New Swiftech TEC Water block released: MCW6500-T

[afireinside]

34C TAT loaded at 1.53v and only a 226W TEC? Much better than I expected! I wonder how well this 320W does. Need to grab a PSU...

Orange> Is your radiator/tubing hot? 226W TEC + conroe is a large heatdump and TECs are very dependant on how well you cool the hot side. The HE rads really need 38mm thick 100+ CFM fans to perform at their best and you don't even have 3 fans on yours! Plus it's highly restrictive so pushing instead of pulling is hurting you as well. If you have some bigger fans you could try you might get a nice gain.

[Orangeblast]

Orange> Is your radiator/tubing hot? 226W TEC + conroe is a large heatdump and TECs are very dependant on how well you cool the hot side. The HE rads really need 38mm thick 100+ CFM fans to perform at their best and you don't even have 3 fans on yours! Plus it's highly restrictive so pushing instead of pulling is hurting you as well. If you have some bigger fans you could try you might get a nice gain.[/QUOTE]


Agreed. I'm eyeing the new PA (series is it??) 120.3 from thermochill...much more quiet fans. I do need to add more fans as this is something that is hurting me right now.

After rethinking the whole TEC concept out in my head. I went ahead and pushed my Vcore to 1.65 (1.61 Actual after VDrop) and threw my cpu to 3.8ghz.

I'm now loaded in to Vista 32 @ 3.8ghz - TAT reports: 13/15 on the cores as I type this.

I have yet to check for stability of course.

Oh and another thing. It almost seems like TAT has a problem monitoring correct temps. Its jumping all over the place from 9 to 15 in and sometimes 19 to 12....?? is this usual with cold temps?

BUT DAMN I'M FINALLY ON TO SOMETHING!

I'll keep you all posted for sure.

Thanks Gabe for causing me to rethink the concept of the TEc. After seeing your TAT temps at load....it makes more sense is all....


UPDATE: well thus far I am Super Pi-32 stable @ 3.775 VCORE = 1.62 TAT reported 28c degrees during 32 testing. Then again its pretty cool in my room right now. Considering my radiator and the water temp, I'd say 28c at load isn't all that bad. Should be improved with a better Rad and some better fans.

Orthos will be a different story. I'll let you know how that beast of a program does with this OC.

sweet so far though.


UPDATE: well...things aren't looking so great after all. My system refuses to POST properly anywhere from 3.5 - 3.750 It acts like a giant memory hole. I don't get it. It wasn't doing that before with my tuniq tower. 3.4 ghz isn't even stable like it use to be and at 3.775ghz Orthos pretty much just laughs at me and it fails within seconds. That or BSOD.

I think my chip is dying. At this point its almost useless to have this TEC block and such. Whats the point? Run 3.1ghz at a 6 degree max load?



I'm actually tempted to just buy another C2D and some how sell off my lapped E6600 to someone....or build another system..?? uhg. garbage.

[gabe]

Thank you so very much for this insightful information.

So its really okay to have your Core temp as high as 30+ at load with a TEC?

You ran the Meanwell at its max voltage? Would that eventually blow the 226 TEC?

What were the other voltages in BIOS for the SPP/MCP/SPP-MCP/FSB?

TEMPS: Yes it's OK. Even higher.

Voltage: no, VMax on the TEC is 15.4v, so its fine. What I do for the meanwell is to put a low flow 80mm in extraction (blowing out) just above the fan of the PSU. That way the PSU never turns on its noisy 60mm, and keeps cool all the time.

BIOS voltages:
CPU FSB 1.5
Memory (depends on your mem) with ours we ran at 2.35
SPP: 1.55
MCP: Auto
HT: 1.55

Note that the above could be tweaked downwards. It's a matter of experimenting.

System UPDATE: Ran the bench at 3.6Ghz (1.53v after droop) 2 hours under TATS, and did several 3DM06 with no signs of instability. temps in the Mid 30's. Note that we also have 2 x 8800 GTX's in this loop.

Important note: I doubt that you can keep your water-temps as low as ours (29C) as this bench is rigged for different applications with much higher power loads, and we didn't have time last night to build a "consummer like" setup. With your thermochill (and 3 x 80~100CFM fans) , you should normally see your water temps in the high 30's. It means that your CPU core temps should be in the mid 40's at this frequency (3.6) and equal CPU voltage (1.53 post-droop). Noting that our CPU is a dog, yours might require less voltage, which would in turn reduce your operating temps. I'd ad to this water-temp subject, that you can get a thermometer with a thermistor at Radio shack for $15 or so, and stick it in your reservoir. That way, you will be able to tell how your loop is cooling the TEC.

Thoughts: if you do not like the noise generated by 3 120's at 80CFM, you will have to settle for either higher temps, or lower overclock, depending what your priorities are, noting finally that the heat load tests that we conduct using TATS are just that: heat load tests. They do not appear to be an absolute indicator of stability, nor to be representative of what the average use will be, meaning that your cores are seldom running at 100% for prolongued periods of time.

[afireinside]

Gabe what are you using to keep the liquid temps at 29C?

[Console]

Any one used one of thse on a AMD setup yet ?>

[Orangeblast]

BIOS voltages:
CPU FSB 1.5
Memory (depends on your mem) with ours we ran at 2.35
SPP: 1.55
MCP: Auto
HT: 1.55

Bah, never mind. System is still unstable. Orthos passes fine but for whatever reasons I get freezing. System just completely freezes even though orthos is passing fine.

I tried some HD diagnostic stuff...all passing, ran memtest....RAM passes..

why the hell would my system just completely freeze up on me like that?

PSU??? I always thought that when a rail is going bad on a PSU the system will usually reboot...?..

And we all know that it can't be heat thats the problem

[ColonelCain]

Have you added some juice to your northbridge and memory?

Just a bit ago (with the set-up in my sig), I was passing superPi, but crashing in-game. Added some voltage to my memory, and it's stable now.

[L33T]

Ok but just because it includes a 226W TEC @ 50x50 [mm] doesn't mean that it won't fit a 62x62 [mm] TEC. I suppose if somebody would flip the water block base and the cold plate around so we could see the other side, and then place a ruler there, we can figure out if there's someway we could stuff a 62x62 [mm] TEC in there -- even if it comes down to rotating the TEC by a small angle.

[ColonelCain]

The question is, would the block be able to efficantly cool a 437w TEC?

[cyberspyder]

The question is, would the block be able to efficantly cool a 437w TEC?

If the Wintsch Labs block needs 2 inlets & outlets, then I don't know how the Swiftech block can handle the heat. BTW, I love your sig, especially the Klipsch part. I have a pair of GMXs, and they are amazing, along with an X-Fi, and Monster Cable upgrades.

Brendan

[m0da]

well this is what is says on the site (regarding size):

Assembly:

Water-block dimensions: including cold plate, excluding fittings: 2.5"x2.5"x1.070". This fits inside of Intel® socket LGA775 no components zone.
Overall dimensions including retention frame: L3.1" x W3.1" x H1.070" (L78.7 x W78.7 x H27.1 mm)

by they way, 2.5" = 63.5mm
Does this mean a 437w would be well off in this block?

edit: but wait, the inlets on the baseplate don't come anywhere near the outer edge of the plate. I guess 62x62mm 437w is out of the question.

[L33T]

If the Wintsch Labs block needs 2 inlets & outlets, then I don't know how the Swiftech block can handle the heat. BTW, I love your sig, especially the Klipsch part. I have a pair of GMXs, and they are amazing, along with an X-Fi, and Monster Cable upgrades.

Brendan

It is really quite capable, as long as its cold enough. I'm getting ready to use the MCW5002 to do it. Of course, heat removal will be fantastic... you guys will see later.

[THE JEW (RaVeN)]

It is really quite capable, as long as its cold enough. I'm getting ready to use the MCW5002 to do it. Of course, heat removal will be fantastic... you guys will see later.

Very interested in your 437watt TEC on a 5002 results. I've got 2 5002's and 2 6002's and I've debated getting a 437watt TEC myself and undervolting it. Your results would greatly help in my decision.

[Pete]

I wana see this as i'm toying with idea of tec my self now

[L33T]

Very interested in your 437watt TEC on a 5002 results. I've got 2 5002's and 2 6002's and I've debated getting a 437watt TEC myself and undervolting it. Your results would greatly help in my decision.

...

Yes as soon as I finish the writeup and everything, I'll post a link or maybe a copy of it from my site.

[ColonelCain]

If the Wintsch Labs block needs 2 inlets & outlets, then I don't know how the Swiftech block can handle the heat. BTW, I love your sig, especially the Klipsch part. I have a pair of GMXs, and they are amazing, along with an X-Fi, and Monster Cable upgrades.

Brendan

Thanks . I went from on-board sound+ crappy $15 speakers, to this beast of a set-up. Constantly my friends are coming over to just see what a certain song sounds like.
When you say monster cable upgrades, do u mean that you changed the wiring?

[Philly_Boy]

Thanks . I went from on-board sound+ crappy $15 speakers, to this beast of a set-up. Constantly my friends are coming over to just see what a certain song sounds like.
When you say monster cable upgrades, do u mean that you changed the wiring?

I hear you on the Klipsch speakers...I have a set of old skool Monsoon ribbon speakers....and an 8" 1200 watt powered Velodyne subwoofer (the whole cube only measures 10" on a side....it's tiny). Crystal clear mids and highs and bowel movement bass....

Anyway....the base of the 6500-T measures approx 58mm square so that'd be a no go with a 62mm x 62mm pelt. You can't leave any part of the pelt uncooled or you'll run into problems.

[m0da]

Anyway....the base of the 6500-T measures approx 58mm square so that'd be a no go with a 62mm x 62mm pelt. You can't leave any part of the pelt uncooled or you'll run into problems.

Check my post a few posts back. On the site it says it's 2.5"x2.5" and 2.5" = ~63mm.
Is the site wrong?

[gabe]

Any one used one of thse on a AMD setup yet ?>

OK, since soeone asked, here are some informal tests.

Key data:
water temps at 27.5c
CPU is uncapped
everything else is pretty much on the screen shots.

Did two stability tests one with burnk6 x 2 for over an hour, and one with orthos x 2 for 10 mn just to check delta in temps between these two tests.

also managed an 11,568 in 3D Mark 06 (published) at slightly higher CPU frequency which places the system 7th in its category.

[Orangeblast]

what could it be?


I have found that if I go to 3.4ghz on my E6600 it is stable...8 hours +


But I get random freezing/hard locking while doing just regular things...I.E, browsing the forums, emailing etc....

what could it be?

If I go down to 3.375ghz, I get no freezing or hard lock. So far at least...

voltages on Evga 680i

VCORE: 1.48 (before vdrop)
FSB: 1.5
VDIMM: 2.2 (dominator 8500)
SPP: 1.45
MCP: 1.50
MCP - SPP: 1.45


Linked and Sync'd

No Failure in Orthos, yet still hardlocks at random.


Ideas??

Is it possible it could be drivers. I did a chkdsk on the c drive and it fixed a hell of alot of files and indexes. It feels like a bad harddrive but every diagnostic test I throw at it says that my drives are passing.

I'm in Vista 32 and I'm not using the Nvidia Storage drivers at all....perhaps I should?

[m0da]

Test your memory. Memory is a pain in the rear.

also, if you were using an intel chipset, i'd tell you to add some voltage to your nb. don't think you have one exactly... add some voltage to its equivalent i guess? worth a try

[theelectic]

OK, since soeone asked, here are some informal tests.

Key data:
water temps at 27.5c
CPU is uncapped
everything else is pretty much on the screen shots.

Did two stability tests one with burnk6 x 2 for over an hour, and one with orthos x 2 for 10 mn just to check delta in temps between these two tests.

also managed an 11,568 in 3D Mark 06 (published) at slightly higher CPU frequency which places the system 7th in its category.

Just what I needed to see - get an AM2 mount version going so I can start freezing my Brisbane

[Console]

OK, since soeone asked, here are some informal tests.

Key data:
water temps at 27.5c
CPU is uncapped
everything else is pretty much on the screen shots.

Did two stability tests one with burnk6 x 2 for over an hour, and one with orthos x 2 for 10 mn just to check delta in temps between these two tests.

also managed an 11,568 in 3D Mark 06 (published) at slightly higher CPU frequency which places the system 7th in its category.

Thx those Idle temps looks very nice . but that seams like a BIG delta -17 to 30 . Very tempting

[mhypolite]

what could it be?


I have found that if I go to 3.4ghz on my E6600 it is stable...8 hours +


But I get random freezing/hard locking while doing just regular things...I.E, browsing the forums, emailing etc....

what could it be?

If I go down to 3.375ghz, I get no freezing or hard lock. So far at least...

voltages on Evga 680i

VCORE: 1.48 (before vdrop)
FSB: 1.5
VDIMM: 2.2 (dominator 8500)
SPP: 1.45
MCP: 1.50
MCP - SPP: 1.45


Linked and Sync'd

No Failure in Orthos, yet still hardlocks at random.


Ideas??

Is it possible it could be drivers. I did a chkdsk on the c drive and it fixed a hell of alot of files and indexes. It feels like a bad harddrive but every diagnostic test I throw at it says that my drives are passing.

I'm in Vista 32 and I'm not using the Nvidia Storage drivers at all....perhaps I should?

Hey Orangeblast, I got a stable over clock at 3.6 with the same mobo and processor as you, my ram is OCZ2t8002GK at 2.2v EPP enabled for now with the the requency unlinked and set at 800MHz, my FSB is 1600Mhz, and my voltages are as follows (read from NVmonitor)

CPU 1.6v (Under volt to 1.56v)
FSB 1.4v
SPP-MCP 1.45v
PCI-E 1.4v
Nforce 1.5v

Have tested running 2 instances of Prime95 (one per core) for 24 hours, all stable. Iw as monitoring this thread to see what the temps are on the TEC, currently I got one of them Thermal take LCS Kandalf cases, got a nice big radiator (3 120mm fans) and a descent pump, currently with the black that came with the case, my temps under full load using OCCT is about 68 degrees Celsius, using Prime95 for 24 hour load testing it reaches as high as 75 degrees Celsius.

Dont know if you want to try my voltages and see if you get any better luck, but I think the TEC temperatures you are getting will work wonders for me.

[cyberspyder]

FYI, Sidewinder has it in stock: http://sidewindercomputers.com/swmc7.html for $155.95.