and I've also noticed some very slight bevels in the heatpipes as previously demonstrated. This could be a good thing for HS contact, but it limits my ability to resell the HS afterwards...

Anyhow, I just wanted to share my experience with other buyers because I'm astonished by the difference between the results in the reviews and what I actually witnessed, and if anyone has a piece of advice, I'd gladly take it.

[spungfoo]

I put one of these suckers in my system last week and I have to say I'm quite satisfied with its performance. 35C at idle, hovers between 50-55C under load on a e6750 with AS5 and a Nexus D12 at 1000 rpm. A helluva lot better than the Intel stock cooler I had on it before. Lightweight and cheap compared to the other tower heatsinks out there.

[Fallsroad]

The SPCR review convinced me to buy one for an upgrade I was doing. Installation was pretty easy except for the fan - a bit of a pain getting those rubber mounts in without bending any fins on the sink.

The pressure I had to put on the pins was about the same (to my perception) as the pressure I had to apply when doing two builds using the E6750 and its stock Intel sink.

Results are very good.

Gigabyte GA-EP25-DS4
E8400 @ 3.6GHz
eVGA 8600GT
P180B case

I have two Noctua exhaust fans running at 600rpm. Idle temps for the cpu are 31C, full load tops out around 45C max, which is better than acceptable. The stock fan on the Xigmatech spins between 850 and 1100rpm, and while not inaudible, is not bad. I'll be replacing it eventually to use something that spins a little slower.

All in all, very impressive. It is rather lightweight, the pins, as hated as they may be, are enough to hold it to the board securely (and the Thermalright back plate will work with it), the fan is reasonably quiet (though I think most here would choose to replace it), and the cooling is excellent.

Thanks to SPCR for a useful, informative review.

Now I can't wait for my ECS passive 9600GT to arrive.

[Ash]

@ Fallsroad

im also looking to get one of these and your temps are impressive. im not a gamer so i will not be putting the cpu under alot of load but i want a quieter cooler.

how have you applied your thermal paste on the cpu?
i know everyone has their preferred way but with your temps i wanted to know how you applied it.

[Fallsroad]

@ Fallsroad

im also looking to get one of these and your temps are impressive. im not a gamer so i will not be putting the cpu under alot of load but i want a quieter cooler.

how have you applied your thermal paste on the cpu?
i know everyone has their preferred way but with your temps i wanted to know how you applied it.

@ Ash

A note about temps: There is no consensus about which programs give accurate (or even semi-accurate) readings for the E8400. I've used many, including Core Temp and Real Temp, which show an exact 10 degree disparity between them. In the end I chose Real Temp because it corresponds most closely to the temps reported by the BIOS. However, that is exactly 10C less than the readings returned by Core Temp, so in my pursuit of an over clock I've kept the +10C figure in my head, to make sure I don't stray into hardware frying territory.

There was a good little piece on a review site about different methods of applying thermal material to the Xigmatech. Unfortunately, I cannot locate it (the site also did a review of the heat sink, but Google is not my friend tonight), but the writer tried several different methods using AS5, then pulled the sink and took photos of the result. One of the things he pointed out is that following the AS5 directions may not work because of the slim notch between the sides of heat pipes and the aluminum "holder".

The best method according to his trial and error was to pre-apply a tiny amount of AS5 (or other TIM) in those areas where the pipes and the metal meet, essentially filling them in, then wiping any excess clean with something flat and stiff, like a plastic card. Then apply a very thin, short line at the center of each heat pipe and install it carefully on the CPU, trying not to slide it around too much.

The result is very close to the photos AS5 has in their directions - a thin, even layer of TIM on the sink and CPU heat spreader.

I likely used a little too much, I'm betting, and possibly could do better and perhaps squeeze a degree or two, but I'm holding off changing it until my fan controller arrives and I pull my entire rig apart and re-install it to the case to better accommodate slower fans, a new passive sinked graphics card, and the less than tidy jumble of cables.

When I do that I'll pull the sink and see how the TIM actually spread.

Regardless of the actual readings I'm getting (be it really 31C or 41C idle), the impressive thing is the decently low temp rise from idle to full load. If the BIOS and Real Temp are both wrong and Core Temp right, then I'm still looking at 55C maximum at full load on a 20% over clock, with a pretty quiet heat sink that is light, and can be made quieter.

Since you are not a gamer, this sink could work very well for you, depending upon your CPU (though I think it can handle most anything competently), case, choice of fan if you go with something other than stock, and airflow.

Do be sure to measure ahead of time - I've read of some folks having problems in some cases, usually those that have a side fan across from the CPU.

Good hunting.

[b0bber]

@fallsroad

I believe you are talking about that article:

http://benchmarkreviews.com/index.php?o ... 0&Itemid=1

Especially that page:

http://benchmarkreviews.com/index.php?o ... mitstart=4

You might want to try the "Intel Thermal Analysis Tool" in order to get your temperature readings too... It might not be better but who knows...

I've tried Real Temp and it gives me 5-6 degrees less per core (compared to speedfan, coretemp, TaT) as the Tjunction is detected at 95 instead of 100 degrees on my q6600, so I'm not really sure it's working perfectly...

Anyhow, I've also realized that my temperature drops 5 degrees when running open case (see previous post), so it's most likely an airflow problem, I'll try to put the HS horizontally and see what the temperatures are like...

[Ash]

thanks for the response mate.

il bare that in mind when i come to installing it.

my cpu is a c2d e6420 overclocked from 2.1 to 3.2, at the moment im using a zalman cooler on it, but temps are around 55 when under full load testing.
my new flat is really warm, and the temp around the computer are getting higher.

im looking to getting a p182 case, and some slipstream fans, probably put a 1200 on the xigmatek aswell.

my gpu is a 8600gt, stock fan is not loud, but im sure il be able to hear it once i replace the noisy psu and cpu cooler.

iv also overclocked my ram, do they need cooling when oc'd or is the airflow in the case enough to cool them.

i just read the links in the articles posted above. the guy applied the thin line on the base rather than the heatpipes themselves. is that what you meant aswell or did you actually apply them to the heatpipes. overall i think it wouldnt make a difference in terms of how much of the cpu they cover.

[Fallsroad]

thanks for the response mate.

il bare that in mind when i come to installing it.

my cpu is a c2d e6420 overclocked from 2.1 to 3.2, at the moment im using a zalman cooler on it, but temps are around 55 when under full load testing.
my new flat is really warm, and the temp around the computer are getting higher.

im looking to getting a p182 case, and some slipstream fans, probably put a 1200 on the xigmatek aswell.

my gpu is a 8600gt, stock fan is not loud, but im sure il be able to hear it once i replace the noisy psu and cpu cooler.

iv also overclocked my ram, do they need cooling when oc'd or is the airflow in the case enough to cool them.

i just read the links in the articles posted above. the guy applied the thin line on the base rather than the heatpipes themselves. is that what you meant aswell or did you actually apply them to the heatpipes. overall i think it wouldnt make a difference in terms of how much of the cpu they cover.

@ b0bber

Those are the right links. I'll give TAT a try, see how it plays out. Thanks!

@ Ash

I filled the gaps as in the photos, but then used three lines, thinner than the reviewer uses, at the center of each of the heat pipes. The overall amount of TIM is likely pretty close, as my lines were thinner and a bit shorter, to account for three. When I redo the rig I'll get a look at it, see how it turned out. As you can see from the pics the reviewer's method gives a pretty even spread.

RAM generally doesn't require specific cooling beyond case airflow unless you have really pushed the voltage and the placement of the RAM puts it out of the flow of the case. The latter is probably unlikely. If you have stability issues you can try cooling the modules directly, though I doubt it would help much. Instability would probably require you to back off the over clock.

Yesterday I swapped out my eVGA 8600GT (which is pretty quiet) in favor of an ECS 9600GT with an Acelero passive cooler (comes preinstalled with it). One less fan makes a bit of a difference, and my GPU temps are way down - from 55C idle / 79C max to 39C idle and 50C max. Pretty astounding stuff, and I didn't add a fan to the passive cooler - and won't need to, it seems.

My only problem now is one of my new Noctua fans is making an unbearable warbling noise, even at lower RPM.

That, and the six hard drives make a bit of a "hum".

[Ash]

i see what you mean, it works out the same in the end.

do let us know when you remove the hs again.

il let you guys know how i get on, when i get around to it aswell.

thanks for you help.

[figment]

I tried a couple different methods using one of the plates I pulled from the 5.25" bays on my P182.

It wasn't really that I didn't believe that website, but rather that I'd never tried before and wanted to see for myself. Sure enough, I have to agree with their conclusions. Two longish, thin lines on the aluminum mount-block work the best. Like others, I figured three lines across the heatpipes would be just as good or better, but for whatever reason, it didn't spread as well. It doesn't make sense at first, but with a little experimentation you can see that the spaces in between the pipes act as channels. If you put the TIM there, it will spread out evenly away from the channels. If you put it on the pipes, it seems to flow to the channels and stop. You get good coverage in the direction perpendicular to the pipes, but poor coverage along the pipes.

I will, however, stress one thing: Don't forget to do a pre-application to fill in the gaps. I did that the first time I installed the heatsink and it seemed to work fairly well. When I removed it, I saw a decent amount of TIM sitting on the chip and so when I remounted it, so I skipped that step, assuming the TIM would squish in by itself. Turns out I was wrong. Or wrong enough. I actually got a 3-4C increase in idle core temps after resetting the heatsink.

So, read through that link above. It really does have good advice for all those HDT users out there.

[Fallsroad]

I filled the gaps first, made sure everything else was squeaky clean, then used the very small, thin lines on the pipes themselves. I suspect with the gaps alongside the pipes already filled, it should have spread pretty neatly.

When my fan controller arrives and I have the time to tear my rig all the way down and reassemble it, I'll check it out and take a pic or two, see what I have wrought.

If it didn't spread cleanly, I'll follow the procedure from the link - which I had read, but was not handy at the time I actually mounted the sink.

[VanWaGuy]

Let us know how it works, that is exactly how I think I would have done it as well. I think that the lines would keep air pockets from forming along the pipes, and with the gaps between the pipes, the single blob method may not work.

[Metalica77]

I'm actually thinking about the model with the Orange LED Fan so i will see how that goes

[Ash]

guys, im desperately trying to get a bolt-thru-kit for my xigmatek 1283. i cant find any instock in the uk.

i have found some metal back plates, but i dont know if they will be strong enough.

http://www.specialtech.co.uk/spshop/cus ... uctid=5863

i am currently using a zalman cooler, and i was wandering if they plastic back plate would be strong enough to hold the xigmatek? what do you think?

i didnt want to use the push pins as they didnt look secure enough too me, in anycase i cant now because iv broken them while removing the pins from the metal.

if i get it shipped from US or anywhere outside uk, the postage cost will be 4 times the price of the kit.

any suggestions?

[NeilBlanchard]

Hello,

Did I miss the links to sound recordings for these Xigmatek heatsinks? I could have sworn that they were there, but look as I might, I can't spot them...

[Ash]

well i just installed my s1283, and it was a nightmare. partly my fault though.

i have mounted the fan on the heatsink first as it was quite tricky to get on, i put a slipstream 1200 on it instead of the original fan.

well as a result, i could mount the hs on the MB, i managed to get a Thermalright bolt thru kit, and the fan was stopping me from mounting 2 screws on one side. the thing is i had already applied the thermal paste (AS5) so when i took it off i had to clean and re apply again.

it took me a while to get the fan off, i didnt want to apply too much force as i didnt want to break the rubber mounts or bend the fins on the HS. i managed to release one side of the fan which was enough to move out of the way and screw it on. but now im worried about the state of the paste, as it moved around a bit while i was mounting it.

as for the temps, they are fluctuating a lot, started on 28 from boot, and since then it has varied between 32 and 42. im hoping it just needs setting in and the temps will then stay the same.

i used the method shown earlier in the thread, by filling in the gaps and then applying 2 strips on the base.

my zalman previously was on 34 on boot and then stayed on 35-39 through most of the day.

anyone else have this problem, the only thing i can think if it carries on is the paste has spread all over as the HS was moving, i guess il only know if i remove it and re apply it.

[Fallsroad]

What RPM is the slipstream turning at?

Those temps seem to indicate that the thermal paste application may be less than ideal. You should see the same or lower temps than the Zalman. Are you using the same temp monitoring program as you were previously?

The sink is a bit of a beast to install, and I did it outside the case, mounting the sink first, then the fan, then putting the board back into the case. Even being careful, my fumble fingers managed to bend a fin or two - they go back in place pretty easily.

I suggest pulling the board from the case and starting over. That way you can leave the fan on and use a flat screwdriver to apply pressure on the pins on the fan side of the sink.

[Ash]

thanks for the reply,

the slipstream 1200, is turning at just over 1200, around 1250 average.
i know the original fan will spin faster, therefore more likely to keep it cooler, but i also wanted to quieten down my system noise.

i am using everest, same as before i installed the xigmatek, iv tried to find out how accurate the temps are but nothing conclusive. i also used TAT (intel program) that shown temps slightly higher than everest.

i did mount the hs with the MB out of the case the first time, i can get around the fan issue again, by just taking one side of the fan off the hs but im just concerned how long the rubber mounts will last.

i had noticed that my vcore keeps fluctuating now, i have set it to 1.33 and it used to just sit on 1.3 all the time before i installed the xigmatek. and the 12v reading also keeps fluctuating but none of the readings used to do that. i have re checked to make sure all power cables are secure but its still happening. vcore will go from 1.28 - 1.3, and the 12v will go from 12.04 - 12.30.

on boot this morning temps were 28 again, and this time they fluctuated between 31-33. i just did a quick 10 min stress test on everest, and the temps reached 52, as soon as i stopped they were idling at 32/33. so im hoping that the paste is setting it self, i will wait untill next week, see if the temps drop before deciding on re applying it.

i have overclocked my system, the original plan was to revert back to default when i first boot with the xigmatek, but i forgot and because it took so long i couldnt be bothered after that.

the vcore was stable before that and i also noticed that while under stress, the vcore ran at 1.25 - 1.26. as soon as i stopped it and it went on idle it rose to 1.3 and fluctuated between 1.28 and 1.3 again. is that normal?

[Fallsroad]

It is normal for there to be small fluctuations in vcore and other voltages. This is in part because those voltages do shift slightly, and also mild inaccuracies in the sensors/programs retrieving the values. vcore will also sag somewhat, motherboard dependent, as a high load is placed on the CPU - this is often called "vdroop".

In an over clocked system those temps are better than average.

Remember that a slower fan may result is slightly higher idle temps - the strength of the Xigmatech lies in keeping a lid on full load temps. Your other fans in your case and the room temp will also play into it. I could get my temps down even further in my rig if I removed 4 of the 5 hard drives - I have good ventilation, but all that heat does affect CPU temps to some degree, even as it is being vented from the case.

CPU-Z is a good monitoring program for CPU speeds and vcore. HWMonitor is a good all round temps, voltages, and fan speed monitor and is written by the same folks.

[Ash]

thanks again for the advice.

i can understand about the fluctuation, the thing that gets me is that it never happened before, i cant help thinking i might have done something as i disconnected the MB from the power supply when removing it from the case.

i have tried other programs aswell, like core temps and hw monitor, the v core doesnt fluctuate on them.

the temps are ok, considering the oc. but i expect them to be better than the zalman, i used to run that fan at 1200 aswell. the xigmatek is also quite sensitive to any activity such as opening programs, internet browsers etc, temps go from idle at 31 to 39/40 and then come back down.

my mb temps are the same as before staying at 33/34 all the time, so nothing else has being effected.

im pretty certain its to do with thermal paste application. did you find the base of your heatsink to be quite rough?