You are here

Hard Drive Silencing: Sandwiches & Suspensions

March 24, 2002

Along with the CPU heatsink-fan and the power supply unit, the hard drive is a major noise source in a typical PC. It whirrs and whines when just sitting idle, and clatters and grinds when in seek or write mode. There are many ways to reduce hard drive noise. The simplest and most obvious of these is to replace your existing drive with one that is quieter, such as a Seagate Barracuda IV, currently by far the quietest hard drive available. Still, even the quietest hard drives make some noise and vibrate, so drive silencing is useful in most PC noise reduction projects.


Sandwiching the Hard Drive

A solution I created some time ago is to clamp the hard drive,
top and bottom, between two heavy aluminum plates, and line all exposed HD surfaces
with high density sound insulation. This technique basically blocks the sound,
and effectively muffles the noise of a hard drive, especially the higher frequency
noise. It works best when the clamped drive assembly is decoupled from the rest
of the PC case to ensure that the vibrations from the drive don't cause case
panels to make noise as well. This is achieved by simply placing the sandwich
HD assembly on a thin piece of foam. A surprising side benefit is a high degree
of cooling for the clamped drive -- the aluminum plates act as heatsinks and
draw the heat away from the drive very effectively.

All of this is detailed and documented in an article originally published at Overclockers.com and updated here in HD Sandwich Redux. The drives I quieted in this way were IBM 75GXPs -- several 30G & 40G models.

Since then, I discovered the Seagate Barracuda IV hard drives. These are all the rage among silent computing aficionados. The single platter models are rated for a noise level of just 2 bels (20 dBA) at idle and 2.4 bels during seek. These numbers are at least 6 dBA better than any cited by IBM for its 75GXP drives. That is a very significant, very audible difference.

After some weeks of experimentation with the 20G and 40G models, I was convinced
of the Barracuda IV's clear noise superiority over not only the IBMs but all
the other recent model 7200 rpm drives I had in my possession: Maxtor, Western
Digital, Quantum. The absence of any high frequency whining, so prevalent with
all the other drives, is remarkable. It also is at least as fast as the IBMs,
which were my favorite for over two years.

Eventually, all the other HDDs in my systems were replaced with Seagate Barracuda
IVs. I felt the improvement in noise was good enough that I abandoned the aluminum
drive sandwiches.

I did notice that mounted normally inside the PC cases, the Barracuda drives still emitted some audible noise. The noise was lower in frequency than with the previous drives, and much quieter, but still noticeable. I discovered that the Barracuda was held in my hand, it was much quieter. This led me to the conclusion that vibrations transmitted from the drive to the case were causing added noise.


Suspending the Hard Drive

The NoVibe drive mounting bracket by Noise Control was something I'd seen on the web before, but now I paid attention. This review at Dan's Data helped me understand fully how the device works and led me to believe I could fabricate a copy that would work just as well

After consulting my better half, I visited the neighborhood sewing supply store and bought several different elastic materials for about $1 a meter. I splurged on 5 meters. :) It comes in different cross sections, shapes, thicknesses and elasticity. It is normally used in clothing, but I used it to rig up a kind of webbing in a 5.25" drive bay. Pictures tell the story so much better here...




Multistranded core and woven poly-fabric sleeve makes it extremely tough,
yet stretchy.



Simple to knot in a loop.

As the pictures show, the drive is essentially suspended on the stretched elastic. The resilience of the elastic stops all vibrations from passing from the drive to the case -- or vice versa, for that matter. The drive cage was rigged it up with two Barracuda drives (20G & 40G) and installed with a couple of machine screws at the bottom of my main PC with a decoupled low power 80mm fan in front of it (at the front panel intake hole). The system became quieter than before, but not dramatically, possibly because the Barracuda drives are already so quiet.

As a simple test, when I touch the drive while it's running, I can feel a fair amount of vibration. But when I touch touch the case right next to the hard drive -- or any part of the case, there is NO vibration from the drive. None at all.

When I showed one of my suspended drive systems to my favorite local dealer, it was the complete absence of vibration in the case that amazed them the most. They could not tell when the PC was turned on by the usual vibration of the case. They found it eerie.

In contrast, as soon as my either of the normally mounted CD drives are spinning (32X and 48X), I can feel it all through the case -- and hear parts of the case buzzing/vibrating. Ditto a backup HD (also a 'cuda) mounted in a removable IDE rack (that normally stays removed).

NOTE: Compared to standard mounting, it appears that suspending a
hard drive increases its temperature somewhat, perhaps around 5-6° C. Standard firm mounting to the chassis does provide for some cooling by conduction. Maximum operating temperature for modern drives is usually specified at ~60° C, and it is not unusual to see temperatures in the 40° C range. All thing
being equal, higher temperatures means increased failure risk and reduced life.
You may wish to monitor HD temperatures using software such as
D Temp
or an external temperature probe. Alternatively, just mount a slow quiet fan to softly blow some outside air across the drive(s); it take very little airflow to drop a HDD's temperature by as much as 10°C.

Sandwiching and Suspending

Of course, being a perpetual tinkerer, I started thinking about combining the
aluminum sandwiching technique with the no-vibration decoupled mounting to see
if it would give me the best of both worlds. ;) It was just a matter of time.
Recently, I tried it. The last picture shows a drive sandwiched and placed on
the elastic suspension. Later, I suspended the drive in the elastic, sandwiched.



Does it work?

Well, I think so. The problem I am faced here, you see, is that the system
was so quiet to begin with that it is really hard for me to tell the difference.
The Barracuda suspended in elastic makes a soft whirr that can be heard when
I get within 18" with the front open. With the drive sandwiched as well, I can
say that qualitatively, the noise does not seem to have changed, but I have
to be within 12" to hear it, not 18". I tried sound level measurements, but
even with the microphone right next to the drive, it is difficult to stabilize
the readings. It looks like my sound level meter is reading 3-4 dB lower with
the drive sandwiched and suspended (rather than just suspended), but I need
to develop a more stable noise measuring technique before I can report conclusively.
For now, let's just say I believe the combo treatment is a little quieter,
and that's good enough for me. ;)

By Mike Chin

 

POSTSCRIPT April 14, 2008

Six years after it was originally posted and some 200,000 pageviews later,
this article has been rediscovered yet again, and a rash of links to it
have appeared on many web sites in the last day or two. Curiously it is
one of the very first I wrote on the topic of quieting a computer, and
the concept has been widely disseminated and employed by SPCR readers
for years. A search in the SPCR forums for the term suspend or
elastic will turn up literally hundreds of posts and threads. It
seems appropriate to add some notes regarding hard drive noise today compared
to 2002, and the relevance of the solutions discussed in the original
article.

First, desktop hard drives have generally become far quieter. The wholesale
adoption of fluid dynamic bearings for desktop drives by all the manufacturers
over the past couple of years has been the biggest cause, but there have
also been advances made in all the moving components. Naturally, the Barracuda
IV pictured in the photos above is long discontinued and displaced
by much faster drives, even though its acoustic performance has not been
surpassed significantly. As a result of the acoustic improvements, the
aluminum plate sandwich noise-blocking and damping technique is no longer
very useful. It is best applied to hard drives that are noisier than most
of today's hard drives.

The elastic suspension I described, however, is as useful as ever. It
reduces HDD noise very effectively, in idle as well as in seek, but particularly in seek. The effect is still often dramatic, even with today's much quieter hard drives.

SATA 2.5" notebook drives are an increasingly viable alternative
to standard desktop drives. The SATA and power connectors are identical,
they generate just a fraction of the heat, they run just about as fast
(now at 7200 and 5400 rpm), and they are usually quieter than even the
quietest 7200 rpm desktop drives. Some still benefit from elastic suspension, however. See other articles in this section (Storage).

There are many related posts and discussion threads in the Silent PC
Review forums which you may find worth exploring. Some are
extended discussions of alternative ways of mechanically decoupling the
drive from the case (which is what the elastic suspension achieves). Much
creativity by forum members appears in these discussions.

The
Silent Storage Forum


HDD
vibration & noise reducing methods - ranked


HDD
Elastic Suspension... Show your pics!

 

Discuss this article in the SPCR Forums.

Sections: 

Google

www SPCR