More detailed sandwich questions

Silencing hard drives, optical drives and other storage devices

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
schnee
*Lifetime Patron*
Posts: 41
Joined: Sat Nov 02, 2002 12:18 am

More detailed sandwich questions

Post by schnee » Sat Nov 16, 2002 4:57 pm

Ok, hate to be so a.r. but some things concern me about the sandwich because I have WD drives without Dtemp support, and I want to clarify some details before I do something to kill $300 hardware and make my life miserable.

I have a Coolermaster 110 aluminum case. The dual WD drives are bolted to aluminum vertical panels, straight in the path of dual 80mm fans. So, this acts like 'side sinks' and has direct airflow over the drives. The sandwiches I build will sit in that same place, vertically oriented, directly on some Magic Fleece (or perhaps suspended).

1) Heatsink properties & cork: It seems to me that having a layer of cork or other insulating material between the aluminum plate and the drive would prevent the heatsink behavior of a sanwich from happening, and may worsen it... I mean, it's like putting a blanket there. Wouldn't it cook the drives?

2) One thing mentioned: sandwiches only really help with foam stuffed on the sides. So, having plates the exact width of the drives on top and bottom, lined with Dynamat for noise-supression, coupled with U-channel side-sinks wouldn't achieve any noise reduction?

3) What about big, thick plates bolted to the sides too? Wouldn't that have the same noise-canceling as foam with the added benefit of heat transfer like a side sink?

SungHyun7
Posts: 149
Joined: Sat Nov 09, 2002 1:45 pm
Contact:

Post by SungHyun7 » Sat Nov 16, 2002 5:25 pm

what's a.r.? anyway here are my inputs.

if you can wait a month or so, wait for seagate v. they're quiet, and very comparable performance-wise to most drives. my point is this. say you get the WD drives, sandwich them, and still think they're loud, you now know you should've gotten the seagate drives. if you get seagate v's, and not satisfied, you can sandwich them, which to most of us becomes 'virtually inaudible'. choice is upto you. but for me, waiting for the barracuda v just seems so sensible at this point. price will be almost the same, they'll be much quieter, they perform as well if not better, they WILL come out with 8mb cache version within a month, etc. etc. you get the best of all worlds.

1) you shouldn't automatically assume that non-conductors are always insulators. in fact, if you use a thin sheet (just enough for mechanical decoupling), they will conduct heat quite well. check out the discussion on 'smart drive' which would prove my point. they may not be great heat conductors, but they do conduct a bit, it seems. if you're that worried, get silicon gel (someone please get the japanese guy's website for this?). silicon gel will conduct heat but not electricity, which would be most ideal.

2) i'm sure there are other creative designs. in any drive silencing experiment, you have three main goals: mechanical decoupling (suspension), noise blocking (enclosure of some kind), and efficient heat dissipation. you just need to find some ground where you can satisfy all three to your taste. sandwich just happens to be a good compromise of those three points that satisfies many of us.

quokked
Posts: 333
Joined: Sun Aug 11, 2002 3:26 pm
Location: Melbourne, Australia
Contact:

Re: More detailed sandwich questions

Post by quokked » Sat Nov 16, 2002 9:20 pm

schnee wrote:I have WD drives without Dtemp support, and I want to clarify some details before I do something to kill $300 hardware and make my life miserable.


One of the cheapest ways I found to measure the temperatures of stuff in my case is to get one of those "inside outside" thermonitors and use the outside probe on parts of my case, the one I got has a thermistor on the outside sensor and I just wedge it into spots that I want to measure, granted it's not the most accurate way of measuring temps in the world but u can also take out the thermistor of its housing and use it direct :)

PS. do NOT try to measure the core temps of a CPU with this setup unless you're absolutey certain that the probe is not interfering with the CPU to Heatsink Contact :)

schnee
*Lifetime Patron*
Posts: 41
Joined: Sat Nov 02, 2002 12:18 am

Post by schnee » Sun Nov 17, 2002 9:59 am

what's a.r.?
Anal Retentive. Psyche joke about someone who worries about detail too much. ;)
you shouldn't automatically assume that non-conductors are always insulators
I can't see cork having better heat transfer than aluminum (heck, I've seen it used as 'counter protectors' for hot pans and such). I assumed the layer of cork would limit the effectiveness of the top and bottom plates as heatsinks. But then again, I'm not an engineer either. (shrugs)

I was hoping you wouldn't say that about buying the Barracudas. I don't want to re-purchase hard drives (and go through the rigamarole of selling the others), but it looks like it's probably my best bet. :( Damn the industry and their new 1-year warranties! :x

Thanks.

Melchior
Posts: 23
Joined: Sun Oct 20, 2002 3:58 pm
Location: Vancouver, BC
Contact:

Post by Melchior » Wed Nov 20, 2002 9:18 pm

Incidentally, as I've mentioned in other threads, cork has a very poor noise-reduction coefficient. Medium-density fiberglass (such as that used in ceiling panels-- cheap at home stores) is an **order of magnitude** better than cork at reducing the amount of noise bouncing around.

schnee
*Lifetime Patron*
Posts: 41
Joined: Sat Nov 02, 2002 12:18 am

Post by schnee » Thu Nov 21, 2002 9:16 pm

Actually, I plan on using Dynamat. The 'cork' was in relation to the original idea.

The fiberglass idea is good. I'm thinking of using it as the 'filler' around the edges of the sandwich. I"ll have to do a bit of research to figure out the best way to construct it... I don't want dust and other crud all over; maybe a resin and black paint treatment, so it looks good too (might put a window on the case).

Post Reply