• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Any NV 5 Silencer install tips?

OPUS1

[H]ard|Gawd
Joined
Apr 25, 2003
Messages
1,766
Just ordered the NV 5 Silencer for my GT.
Any ideas,tips,suggestions,hints, or just general heads up on the install process?
 
its very, very simple. you just put the 4 pegs through the holes and tighten. peice of cake

be SURE to use as5 or equivalent though, the crap that comes with it will up your temps about 8c.
 
-When you unscrew the old HSF, pull it off the board SLOWLY. The old goop isn't exactly glue, but it will resist being pulled off with a kind of suction effect between it, the board, and the HSF. It also can get a little "stiff" like dried-out chewing gum. So don't just yank it off hard or you might damage something.

-99% Isopropyl alcohol removes the old thermal paste easily. Use a lint-free cloth so you don't leave bits of crud all over your board.

-Use AS5, you may need to put it a little thicker on the RAM chips go get 100% good contact with the heatsink, but generally a little dab spread out evenly over the chip will do. It is definitely better than the usual silicone goop. And the tiny tube you get for $9.99 is plenty.

http://www.arcticsilver.com/arctic_silver_instructions.htm

Note that AS5 has a "break in" period before it is giving you the best thermal conductivity, so you may not get your 100% optimum cooling until you have run it throuhg several cycles of cooling off and running under load with a 3D program.

-Don't tighten the screws too much. Tigthen one screw a little, then another, work your way around, and repeat.

-Look at the board to make sure you are getting full contact on all the surfaces. You may need to gently "squish" the board and the HSF together with your fingers, or tighten the screws a little bit more to get this. Bear in mind that the makers of the boards are pretty sloppy about this, so you have to screw up pretty badly to mess this up.

Overall, it is very simple, it took me about 30 minutes, mainly because of time spent cleaning up thermal goo. Just take your time.

My results were as follows:

I ran RTHDRIBL in a window, turned the settings up till it only ran ~ 30 FPS, and let it run to see how hot it would get. The results are listed below, with the stock temps for comparison. The card is clocked at the stock (for BFG) 370 MHz/1.0 GHz.

STOCK HSF

Startup: 56/38C core/ambient
Load: 82/62C
Post-load idle: 60/44C

NV Silencer 5

Startup: 53/36C
Load: 72/54C
Post-load idle: 55/39C

Overclocking test: I increased the clock rate to Ultra levels (400 MHz core/1.1 GHz memory). Here are the results:

Stock cooler: 85/64C core/ambient
NV Silencer 5: 73/55C core/ambient

http://www.daionet.gr.jp/~masa/rthdribl/
 
Use AS3 or AS5 _ONLY_ on the core. Use AS Ceramique on the ram chips.....very thin coating on any thermal compound too.

Dont risk using AS3/5 near the rams, if any gets over the edge and under the chip it will short out pins on the BGA and you will be in a world of hurt. Use Ceramique for the rams, it's non-conductive and a mistake wont ruin your day.
 
NV-5 on a Evga 6800 Ultra:

I found that the supplied rubber washers don't cut it.. so I used the orginal back plat as a washer or spacer...

The NV - 5 has 2mm studs and the stock cooler has 2.6mm bolts that screwed into the back plate..

I took a small drill bit and reemed out the original backplate holes a little and it slid right over the NV-5 2mm Studs perfectly.. I then tightened it down slowly...

I found that if you hold the card up into the air faced directly at a light and look between the heatsink and the the pcb that you see the four corners of the metal shim that surrounds the core...So I tighten the four nuts up until each corner of the shim barely touched the nv5 copper base....

NOTE: I also Lapped the base w/ 1500 then 2000grit.... I started with 1500 then onto 2000 grit.. The base is pretty flat, but kind of rough textured. 1500grit had no problem smoothing out the base, and the 2000 grit allowed for a shine...

I also used AS5 on both the GPU and the Memory..

Results:

Ambient: 35C

Stock Cooler:
- Idle - 57/58C
- Load - 78/80C

NV - 5
- Idle - 48/49C
- Load - 66/68C

When I first installed the NV-5 with the rubber washers I was Idling at 53/54C....

I have edited the original bios to 450mhz /1.16ghz w/1.5v Vgpu "3dmark2k1, 2k3(Looped 3 times), and Doom3 (2 hours) Stable"
 
uwackme said:
Use AS3 or AS5 _ONLY_ on the core. Use AS Ceramique on the ram chips.....very thin coating on any thermal compound too.

Dont risk using AS3/5 near the rams, if any gets over the edge and under the chip it will short out pins on the BGA and you will be in a world of hurt. Use Ceramique for the rams, it's non-conductive and a mistake wont ruin your day.

AS5 isn't an electrical conductor. Says so right on their own web site.
 
CastleBravo said:
AS5 isn't an electrical conductor. Says so right on their own web site.

it also says to be careful with it, because it CAN be a slight electical conductor becuase of the silver. it is in the directions of application. it warns not to connect two traces with as5.
 
the stock rubber grommets are fine, you just have to tighten the nuts enough (I tightened mine ALOT, but don't go overboard, you could snap them). Use a thin layer on the GPU and a thicker layer on the memory (I used AS5).

Don't forget to plug the fan in before putting the entire unit on.

If installed correctly you should see a noticable drop in temps across the board.

Mine went from idle 59C (@ 410/1100) to 52C idle (@ 435/1200) and max temps went from 83C (@ 410/1100) to 72C (@435/1200) both after 2 straight hours of rthdribl.
 
The rubber washers will work, but not as good as using the old backplate, by using the backplate you won't bottom out the nuts that come with the NV-5.."2mm" If you were to bottom out the nuts it would be really easy to break the studs.....

About the big dual slot bracket, I made a line in the middle and cut it in half.. A hack saw worked great.. I removed it from the video card to do this, so don't try to cut it off while still on the video card, unless you don't mind buying a new one..
 
either way, just be careful when you tighten the nuts, backplate or not, the potential will exist for breakage.

I chose the rubber grommets as the are rubber an can flex/crush under load...the metal backplate will not, but may offer better pressure dispersion (due to its surface area).

I wish AC would test the bolts for tolerance, and give a torque rating (although it would be hard to check the torque on a flathead nut).
 
My concern is that alot of pepole where having problems with there ATI installs,being that the core is naked ,what are the chances of cracking the GPU?
 
OPUS1 said:
My concern is that alot of pepole where having problems with there ATI installs,being that the core is naked ,what are the chances of cracking the GPU?

i didnt even know about this when i did it, and nothing bad happened.. so id say slim, it has a shim around it that has the model name of the card one it. i havent heard of anyone cracking the core with the nv silencer5, then again, there are alot of epople that cant get their hands on one yet..
 
The difference between the rubber washers and using the original metal backplate is anywhere from 3-6C...I believe looking between the NV-5 and the video card and barely touch the four corners of the shim to the base of the NV-5 is very important, but I believe you would bottom out the nuts using the rubber washers and risk breaking the studs....

I had time to do some more testing and have gotten 460mhz/1.2ghz w/ 1.5v 2D & 3D as when Ocing with the 2D and 3D settings different it causes instability...
I tested for stability by looping 3dmark2k1 3 times and then playing Doom3 w/o problems for at least 2 hours...
 
Back
Top