Nice Sinks saw them at Petras
Printable View
Nice Sinks saw them at Petras
Out of curiosity, is there any possibility you'd be willing to sell just the MOSFET part alone?
If so, how much?
What kind of TIM is everyone using when putting these new sinks on? How are you applying it? Using the same TIM on both memory and Vregs?
i use tuniq , tx2 i think its called
Here is a pictures of the installed sinks. I used MX-2. From viewing a couple of pictures of other installs it seems that I might need to back out my screws a bit. Haha. :eek:
That's because it's on a GTS, this time. :p:
Will these fit MSI 8800GT OC edition?
I am the new owner of Iandh's first set of sinks and his custom Waterblock...fantastic work. I feel priviledged :)
So these sinks only work on the 8800GT and 8800GTS G92 Version?
Do you work with the original 8800GTS?
Pic w/ the fan.
Looks like I have to reconsider my new dream setup: Ian's sink + VF900. I've been performing a bunch of heat tests (oZone3D fur, ATITOOL artifact scanning, running various benchmarks, etc.) tonight with two 8800GTS G92s (one card is stock and one card was modified with Ian's sinks + VF900). The results aren't pretty and that's why I'm still up at almost 3am still testing the cards. Haha. I think the problem lies with the VF900. Of all the tests oZone3D seems to generate the most heat. Running it in stability mode @ 1280x1024 windowed with MSAA 0 the modified card (fan @ 100%) will run up, fairly quickly compared to the stock card, to 85-90C+ and eventually cause the program to crash. Performing the same test with the stock card (fan @ 45%) temps will rise, quite slower then the modified card, to 75-79C+ (the temp will eventually settle around 78C). There's no crashing with the stock card. I used MX2 on the modified card and reseated the VF900 twice now. :confused:
I'm guessing with water Ian's sinks rock. I'm interested to hear how others have fared with air.
Your results are very different than mine with air cooling this card.
I had a vf700 on it for about a week and was topping out at 55C load, and now with an accelero S1 I am topping out just over 50C.
Almost sounds like poor contact on the gpu. Those temps are strangely high.
What exactly does an anodized finish look like? Anyone have good pics. of something, anything that's anodized? I kind of like the bead-blasted finish.
Most ram heatspreaders are anodized, many NB heatsinks are as well, Maglite flashlights, all kinds of different stuff.
The "clear anodized" sinks will look very similar to the beadblast finish.
I would like to think it's a contact issue between the VF900 and CPU but testing the modified card with games (high res/max settings), 3DMark06, ATITOOL, etc. to generate heat the temps stay lower then the stock card. This is what I expected and what lulled me into a false sense of security. But once I ran the fur test (with the settings in the post above) the picture quickly changed and what I thought was stable quickly turned into a mess. The amount of heat that the fur test can cause to be generated is impressive.
You mention hitting 55C at load. What do you define as load?
My new definition of load is running the oZone3D fur test with the settings above at the same time as the ATITOOL artifact scanner (for heat generation). Haha.
Thanks.
I just stopped by the machine shop, he has somewhere around 80 sets complete at the moment, so there should be no more stock issues :D
Parts ship monday for anodizing, and should be back within two days, shipment to Petra's should be wednesday or so.
These are gonna be SWEET!!! They look totally awesome.
I'll try to get some video of the parts running if possible.
iandh, you mentioned that you still have a few sets of your original bead blasted finish sets sitting around. Could you save one of those sets for my order + waterblock mount? I like the look of the beadblasting. Plus, I plan on placing a fan card adjacent to the video card to remove ramsink heat. The beadblast finish will be better for generating air turbulence on the ramsink surface which will increase its cooling efficiency.