Waterblocks for nvidia 6800
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Waterblocks for nvidia 6800
Starting this thread to try to get a handle on the waterblocks that will work with nVidia 6800 cards. So far we have:
DangerDen $125!! Ouch.
SilverProp FusionHL ~$65 (in the US).
PolarFlo TT VGA $50
Does anyone know how good the SilverProp of the PolarFlo are? More interested in the PolarFlo, as it's much easier to get in the US.
Also, I've read that there may be issues with only watercooling the GPU on these cards. There's some sort of power circuit that the normal air cooling blows over. Anyone have direct experience with this?
Steve
DangerDen $125!! Ouch.
SilverProp FusionHL ~$65 (in the US).
PolarFlo TT VGA $50
Does anyone know how good the SilverProp of the PolarFlo are? More interested in the PolarFlo, as it's much easier to get in the US.
Also, I've read that there may be issues with only watercooling the GPU on these cards. There's some sort of power circuit that the normal air cooling blows over. Anyone have direct experience with this?
Steve
-
- SPCR Reviewer
- Posts: 2696
- Joined: Thu Dec 11, 2003 9:53 pm
- Location: Scarsdale, NY
- Contact:
If you browser the HighSpeedPC site, there's an Innovatek block the covers both ,the GPU and the GDDR3. Of course it's the most expensive of all, at about $130 for the standard model, and like $150 for the LED-lit model (purely bling, zero added functio).
There's a secret I discovered about the narrow sink that covers the MOSFETs on the GT (and likely the Ultra) boards when I dismantled the cooling system to redo the TIM, as the factory TIM was junk. Underneath the sink (contact surface side), at the two ends where the screwholes are, there are elevated portions with some nonconductive barrier tape added; this, ledge, I'm quite sure, was designed to prevent crushing the MOSFETs with excessive screwdown pressure. Here's the problem: the ledges are WAY too high!!! When I cleaned off the factory TIM and applied Arctic Silver V to both surfaces, and then placed the sink down on top and lifted it to check contact, there was ZERO contact!!! NONE at ALL!!! For reference point, mine is an XFX brand 6800 GT card and does not deviate from the reference design one bit.
What I did to remedy this problem, was I peeled off the two pieces of plastic to reveal the ledges, and then I utilized my rotary tool to grind down the ledges about 1.5-2mm. I then cleaned all surfaces, reattached the plastic anticonductive pads, reapplied the TIM and then did another contact check...
Perfect! All three MOSFETs (for Ultras, all six) now make good contact. To finish it off, I simply screwed the sink down being careful not to overtighten; when I did a visual check to see the gapping on the ledges to the board after I grounded it down (before rescrewing), I found that I had removed enough of the ledges to allow the MOSFETs full contact, but I had not removed an excessive amount of the ledges; there was almost no gapping from the board. This way, even if I did overtighting, it is still not very likely to do major harm.
I recommend, to 6800 GT/Ultra owners, to check the contact on their PWM heatsink, because we're not talking poor contact here, we're talking ZERO contact!
And now on to lighter matters, here are some photos I have shot of the NV-68 block from Danger Den; it arrived to me this afternoon!
Per usual, clicking on a photo will bring up a higher resolution version.
As shipped, from the bottom side (top side when viewed in standard tower configuration once installed into the system):
The contact surface:
The top of the cooler, particularly when installed in Lian-Li V-series, G5-like cases:
Sorry for not doing my usual soft edges; in a bit of a hurry today.
-Ed
There's a secret I discovered about the narrow sink that covers the MOSFETs on the GT (and likely the Ultra) boards when I dismantled the cooling system to redo the TIM, as the factory TIM was junk. Underneath the sink (contact surface side), at the two ends where the screwholes are, there are elevated portions with some nonconductive barrier tape added; this, ledge, I'm quite sure, was designed to prevent crushing the MOSFETs with excessive screwdown pressure. Here's the problem: the ledges are WAY too high!!! When I cleaned off the factory TIM and applied Arctic Silver V to both surfaces, and then placed the sink down on top and lifted it to check contact, there was ZERO contact!!! NONE at ALL!!! For reference point, mine is an XFX brand 6800 GT card and does not deviate from the reference design one bit.
What I did to remedy this problem, was I peeled off the two pieces of plastic to reveal the ledges, and then I utilized my rotary tool to grind down the ledges about 1.5-2mm. I then cleaned all surfaces, reattached the plastic anticonductive pads, reapplied the TIM and then did another contact check...
Perfect! All three MOSFETs (for Ultras, all six) now make good contact. To finish it off, I simply screwed the sink down being careful not to overtighten; when I did a visual check to see the gapping on the ledges to the board after I grounded it down (before rescrewing), I found that I had removed enough of the ledges to allow the MOSFETs full contact, but I had not removed an excessive amount of the ledges; there was almost no gapping from the board. This way, even if I did overtighting, it is still not very likely to do major harm.
I recommend, to 6800 GT/Ultra owners, to check the contact on their PWM heatsink, because we're not talking poor contact here, we're talking ZERO contact!
And now on to lighter matters, here are some photos I have shot of the NV-68 block from Danger Den; it arrived to me this afternoon!
Per usual, clicking on a photo will bring up a higher resolution version.
As shipped, from the bottom side (top side when viewed in standard tower configuration once installed into the system):
The contact surface:
The top of the cooler, particularly when installed in Lian-Li V-series, G5-like cases:
Sorry for not doing my usual soft edges; in a bit of a hurry today.
-Ed
-
- Posts: 135
- Joined: Tue Jun 01, 2004 10:09 am
- Location: Fort Worth, Texas
Just an update. Got an email back from PolarFlo. So now the only question is how does the TT VGA compare to my existing Maze4GPU:
I believe that our TT VGA will work with it. We have had a few customers order them for that card and we have not heard of any incompatibilities. Thanks
----- Original Message -----
To: '[email protected]'
Sent: Tuesday, August 10, 2004 6:57 PM
Subject: nVidia 6800 block...
Do you guys make a water block that will work with nVidia 6800 boards?
-
- Posts: 34
- Joined: Fri Jan 03, 2003 1:44 am
- Location: San Francisco, CA
So far the concensus (procooling, etc) is that PolarFlo is all bling and no zing.Sklug wrote:Just an update. Got an email back from PolarFlo. So now the only question is how does the TT VGA compare to my existing Maze4GPU:
Its pretty hard to tell since they won't send the block to be reviewed by any of the reputable sites. But think of it this way, according to their own data, the TT CPU block isn't any better than a maze 4 cpu block. I would hold off supporting/buying anything from PolarFlo until they are tested by a reputable website such as Overclockers or ProCooling.
Aaron Spink
speaking for myself inc.
Agreed...The silverprop seems like a better bet, though still relatively expensive, even though it's only HALF the cost of the DangerDen.
Going to wait a bit on the new vid card anyway, as the 6800 GTs are still selling for mostly above list, and ATI hasn't really released a resopnse to the 6800 GT yet anyway (i.e., a less expensive 16 pipeline part).
I may have to stop waiting when Half Life 2 comes out though. Hopefully there will be some better waterblock options by then.
Steve
Going to wait a bit on the new vid card anyway, as the 6800 GTs are still selling for mostly above list, and ATI hasn't really released a resopnse to the 6800 GT yet anyway (i.e., a less expensive 16 pipeline part).
I may have to stop waiting when Half Life 2 comes out though. Hopefully there will be some better waterblock options by then.
Steve
-
- Posts: 135
- Joined: Tue Jun 01, 2004 10:09 am
- Location: Fort Worth, Texas