dBA and Bels

Folkert Rienstra wrote in message
...
Mark M wrote
John H. wrote

If any of them answer my original question,
how to convert bels to dBA, I missed it.

From http://tinyurl.com/xis6 ...

========START QUOTE
Values in bel should be reserved for when talking on sound power
level, while values in decibels should be used for when talking on
sound pressure level. Values in bels can also be written followed by
a capital B, but should be written in all small letters when stated
as bels. The abbreviation for decibel is dB (always spelled with a
small d and a capital B). A decibel is a tenth of a bel: 1 B = 10 dB.
While possible to convert between bel and decibel, is it not possible
to convert between sound power level and sound pressure level.
======= END QUOTE

Aha, now I get it. You and CWatters are one and the same
person, It just takes you a day to switch from one to the other.

Or thats what turns up using google, cretin ******.

On Thu, 04 Dec 2003 16:50:18 GMT
John H. wrote:

On Wed, 3 Dec 2003 00:45:11 -0500, "J.Clarke"
wrote:

The difficulty with all of this is that there is no established
standard for measuring the sound pressure or sound power levels of
either drives or fans. Therefore any differences in the reported
measurements may be the result of differences in the test procedures
rather than in the devices themselves.

Actually I think there is.

"ISO 7779 specifies operating and installation conditions in an
acoustical lab in order to have reproducible and repeatable values.
The two noise metrics in ISO 7779 are the A-weighted sound power level
and the A-weighted sound pressure level at specified locations.

ISO 9296 specifies the declaration of noise emissions from information
technology products. ISO 9296 specifies reporting statistical maximum
values of the A-weighted sound power levels based on measurements
taken according to ISO 7779"

Problem is, you have to pay dearly if you want to download the pdf
files.

I believe that you will find that these apply to assembled machines, not
to internal components such as fans and disk drives. In any case, are
those standards "generally accepted", i.e. do all or most manufacturers
adhere to them?

I don't know what you're talking about here. Bel is used for
acoustic power, dB or dBA for acoustic pressure. I haven't used B
anywhere.

That statement makes no sense. "dB" is the abbreviation for
"decibel", which is one tenth of a bel, just as "dm" is the
abbreviation for"decimeter", one tenth of a meter. "Bel" is "B" and
"B" is "Bel" when discussing acoustic measurements. B or Bel is
used for acoustic pressure and acoustic power and many other things.
dB is used for
acoustic pressure and acoustic power and many other things. dB(A)
is normally used for acoustic pressure because that's the only area
in which weighting to provide a measurement comparable to the
response curve of the human ear has real relevance.

Not true. See http://www.silent.se/

Huh? I'm sorry, but I find nothing on that page which contradicts my
statement. But I do find that it contradicts your assertion that there
are generally accepted standards.

For some reason some disk manufacturers have been using B for
acoustic power and dB for acoustic pressure but that is not anything
that they are required to do by any standard or convention.

The reason is because it is a standard. They don't use B, they use
bel. It's supposed to be spelled out.

http://www.silent.se/labels.php

Would you be kind enough to provide a quotation from that page to
support your argument? It appears to me that all labels are to use "B"
rather than the word "Bel"--at least that is what appears on the samples
they provide. And what leads you to believe that that web site is
authoritative in any way?

"The computer industry have intelligently choosen to use the unit bel
to express sound power level values to avoid confusion between
decibels for sound power level and decibels for sound pressure level."

And who is required to comply with this standard?

Which may be why Papst for example doesn't feel any need to tell you
that their bel rating is for power and their dBA rating is for
pressure- it's the standard (for now).

Excuse me but Papst makes fans, they don't make "computer equipment", so
why should they be controlled by a standard that is intended to apply to
the computer industry and to which even most of the computer industry
does not appear to adhere? Or perhaps you are unaware that there are
many other kinds of device in which fans are used. In any case, the
Papst catalog explicitly states: "1. Noise Pressure Level-dB(A) Noise
Ratings of the fan in free air operation, i.e. at maximum flow rate. 2.
Sound Power Level - bels Extent of the overall sound radiation of the
fan. The sound power level is determined in the optimum operating
range." Furthermore the Papst catalog contains no reference to either
7770 nor 9296, leading one to suspect that they are in fact adhering to
neither standard.

Note by the way that this statement alone demonstrates one of the
pitfalls of sound level measurement. The noise level of a fan varies
depending on the backpressure--so at what level of backpressure do you
measure? And do your ISO standards specify this?

I've been following the same
standard. All my bels (not B) are power and all my decibels are
pressure. I'm not sure about the "avoiding confusion" part though,
sure hasn't helped you and Rod any. :-)

Which standard have you been following and please quote the section
which forbids the use of "B" to denote the bel?

--
--
--John
Reply to jclarke at ae tee tee global dot net
(was jclarke at eye bee em dot net)

On Thu, 04 Dec 2003 16:50:28 GMT
John H. wrote:

On Thu, 4 Dec 2003 07:57:41 +1100, "Rod Speed"
wrote:
=20
I agree that a drive with a lower 'belA' rating could still have
higher sound pressure (dBA) in *some* directions (compared
to the drive with the higher 'belA' rating) even though the
average of *many* directions is lower. But the same is true for
bel ratings.
=20
You're mangling the entire story utterly. The only difference
between bels and decibels, B and dB is the usual metric
scale factor. Just like between M and KM etc.
=20
I know the units. I'm going by the industry standard, bel for power,
dBA for pressure, which is even stated in the link you provided and
that was posted twice in this thread. There IS a difference between
bel and decibel (for the computer industry anyway).

Just to be pedantic, (a) the bel is not a metric unit, it is a Bell
Laboratories unit which has come into common use. (b) the bel is
different from the decibel in the same fashion that the meter is
different from the decimeter. Now it may be that by convention some
distances are measured in bels and others in decibels just as it is
convention that distances between cities are measured in kilometers not
meters, but that does not mean that the units are different in kind,
only in magnitude.

The A signifys a quite different application of the A law to the
sound being measured to allow for the ear's frequency response.
=20
I know that.
=20
At the extremes you get quite different effects when
the tip speeds are approaching the speed of sound too.
=20
Which fan is that? For the most powerful 80mm Papst fan (3600 RPM),
the tip speed would be 2.9"*pi*3600*60/12/5280 =3D 31 MPH. If I remember
right, the speed of sound is 746 MPH at sea level.
=20
I still think it's possible [to have a belA value]
=20
You're wrong. It aint.
=20
Oh yes it is. If I had remembered what I read in the
silentpcreview article (that I provided a link to) I could have told
you earlier that the two noise metrics used in the ISO 7779 standard
are belA (A-weighted sound power) and dBA (A-weighted power pressure).
ISO doesn't use the
bel word though, so when this standard starts to be used a drive might
be labeled as "xxdB ISO" which will be understood to mean sound power.
Of course there will still be people who'll say "dB can be used for
both power and pressure" and be confused all to hell.

Interesting. Did you not say earlier that "the standard" requires that
the word "bel" be spelled out?
=20
Thats the whole point of the A letter in dBA, it signifys that
THAT value has the A law filter applied.
=20
Yes but that's pressure, not power. bel can have the A letter too.
=20
B is just Bel. dB is decibel. dB aint the same as dBA.
=20
I _know_.
=20
This seems to make a dBA rating a better indicator of
how loud a drive will appear than what a bel rating is,
=20
Yes.
=20
Only for people "still not informed by those knowing better."
=20
Check out this site: www.silent.se (It appears that a computer was
used to translate to English but it's still very readable)
=20
Some excerpts and comments:
=20
http://www.silent.se/iso-9296.php
"The users still asking for sound pressure level figures seems to be
the ones still not informed by those knowing better."
=20
We're both uninformed. Knowing the power is better than knowing the
pressure because power "values don't depend on distance or user
position." (but I think we both knew that already)
=20
"A-weighted sound power level (LWAd) in bels (B) is the measure best
suited for comparision of noise emissions"
=20
Damn, somebody ripped off my idea. And you said it was impossible.
It seems that A-weighted power values are the ones used most too. I
wonder if bel values for hard drives are already A-weighted (they
*should* be) - no way of knowing for sure.
=20
"When we describe IT acoustic noise emissions in bels is the use of
the A-weighted filter mostly also included, but we seldom state this
fact as BA or bels(A): this because, as said above, when using bel
values for information technology noise emissions are we talking on
sound power level values, and this is stated by putting an "LWAd"
before the values: The A in LWAd states that the A-weighted filter has
been used; making a second A after the B or bels unnecessary."
=20
At one point he says "it not possible to convert between sound power
level and sound pressure level," and at another point he says "sound
power levels are useful...for calculating the sound pressure level
from a machine at a given distance."
=20
http://www.silent.se/labels.php
=20
"The computer industry have intelligently choosen to use the unit bel
for to express sound power level values to avoid confusion between
decibels for sound power level and decibels for sound pressure level."
=20
Which is what I keep saying: bel is used for power, dBA for pressure.
But it looks like this may change soon.
=20
"However, today the computer industry is the only product group that
uses sound power in bels, and if the idea of using dB ISO or dB IEC
will come true will they find themselves using decibel (dB) instead of
bel(B) values for sound power level as intended in their precious
standards. Thus an other maybe better option would be to choose to use
"bels ISO", "B ISO", "bels IEC" or "B IEC" for a simplified marking
standard"=20
=20
So a '3.5 bel' sound power rating today might become '35 dB ISO' in
the future. Since both the noise metrics uses in ISO 7779 are
A-weighted, I guess '35 dB ISO' would mean A-weighted too, with
'A-weighted power' understood.

Here are the tidbits that you missed on your beloved page:

"Most ordinary people are not aware of that a certain dB figure has to
be accompanied with information on how it has been obtained and
declared, for to be able to use it as anything better than the relative
notions "quiet" and "almost=A0silent"."

"Distance, background noise and room conditions are important factors
that will affect measured noise figures. Therefore is it of utmost
importance that one knows what standard has been used for measurement
when talking in exact noise figures. Telling just "noise=A0emission:
20=A0dBA" without informing how this value has been obtained doesn't mean
anything, but since unaware customers will interpret it as information,
is it easy to understand that manufacturers or resellers using figures
this way will mislead them not understanding better."

"To conclude: IT manufacturers or resellers not stating what standard
they have used for noise declaration, and forgetting to declare when
they state sound power level or sound pressure level values, will be the
ones that mislead their customers."

"We are today not lacking standards for how to measure and declare
information technology noise emissions, but we are most often still
lacking the adoption of them:"

And then there is "The ISO 10302 standard, "Acoustics -- Method for the
measurement of noise emitted by small air-moving devices", is the
international accepted one used for noise measurement of fans." This
would in principle be the one that Papst uses. I wonder why their
catalog does not mention it?

Now, do any of the disk manufacturers claim to test in accordance with
the standards that "The Silent PC" thinks they should apply? "The
Silent PC" doesn't seem to think so so why should you, after using them
as your sole reference, disagree with them?


=20
=20


--=20
--=20
--John
Reply to jclarke at ae tee tee global dot net
(was jclarke at eye bee em dot net)

On Thu, 04 Dec 2003 16:50:31 GMT
John H. wrote:

On Wed, 3 Dec 2003 00:26:29 -0500, "J.Clarke"
wrote:

What would really help with HD noise is switching to 2.5"
drives, even 10,000 RPM models.

Why would that help? What leads you to believe that a 2.5"
drive with a given capacity and performance level will be
quieter than a 3.5" drive with the same capacity and performance
level?

Would you have said the same thing back when only 5.25" drives
were available and someone mentioned 3.5" drives?

Yes. But why do you feel that that is relevant?

Because I believe 3.5" drives are quieter than 5.25" drives if the
specs are in the same ballpark.

And why do you believe this?

Lower mass should help a lot

Why would this be? Why do you believe that reducing mass will
reduce noise?

1. A smaller motor for the spindle. Smaller usually means less noisy
if the quality level is the same.

Unless it doesn't. Resonances are strange things--massive objects tend
to have lower resonant frequencies than less massive ones for example.
If reducing the mass puts the resonant frequency into a range which
corresponds to the rotational velocity then you end up with more noise,
not less.

2. Lower spinning mass (which can never be 100.0000% perfectly
balanced) should mean less noise caused by vibration.

So, are 15,000 RPM drives quieter than 10,000 RPM drives? They have
smaller platters so by your reasoning they should be.

3. A shorter arm for the heads. I'd guess that the mass for an arm
increases by the cube of the length. So if the arm is 20% shorter it
may have only half the mass, meaning a _lot_ less energy needed for
seeks and therefore less noise.

Again, do 15,000 RPM drives exhibit the behavior you predict?

If you can't see any difference between 2.5" and 3.5", compare 3.5" to
12" or bigger. I think you'd readily agree that 3.5" drives are much
quieter (not to mention much faster). Why? Because they're smaller.

Uh, doesn't being about 20 generations newer and targetted at a
different market that is more sensitive to noise have anything to do
with it?

Of course this is just an end-user's opinion and I could be 100%
wrong.

and also allow higher performance.

Why do you believe that reducing mass will allow higher performance?

A lighter arm is easier to accelerate and decelerate and the travel
distance would be shorter, meaning a faster access time should easily
be possible.

So how much faster is the access time (neglecting latency, which is not
affected by the arm mechanics) for 15,000 RPM drives with their reduced
platter size than for 10,000 RPM drives? And do current 2.5" drives
exhibit the shorter access times (again neglecting latency) that you
predict?

And what do you believe that I would gain by such a combination over
two 3.5" drives?

Higher performance for your system drive (assuming the 2.5" is
faster).

That's a big assumption.

If it's faster, quieter, and consumes less power than any 3.5"
drive, the market would include just about everybody.

Even if it cost $20,000 each?

No.

If it's faster, quieter, consumes less
power, provides the same amount of storage, and costs the same, then
it has a market. Until then it would be a low-volume niche product
that wouldn't repay the development cost.

It wouldn't need the same amount of storage. The WD 37.5GB 10000RPM
Raptor is very popular and yet cost 4X as much per GB.

How popular is it? Do you have any inforamation about quanties sold?
Does it sell well compared to other IDE drives, or just when compared to
SCSI drives in the enterprise market?

But you have yet to provide a convincing argument that it would be
faster, quieter, or consume less power.

I've leave the "convincing" part to others who know the real facts.

--
--
--John
Reply to jclarke at ae tee tee global dot net
(was jclarke at eye bee em dot net)

On Mon, 01 Dec 2003 18:56:53 GMT, John H. wrote:

I need a physics lesson. Assuming sea level, how do you convert between
sound power (bels) and sound pressure (dBA)? I think you'd have to assume
that the acoustic power is being emitted equally in all directions (which
may not be true for HDs).

WD tells you only the pressure (34 dBA for the 250 GB Caviar SE) while most
other manufacturers tell you only the power (2.5 bel for the DiamondMax Plus
9 and Barracuda V). How do you compare apples and oranges? Which is better
to know, bels or dBA?

I'm thinking that (for HDs anyway) 2.5 bels is less than 25dBA SPL, which
makes the WD very noisy by comparison. (I have a cooling fan that's 3.5
bels and 12dBA)


Sound level specs are notoriously non comparable mostly due to
marketing department baloney.

However, even if you wanted to produce and publish an honest
measurement, you'd have a hard time doing it. The problems aren't
physics lessons, it's the various official test methods and how people
interpret them

I've seen sound testing labs and all that goes into it and it's an art
form.

A lot of people out there are finger painters.

Your ears are your best guide, and remember, if you can't hear it,
it's not a problem.

John H wrote in message
...
Rod Speed wrote

John H wrote

I agree that a drive with a lower 'belA' rating could still have
higher sound pressure (dBA) in *some* directions (compared
to the drive with the higher 'belA' rating) even though the average
of *many* directions is lower. But the same is true for bel ratings.

You're mangling the entire story utterly. The only difference
between bels and decibels, B and dB is the usual metric
scale factor. Just like between M and KM etc.

I know the units. I'm going by the industry standard, bel for power,
dBA for pressure, which is even stated in the link you provided and
that was posted twice in this thread. There IS a difference between
bel and decibel (for the computer industry anyway).

You're still mangling the story completely. There is certainly
a difference between sound pressure values and sound power
levels, but its completely silly to mangle that into the completely
separate use of metric units with bels and decibels.

That silentpc site is clearly where you got that
from, but its still comprehensively mangled.

The A signifys a quite different application of the A law to the
sound being measured to allow for the ear's frequency response.

I know that.

Superficially you do, but if you had actually grasped the concept
you wouldnt be claiming that it is ever going to be possible to
have a single conversion factor between dBA and bels.

I still think it's possible [to have a belA value]

Thats dishonest. You were talking about a
fixed ratio between dBA and bels, NOT belA.

And you're still mangling the difference between sound
pressure levels and sound power levels into the units.

You're wrong. It aint.

Oh yes it is.

Like hell it is.

If I had remembered what I read in the silentpcreview article
(that I provided a link to) I could have told you earlier that the two
noise metrics used in the ISO 7779 standard are belA (A-weighted
sound power) and dBA (A-weighted power pressure).

You originally asked for a conversion between the
two units HARD DRIVE MANUFACTURERS CITE.

Thats nothing like this other distinction between two
different values that have the A law applied to them,
sound pressure levels and sound power levels.

AND that article clearly says that there aint no simple
ratio between even these you are NOW discussing.

If there was, there wouldnt be any need to spend the
substantial amount of money to have the drive tested using
the protocol in the standard and everyone could just apply
the factor to the value the manufacturer chooses to cite.

That Silent PC article clearly states that that isnt possible.

ISO doesn't use the bel word though,

Yep, thats the mangling introduced by Silent PC.

so when this standard starts to be used a drive might be labeled
as "xxdB ISO" which will be understood to mean sound power.

Correct. And thats the correct way to do it, not play silly
buggers with dB and bels. Because there is a lot more
involved than just whether sound pressure levels or sound
power levels are being cited, or whether the A law has been
applied or not. The standard standardises MUCH more than
just that stuff to produce a purportedly standardised test situation.

BUT, whatever is done at that physical level,
IT STILL DOES ABSOLUTELY NOTHING
ABOUT THE MIND'S PERCEPTION OF NOISE.

Which just happens to be a MUCH more important
factor with drives that are already very quiet.

Of course there will still be people who'll say "dB can be used
for both power and pressure" and be confused all to hell.

Its those who attempt to mangle sound pressure and sound
power into the unit used that are massively confused.

bels and decibels arent even unique to sound levels of either type.

Thats the whole point of the A letter in dBA, it
signifys that THAT value has the A law filter applied.

Yes but that's pressure, not power. bel can have the A letter too.

Again, you're massively confusing sound
power and sound pressure and the UNIT used.

The standard does not do that.

This seems to make a dBA rating a better indicator of
how loud a drive will appear than what a bel rating is,

Yes.

Only for people "still not informed by those knowing better."

Utterly mangled all over again.

Check out this site: www.silent.se (It appears that a computer
was used to translate to English but it's still very readable)

Still comprehensively mangles the entirely separate question
of sound pressure and sound power into units used.

The standard doesnt do that.

Some excerpts and comments:

http://www.silent.se/iso-9296.php
"The users still asking for sound pressure level figures seems
to be the ones still not informed by those knowing better."

We're both uninformed.

You have always been. I never ever said anything about that.
I JUST rubbed YOUR nose in the FACT that there cant be
any simple conversion factor and that whatever is quoted is
useless when trying to decide which drive is quieter to a
human, particularly when comparing already quiet drives.

Knowing the power is better than knowing the pressure because
power "values don't depend on distance or user position."

Thats comprehensively mangled all over again, particularly
with user position at the same distance from the drive.

HARD DRIVES DO NOT RADIATE
NOISE UNIFORMLY IN ALL DIRECTIONS.

(but I think we both knew that already)

I'm not silly enough to buy that mangling.

"A-weighted sound power level (LWAd) in bels (B) is the
measure best suited for comparision of noise emissions"

And that mangles the unit story completely. Doesnt matter
a damn if its bels (B) or decibels (dB), what matters is that
its the sound power level, A weighted, thats being stated.

AND to be validly comparable between different drives,
you also need to state the standard protocol that was
used to measure it. And that cant be mangled into the
units used either, it has to be stated explicitly.

AND YOU STILL CANT HAVE A SIMPLE RATIO
BETWEEN THAT STATED MEASURE VALUE AND
THE OTHER ONES THAT APPEAR IN DATASHEETS.

Damn, somebody ripped off my idea.

Like hell they did.

And you said it was impossible.

I said it is impossible to HAVE A SIMPLE RATIO
BETWEEN bels and dBA cited in datasheets.

And that neither allow for the MIND'S PERCEPTION
OF SOUND WITH ALREADY QUIET HARD DRIVES.

It seems that A-weighted power values are the ones used most too.

No surprises there when it would be completely silly
to not allow for the frequency response of the ear.

I wonder if bel values for hard drives are already A-weighted
(they *should* be) - no way of knowing for sure.

Again, thats just plain wrong. Those who measured
it must know if that A weighting was applied.

"When we describe IT acoustic noise emissions in bels is the use of the
A-weighted filter mostly also included, but we seldom state this fact as
BA or bels(A): this because, as said above, when using bel values for
information technology noise emissions are we talking on sound power
level values, and this is stated by putting an "LWAd" before the values:
The A in LWAd states that the A-weighted filter has been used;
making a second A after the B or bels unnecessary."

Pity about the situation where just bel
is used. Utterly mangled all over again.

At one point he says "it not possible to convert
between sound power level and sound pressure level,"

That is correct.

and at another point he says "sound power levels
are useful... for calculating the sound pressure
level from a machine at a given distance."

There isnt any conflict between those.

http://www.silent.se/labels.php

"The computer industry have intelligently choosen to use the unit bel
for to express sound power level values to avoid confusion between
decibels for sound power level and decibels for sound pressure level."

Thats mindlessly silly.

Which is what I keep saying: bel is used for power, dBA for pressure.

Pity thats just plain wrong.

But it looks like this may change soon.

Its never been as true as they claim.

"However, today the computer industry is the only product
group that uses sound power in bels, and if the idea of using
dB ISO or dB IEC will come true will they find themselves
using decibel (dB) instead of bel (B) values for sound power
level as intended in their precious standards. Thus an other
maybe better option would be to choose to use "bels ISO",
"B ISO", "bels IEC" or "B IEC" for a simplified marking standard"

That last is the only approach that makes any sense at all.

So a '3.5 bel' sound power rating today might become '35 dB ISO' in the
future. Since both the noise metrics uses in ISO 7779 are A-weighted, I
guess '35 dB ISO' would mean A-weighted too, with 'A-weighted power'
understood.

Sure. AND THERE IS STILL NO SIMPLE FACTOR THAT CAN
BE APPLIED BETWEEN THE STATED VALUES IN DATASHEETS.

In spite of what you have claimed all along.

John H wrote in message
...
J.Clarke wrote

What would really help with HD noise is switching
to 2.5" drives, even 10,000 RPM models.

Why would that help? What leads you to believe that a 2.5" drive
with a given capacity and performance level will be quieter than a
3.5" drive with the same capacity and performance level?

Would you have said the same thing back when only 5.25"
drives were available and someone mentioned 3.5" drives?

Yes. But why do you feel that that is relevant?

Because I believe 3.5" drives are quieter than
5.25" drives if the specs are in the same ballpark.

You're wrong. Again.

There are plenty of examples of 2.5" drives that are NOISIER
than 3.5" drives that have much better performance.

Lower mass should help a lot

Why would this be? Why do you believe
that reducing mass will reduce noise?

1. A smaller motor for the spindle. Smaller usually
means less noisy if the quality level is the same.

Mindlessly superficial, particularly when the 3.5"
drive is already inaudible when just rotating.

Pointless paying more for the 2.5" drive in that situation.

2. Lower spinning mass (which can never be 100.0000%
perfectly balanced) should mean less noise caused by vibration.

Pity about the situation where the 3.5"
drive is already inaudible vibration wise.

Pointless paying more for the 2.5" drive in that situation.

3. A shorter arm for the heads. I'd guess that the mass
for an arm increases by the cube of the length. So if the
arm is 20% shorter it may have only half the mass, meaning
a _lot_ less energy needed for seeks and therefore less noise.

Pity about the situation where the 3.5" drive
is already inaudible head movement wise.

Pointless paying more for the 2.5" drive in that situation.

If you can't see any difference between 2.5" and 3.5", compare 3.5"
to 12" or bigger. I think you'd readily agree that 3.5" drives are much
quieter (not to mention much faster). Why? Because they're smaller.

Utterly mangled all over again.

When you have a 3.5" drive thats so quiet that you have to feel
the drive to check if its actually working, its pointless spending
more to have the 2.5" drive with the same performance if you
dont care about power use, as you dont with a desktop system.

Of course this is just an end-user's opinion and I could be 100% wrong.


Yep, yet again.

and also allow higher performance.

Why do you believe that reducing mass will allow higher performance?

A lighter arm is easier to accelerate and decelerate
and the travel distance would be shorter, meaning a
faster access time should easily be possible.

There is more involved with access times than just that.

If it's faster, quieter, and consumes less power than any
3.5" drive, the market would include just about everybody.

Have fun explaining why we dont use 2.5" drives exclusively.

There might just be a reason why we dont.

If it's faster, quieter, consumes less power, provides the
same amount of storage, and costs the same, then it has
a market. Until then it would be a low-volume niche
product that wouldn't repay the development cost.

It wouldn't need the same amount of storage.
The WD 37.5GB 10000RPM Raptor is very popular

Only with fools that havent got a clue.

Its a niche market dud.

and yet cost 4X as much per GB.

Yep, and only fools buy it.

But you have yet to provide a convincing argument
that it would be faster, quieter, or consume less power.

I've leave the "convincing" part to others who know the real facts.

Completely trivial to compare the currently available
2.5" and 3.5" drives on faster and price right now.

Three guesses which format wins ? And when there are 3.5"
drives that are already so quiet that it isnt easy to work out
if they have spun up or not, there's no point in spending more
for the same performance in 2.5" format unless you need the
smaller form factor for a tight packaging situation.

On Fri, 05 Dec 2003 16:27:04 GMT
John H. wrote:

Let's just wait and see what happens in the years ahead. You and Rod
obviously think that 3.5" is the lower size limit for high performance
hard drives.

In what post did I make such a statement? Asking someone to support
his arguments does not necessarily mean that I disagree with them.

You may be right, then again you may not. Time will
tell.

What I'd really like to see is the elimination of all rotating parts
in a computer, HDs, DVDs, fans, the works. This will happen too
someday(although I'm sure you guys would disagree ).

If you will review the thread you will find that I made such a
suggestion myself a few posts back.

On Wed, 3 Dec 2003 22:32:06 -0500, "J.Clarke"
wrote:

On Thu, 04 Dec 2003 16:50:31 GMT
John H. wrote:

On Wed, 3 Dec 2003 00:26:29 -0500, "J.Clarke"
wrote:

What would really help with HD noise is switching to 2.5"
drives, even 10,000 RPM models.

Why would that help? What leads you to believe that a 2.5"
drive with a given capacity and performance level will be
quieter than a 3.5" drive with the same capacity and
performance level?

Would you have said the same thing back when only 5.25" drives
were available and someone mentioned 3.5" drives?

Yes. But why do you feel that that is relevant?

Because I believe 3.5" drives are quieter than 5.25" drives if the
specs are in the same ballpark.

And why do you believe this?

Lower mass should help a lot

Why would this be? Why do you believe that reducing mass will
reduce noise?

1. A smaller motor for the spindle. Smaller usually means less
noisy if the quality level is the same.

Unless it doesn't. Resonances are strange things--massive objects
tend to have lower resonant frequencies than less massive ones for
example. If reducing the mass puts the resonant frequency into a
range which corresponds to the rotational velocity then you end up
with more noise, not less.

2. Lower spinning mass (which can never be 100.0000% perfectly
balanced) should mean less noise caused by vibration.

So, are 15,000 RPM drives quieter than 10,000 RPM drives? They have
smaller platters so by your reasoning they should be.

3. A shorter arm for the heads. I'd guess that the mass for an arm
increases by the cube of the length. So if the arm is 20% shorter
it may have only half the mass, meaning a _lot_ less energy needed
for seeks and therefore less noise.

Again, do 15,000 RPM drives exhibit the behavior you predict?

If you can't see any difference between 2.5" and 3.5", compare
3.5" to 12" or bigger. I think you'd readily agree that 3.5"
drives are much quieter (not to mention much faster). Why? Because
they're smaller.

Uh, doesn't being about 20 generations newer and targetted at a
different market that is more sensitive to noise have anything to do
with it?

Of course this is just an end-user's opinion and I could be 100%
wrong.

and also allow higher performance.

Why do you believe that reducing mass will allow higher
performance?

A lighter arm is easier to accelerate and decelerate and the
travel distance would be shorter, meaning a faster access time
should easily be possible.

So how much faster is the access time (neglecting latency, which is
not affected by the arm mechanics) for 15,000 RPM drives with their
reduced platter size than for 10,000 RPM drives? And do current
2.5" drives exhibit the shorter access times (again neglecting
latency) that you predict?

And what do you believe that I would gain by such a combination
over two 3.5" drives?

Higher performance for your system drive (assuming the 2.5" is
faster).

That's a big assumption.

If it's faster, quieter, and consumes less power than any 3.5"
drive, the market would include just about everybody.

Even if it cost $20,000 each?

No.

If it's faster, quieter, consumes less
power, provides the same amount of storage, and costs the same,
then it has a market. Until then it would be a low-volume niche
product that wouldn't repay the development cost.

It wouldn't need the same amount of storage. The WD 37.5GB
10000RPM Raptor is very popular and yet cost 4X as much per GB.

How popular is it? Do you have any inforamation about quanties
sold? Does it sell well compared to other IDE drives, or just when
compared to SCSI drives in the enterprise market?

But you have yet to provide a convincing argument that it would
be faster, quieter, or consume less power.

I've leave the "convincing" part to others who know the real
facts.

--



--
--
--John
Reply to jclarke at ae tee tee global dot net
(was jclarke at eye bee em dot net)

On Fri, 05 Dec 2003 16:27:21 GMT
John H. wrote:

John and Rod,

Your comments on sound have been interesting and some even helpful,
but since we can't even agree on things that should be undisputable,
e.g. the computer industry using bel for power and dBA for pressure,
there's little hope for agreement on much else. So I'll take a pass
on any more detailed replies. Thanks--JH

Therein lies the problem--you are taking as "undisputable" things which
(a) don't really merit dispute and (b) are most assuredly "disputable".
You can use bels, decibels, centibels, millibels, exabels, petabels, B,
dB, cB, mB, or any other abbreviation for or multiple of the bel and you
still end up with the same value--it doesn't matter which you use from a
computational or engineering viewpoint except to the extent that picking
the right unit means that you have to write or key fewer zeros than with
the others. And as for it being "undisputable", I don't think that
anybody disputed that the bel was commonly used for sound power levels
and the A weighted decibel for sound pressure levels, it's your
insistence that the bel is different from its abbreviation B and that
the A weighted decibel is somehow different from its abbreviation
"dB(A)" or "dBA" that is being disputed. That is like saying that I am
somehow a different person if I call myself "JC" than if I call myself
"John Clarke".

On Wed, 3 Dec 2003 21:58:11 -0500, "J.Clarke"
wrote:

On Thu, 04 Dec 2003 16:50:18 GMT
John H. wrote:

On Wed, 3 Dec 2003 00:45:11 -0500, "J.Clarke"
wrote:

The difficulty with all of this is that there is no established
standard for measuring the sound pressure or sound power levels
of either drives or fans. Therefore any differences in the
reported measurements may be the result of differences in the
test procedures rather than in the devices themselves.

Actually I think there is.

"ISO 7779 specifies operating and installation conditions in an
acoustical lab in order to have reproducible and repeatable
values. The two noise metrics in ISO 7779 are the A-weighted sound
power level and the A-weighted sound pressure level at specified
locations.

ISO 9296 specifies the declaration of noise emissions from
information technology products. ISO 9296 specifies reporting
statistical maximum values of the A-weighted sound power levels
based on measurements taken according to ISO 7779"

Problem is, you have to pay dearly if you want to download the pdf
files.

I believe that you will find that these apply to assembled machines,
not to internal components such as fans and disk drives. In any
case, are those standards "generally accepted", i.e. do all or most
manufacturers adhere to them?

I don't know what you're talking about here. Bel is used for
acoustic power, dB or dBA for acoustic pressure. I haven't
used B anywhere.

That statement makes no sense. "dB" is the abbreviation for
"decibel", which is one tenth of a bel, just as "dm" is the
abbreviation for"decimeter", one tenth of a meter. "Bel" is "B"
and"B" is "Bel" when discussing acoustic measurements. B or Bel
is used for acoustic pressure and acoustic power and many other
things.
dB is used for
acoustic pressure and acoustic power and many other things.
dB(A) is normally used for acoustic pressure because that's the
only area in which weighting to provide a measurement comparable
to the response curve of the human ear has real relevance.

Not true. See http://www.silent.se/

Huh? I'm sorry, but I find nothing on that page which contradicts
my statement. But I do find that it contradicts your assertion that
there are generally accepted standards.

For some reason some disk manufacturers have been using B for
acoustic power and dB for acoustic pressure but that is not
anything that they are required to do by any standard or
convention.

The reason is because it is a standard. They don't use B, they
use bel. It's supposed to be spelled out.

http://www.silent.se/labels.php

Would you be kind enough to provide a quotation from that page to
support your argument? It appears to me that all labels are to use
"B" rather than the word "Bel"--at least that is what appears on the
samples they provide. And what leads you to believe that that web
site is authoritative in any way?

"The computer industry have intelligently choosen to use the unit
bel to express sound power level values to avoid confusion between
decibels for sound power level and decibels for sound pressure
level."

And who is required to comply with this standard?

Which may be why Papst for example doesn't feel any need to tell
you that their bel rating is for power and their dBA rating is for
pressure- it's the standard (for now).

Excuse me but Papst makes fans, they don't make "computer
equipment", so why should they be controlled by a standard that is
intended to apply to the computer industry and to which even most of
the computer industry does not appear to adhere? Or perhaps you are
unaware that there are many other kinds of device in which fans are
used. In any case, the Papst catalog explicitly states: "1. Noise
Pressure Level-dB(A) Noise Ratings of the fan in free air operation,
i.e. at maximum flow rate. 2. Sound Power Level - bels Extent of
the overall sound radiation of the fan. The sound power level is
determined in the optimum operating range." Furthermore the Papst
catalog contains no reference to either 7770 nor 9296, leading one
to suspect that they are in fact adhering to neither standard.

Note by the way that this statement alone demonstrates one of the
pitfalls of sound level measurement. The noise level of a fan
varies depending on the backpressure--so at what level of
backpressure do you measure? And do your ISO standards specify
this?

I've been following the same
standard. All my bels (not B) are power and all my decibels are
pressure. I'm not sure about the "avoiding confusion" part
though, sure hasn't helped you and Rod any. :-)

Which standard have you been following and please quote the section
which forbids the use of "B" to denote the bel?

--



--
--
--John
Reply to jclarke at ae tee tee global dot net
(was jclarke at eye bee em dot net)

Previously John H. wrote:
On 1 Dec 2003 20:56:06 GMT, Arno Wagner wrote:

[...]
dBA is actually deci Bel (A). So disregarding the (A) for the moment,
10 dB = 1 Bel. The (A) is a weighting curve that reflects the human
ear's sensitivity, so dB(A) is more honest than dB or Bel. The problem
is that the human ear's sensitivity is only loosely conectet to the human
mind's resulting anoument. For that you need a measurement in "Sone".
The german computer magazine c't regularly lists dB(A) and Sone in
its HDD tests and there are drives that have good dB(A) ratings but
only not so good Sone ratings. E.g. a high-pitched whine will cause
that.

If I read this article right, sones can directly be converted to phons, and
there's not much difference between phons and dBA for low to moderate sound
levels (which hopefully includes the HD).

http://www.phys.unsw.edu.au/~jw/dB.html

I still would not trust dB(A) too much.

Arno
--
For email address: lastname AT tik DOT ee DOT ethz DOT ch
GnuPG: ID:1E25338F FP:0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
"The more corrupt the state, the more numerous the laws" - Tacitus