First attempt, A64 w. 3700BQE

Show off your quiet rig.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
freka586
Posts: 51
Joined: Sun Nov 21, 2004 11:04 pm

First attempt, A64 w. 3700BQE

Post by freka586 » Sat Apr 02, 2005 4:46 am

This is my first attempt at a reasonable compromise between noise, cost and performance.
Comments and answers to my questions on RMClock and resonance noise are most welcome!

Components:

Case: Antec SLK3700BQE + Acoustipack Pre-cut
PSU: Stock SLK3700BQE PSU
CPU: Winchester A64 3500+, Zalman CNPS 7000-B CU-Led
Motherboard: ASUS A8N-SLI Deluxe + Zalman NB47J
Graphics card: Leadtek GF6600 GT + Zalman VF700-CU
HDD: 2 x Samsung 160 GB SATA (1 Nidec, 1 JVC)
RAM: 4 x 512 PC3200 TwinMOS CAS2.5
Audio: SB Audigy 2ZS
Optical: Pioneer DVD-RW DVR-109

Pictures:

Image

Image

I guess the cable-gami could be improved but I was getting eager to start wrapping up.

Temps:

Full load (sev. hours of Prime95 torture):
* CPU 52 C
* NB 36 C
* GPU 52 C (IIRC, from Futuremark 2005)

Idle:
* CPU 32 C
* NB 36 C
* GPU 38 C
* Ambient: 32 C (?)

Temps are reported from Asus Probe for CPU and NB, and from Nvidia Display Properties for GPU and ambient.
How do these temps sound to you? If I in the future go for slower/quiter fans, how's my "headroom" temp wise?


I've been playing a bit with RMClock and Cool'n'Quiet.
My corrent config uses a tad OC on the FSB (215 vs 200 stock) but also reduced voltage (1.3 vs 1.4)
I have set PST to [email protected] for minimal state and [email protected] for maximal.
How do these figures sound?
One thing is odd though. When I run Prime95, RMClock reports 1.3V (my max state) but Asus probe reports 1.42V.
Is such a difference normal, or is maybe the MB somehow forcing the CPU to override RMClocks settings for the normal ones?


Things that should be done without additional costs is mainly taping/sealing the unused 5.25 an 3.5 slots.
I could also swap the case fan for a Nexus or something and of course swap the PSU.
Since I at this point don't feel like opening the case again and start messing, what do you people reckon is the noisiest component in the setup above?

For the most part the setup is very pleasing acoustically. There is however a very slow, pulsating sound possible from some kind of resonance.
The noise frequency is something like a bit less than once a second.
The resonance seems to be reduced if I open the front door, but if I do that other sounds increase a bit.
Any clues?

ilh
Posts: 313
Joined: Sat Jan 08, 2005 3:31 pm
Location: Acton, MA, USA -- Folding for SPCR

Post by ilh » Sat Apr 02, 2005 6:28 am

My guess on the noise is that it is the noises from the two hard drives "beating" against eachother. If you try to run with one drive off, does it sound different?

If so, a softer mounting (e.g., sorbothane or suspension) would likely knock it out.

len509
Posts: 101
Joined: Mon Aug 30, 2004 12:28 am
Location: Central U.S.

Post by len509 » Sat Apr 02, 2005 8:57 am

Good job freka586. You should definately try to suspend your hard drives. It should be easy on a 3700BQE. String up some Stretch Magic, Mini Bungee Cords or elastic across the drive bays.

Your temps look good. I'd go ahead and start slowing your fans down and try to to get your CPU around 38/39 C.

m0002a
Posts: 2831
Joined: Wed Feb 04, 2004 2:12 am
Location: USA

Post by m0002a » Sat Apr 02, 2005 9:30 am

If you are not comfortable suspending your drives, try this method which is specifically designed for the drive trays you have.
http://forums.silentpcreview.com/viewtopic.php?t=10450

freka586
Posts: 51
Joined: Sun Nov 21, 2004 11:04 pm

Post by freka586 » Tue Apr 05, 2005 2:32 am

The sorbothane thingy seems like a good solution.
Problem is where to find it in Sweden/Europe without spending too much...

Aleksi
Friend of SPCR
Posts: 889
Joined: Sun Dec 05, 2004 11:34 pm
Location: Finland -- Folding For SPCR

Post by Aleksi » Tue Apr 05, 2005 4:28 am

Dorothy Bradbury brought up one source: British audio products. 15x15cm sheet is ~25€ delivered to scandinavia. Well, not cheap by any means, but one of the few sorbothane sources in europe.

m0002a
Posts: 2831
Joined: Wed Feb 04, 2004 2:12 am
Location: USA

Post by m0002a » Tue Apr 05, 2005 6:42 am

freka586,

I notice that you are reporting your GPU temps. Did you have to zap your video card bios to set the flag on to report that, or did the Leadtek GF6600 GT report the temps out of the box?

freka586
Posts: 51
Joined: Sun Nov 21, 2004 11:04 pm

Post by freka586 » Tue Apr 05, 2005 7:55 am

On the sorbo front I have found some in Sweden for 200 SEK (about 20 euros), I guess that is pretty OK.
Not sure when I get round to buying some though...
OT: How do I type the euro sign?

I have not done anything special to get GPU temps, they are shown in the corresponding tab of the NVIDIA display properties.
Not 100% sure they are correct, but they seem reasonable and rise some whenever I play games/tech demoes so....
Short answer: Yes, Leadtek seems to report temps out-of-the-box.

Tibors
Patron of SPCR
Posts: 2674
Joined: Sun Jul 04, 2004 6:07 am
Location: Houten, The Netherlands, Europe

Post by Tibors » Tue Apr 05, 2005 9:22 am

OT: ALT-0128 -> €

gasman
Posts: 3
Joined: Wed Jan 05, 2005 8:41 pm
Location: West of Middle-earth

Re: First attempt, A64 w. 3700BQE

Post by gasman » Thu Apr 07, 2005 8:37 am

freka586 wrote: I've been playing a bit with RMClock and Cool'n'Quiet.

One thing is odd though. When I run Prime95, RMClock reports 1.3V (my max state) but Asus probe reports 1.42V.
Is such a difference normal, or is maybe the MB somehow forcing the CPU to override RMClocks settings for the normal ones?

Any clues?
hi freka
nice!
just wondering:
do you enable both RMClock and C'n'Q?

as for temp difference, apparently ok - up to 0.1V variance common, other Asus boards do the same, monitoring issue only. I'm getting 0.85V on Asus utility and 0.9V on RMClock, SiSandra says 0.9V as well.

cheerio

freka586
Posts: 51
Joined: Sun Nov 21, 2004 11:04 pm

Post by freka586 » Thu Apr 07, 2005 10:40 am

Well, I assume CnQ is applied automatically once enabled in bios and the OS driver is in place.
At least the state transitions seem to work properly...

I guess I should try a few more monitoring programs to see if the numbers make sense...

The computer seemed to have problems getting past POST sometimes with the increased FSB, so I turned it down a notch.
Since I'm not that big a power-user I don't really need the extra few MHz..

Post Reply