I would like to know the way to reduce the speed of a fan inside a
small device in order to cut down the noise from the fan. If I
understand this correctly, I am supposed to do this by undervolting the
fan from 12-volt it is now to something like 6-volt using some kind of
"inline resistor" or someting. I can soldering wires together. But I
really don't know much about resistor and such. I am hoping someone can
give me the "exact" information about the way to do this.

The small device is a LinkSys gigabit switch that has a very noisy 40mm
fan in it. I have replaced it with a slightly quieter Papst 60mm fan.
My intention is to oversize the fan and then reduce the fan speed; then
I will be able to maintain the same air flow as the 40mm fan but with
less noise. Now, I have the oversized fan mounted inside the gigabit
switch. And I can feel that its air flow is definitely stronger than
the 40mm fan. This means I can go ahead to reduce its speed without
worrying about the possibility of not getting enough air flow. I need
to figure out how to reduce its speed.

My questions are:

-  I am under the impression that I should use something
   called an "inline resistor". Is it the right product?

-  Because the gigabit switch has very limit space
   inside, the "inline resistor" must be small enough
   to fit inside. Will it be small enough?

-  The fan is rated as 12V and 0.6 watt with two wires.
   What type of "inline resistor" should I use? I am
   under the impression that inline resistor is rated
   by "ohms". Can someone give me a RadioShack
   product-number or something like that?

-  Which wire should I solder the inline resistor to the
   fan? Red-wire or the blue-wire of the fan?

-  Does this matter which way I orient the inline resistor?

Thanks in advance for any information.

Jay Chan

jaykc***@hotmail.com wrote:
Show quoteHide quoteWell, at least you know you don't know anything about electricity.
If the figures you give are accurate (which they aren't, they are
probably maximums) you want about the same ohmage as what the fan
appears to be.  This would be about 12 v. / 0.05 A = 240 ohms.  All
that really does is give you a ball park number.  You may want
anything from about 100 ohms to that.  They should probably be at
least 2 watt rated for safety, but have virtually no tolerance
requirements.  Get a set of 2 watt carbon resistors, worth about 5
to 10 cents each (but expect to pay more unless you can raid
someones junk box), say 100, 150, 220, 330 ohms and try them out.
No, it doesn't matter which lead they appear in, nor which
direction they are connected in.  Ohms and watts are an adequate
specification, and you don't need anything better than 10 or 20%
tolerance.

To try them out you don't need to solder anything.  Clip leads with
alligator clips on the ends will be useful.  Feel the resistor
while running with it - it should not be too hot to touch.  If so,
it is a sign you need a higher wattage.  But I expect 2 watts will
be more than adequate.

<snip>

Show quoteHide quoteJust wondering... When trying to slow a fan in a PC, which is better - using
5v and 12v to get 7v to the fan, or adding a resistor inline?

I've got a Media PC that's very quiet except for one case fan.

Thx

Troll

Look who's talking

On Wed, 16 Mar 2005 16:15:30 GMT, "Noozer"
<dont.spam@me.here> wrote:


The resistor, assuming a reliable connection is made
(soldered or good crimp, not just twisted bare wires) and
electrically isolated (like heatshrink or similar, as
electrical tape can degrade and unravel especially in hotter
environments). 

In practice, either has worked fine for most people trying
(either).  The resistor is certainly a higher level of
control, being able to choose some other voltage... 7V is
too slow for some applications and 5V too low for some fans
to even reliably spin-up.  With today's larger 92-120mm fans
becoming more common, in some cases running the fan from 5V
(5V rail & ground)  could be a good alternative instead.

Only issue there is being aware of the resultant temp
changes that come from flow reduction.

Show quote Hide quote Thanks. I felt that the resistor was better, but I wasn't sure if there was
anything about it that I had overlooked.

Machine runs cool. I'd just disconnect the fan completely, but I think that
would be a bit drastic as the chassis isn't the best design.

Noozer wrote:
Show quoteHide quoteYou can only use the 12-5=7 trick when the 5 V line has enough load
on it to sink the 7 V current.  That will usually be so.  However,
that also gives you only one possible operation point, while the
resistor allows you to select what you want, at the cost of a watt
or two.

Please try to get in the habit of preserving attributions for
material you quote.

Thanks for specifying the range of inline resistors that I can try.
This is especially useful if I need to mail order them.  Then, I can
order a range of inline resistors instead of just one type. Otherwise,
the shipping cost will kill me if I need to order various types
multiple times.

Anyway, I will look for them in the local RadioShack first.  If not, I
may have to mail order them.

Jay Chan

On 15 Mar 2005 20:18:20 -0800, jaykc***@hotmail.com wrote:

Show quoteHide quotethere should be also a +5 V somewhere inside the case available!
Find it & tray that instead of 12V ...
Show quoteHide quote

I only see one single two-pins connection on the circuit board of that
small gigabit switch.  This means this option is not available.

Moreover, I believe 5-volt may be too low, and the fan "may" have a
problem to start spinning according to what I read in other newsgroup
about undervolting a fan to keep a PC quiet.

Jay Chan

jaykc***@hotmail.com wrote:
Without quotes and attributions your messages make no sense.  See
my sig below for a means of sane use of google.

How does it look now?  Does it show up the right way?

Jay Chan

jaykc***@hotmail.com wrote:
*>> "If you want to post a followup via groups.google.com, don't
use
*>>  the broken "Reply" link at the bottom of the article.  Click
on
*>>  "show options" at the top of the article, then click on the
*>>  "Reply" at the bottom of the article headers." - Keith
Thompson
*> Jay Chan

That looks fine.  Very good.  Now you just need to learn to snip
the useless portions that are not germane to the reply.  In this
case it would be all the lines preceded by * above.

OK, I will reply through the Show-Options button from now on.

I am still wondering what happened to my previous reply that doesn't
show up.

Jay Chan

jaykc***@hotmail.com wrote:
Usenet is a fundamentally unreliable system.  An article is sent to
some list of entities, which in turn send it on to other lists they
know about.  A glitch in transmisssion, or a mistake in a list,
means the article stops promulgating.  Google is just one of many
entities receiving usenet articles, in something like 50,000 or
more newsgroups.

Once your article gets to one of those many servers, some user can
come along and query the server about what it has available.  These
are usually classified by author, date, subject, newsgroup, and
thread.  There are standards defining these things generally known
as RFCs.  Your article may never get read, or it may only be read
at a small fraction of the servers receiving it, or it may be
exceedingly popular.  Thus it is important to trim it down to the
essentials, and to follow the established conventions.

On 15 Mar 2005 20:18:20 -0800, jaykc***@hotmail.com wrote:

Take original fan and cut it's power + lead in half, maybe
in the middle so you have enough slack to work with.

Next look at the length of the resistor  you'd use,
accounting for a few (maybe 3) millimeters on each of the
leads, and cut off that much more from one of the ends of
the wire you'd just cut.  Essentially you're shortening the
lead such that both leads will be same length after the
resistor is soldered on, which is theoretically unnecessary
but looks better.

Get a 2W resistor, the value depends on the fan but a
ballpark range would be 47-200 Ohm.  A value of about 100
Ohm is a good first guess with no further info.  Since you
don't yet know what the resultant RPM & noise will be at
"6.0V", it could be that you don't actually want exactly 6V?

Another alternative could be to start out with a 200 Ohm (or
make-do with whatever is available) rheostat, dialing in the
resistance value you find to be optimal then measuring the
resistance, to "size" the resistor value you need.  Then
choose the closest commonly available value (or whatever you
have on hand).

So you've found the resistor value you want, simply slip a
length of heat-shrink tubing on each of the cut wire ends,
back as far as possible away from the heat of soldering.
Then solder on the wires-  could help to tin the wires first
and/or clamp them to the resistor while soldering.  In theor
a mechanical joint is best before soldering, but a good
solder joint alone with heatshrink over it is sufficient for
a non-stressed connection (no later mechanical force will be
applied to it).

After resistor is soldered on, slip the heatshrink over the
soldered, bare metal area and heat it up to shrink it.


Show quoteHide quoteInline simply means it's a serial connection, that you cut
the power lead and put each cut end on opposite ends of the
resistor.  It would be called a "2W" resistor.  There are
different resistor compositions like metal or carbon and
ceramic, etc.  They don't matter, just use smallest 2W
resistor you can find, OR the cheapest, or whatever's
available where you normally purchase parts (to combine with
other orders since resistor is tiny fraction of the cost to
ship it) or whatever you have handy. 

Yes, so long as you don't get a giant cermet (cement filled)
wire-wound type.  They are usually rectangular and white
colored, and quite overkill.  The typical resistor you might
seek is about the diameter of a thin pencil and just under 2
cm long.

It's a relatively slow fan already, again I'd try around 100
Ohm resistor first, maybe slightly higher.


Well Radio Shack might have them, I'll check that in a
moment but these are common parts, any fully stocked
electronics outfit should have something suitable.  Seems
like Radio Shack brick-n-mortar stores carry fewer misc.
parts than they used to, you might not be able to assume
what's available online is actually stocked in any
particular store- I'd call ahead first to ask as one store
might have it but another not.

I didn't see any on Radio Shack's website, maybe I
overlooked them but I dug up another source, might as well
get a good deal on 'em?

http://www.goldmine-elec-products.com/prodinfo.asp?number=G14511&variation=&aitem=2&mitem=11

You might call and ask them if they'd just throw 5 (pieces
instead of 200) of them in an envelope, if you're trying to
get ultra-cheap shipping.

Since the fan has such a low RPM already, it's possible that
it's not a good candidate for (around) 6V operation, but it
might be suitably quiet at higher voltage.  You might try a
68 Ohm resistor, or solder two 120 Ohm resistors in parallel
(for 60 Ohms), which is just connecting both ends of each
resistor together and one end of the cut wire (mentioned
previously) to each junction.


Positive, usually red-wire is the typical way, but if you
wanted to solder it to the blue wire instead that would work
too with no drawbacks since this is only a two (power) lead
type fan connection.

no

Thanks for the detailed reply that you have spent the time to prepare.
I appreciate your help.

I especially appreciate the fact that you have gone through the trouble
to find a picture of the 120 ohm resistor.  Now, I have an idea of what
I should be looking for.

Also thanks for telling me the details on the way to attach it to the
wiring.

I will buy a range of inline resistors plus the 68 ohm resistor that
you have suggested, and several alligator clips.

Jay Chan

I use a Zalman Noiseless Resistor Cable  from
http://www.quietpc.com/uk/casefansacc.php#rc56 and it has made my PC quiet.

Mike

Glad to hear that you have good result with using that Zalman resistor
cable.

I came across the exact same item last night when I was looking around
for a solution.  But I was not exactly sure how I could use this
3-wires item into my 2-wires configuration.  Moreover, I didn't know
whether the resistor was the right item for my 60mm fan.  Therefore, I
figured I should ask first before spending the money.  Now that I have
received some helpful info here.  I have a feeling that the Zalman
resistor cable could work.  I just needed to cut out the plugs and only
used the resistor.  But if I needed to cut out the plugs, I might as
well order the resistor alone from elsewhere.  Having said that I
should add that that Zalman resistor cable should be useful for people
who need to reduce the noise of a case fan without soldering.

By the way, what's the ohm rating of that Zalman resistor cable, how
many watt, and what size fan do you use the cable for?  I am just
curious.

Jay Chan

On 16 Mar 2005 09:53:19 -0800, jaykc***@hotmail.com wrote:

The optimal type would have that semi-barbell shape, but
usually are not cut (leads) and formed (bent leads) yet,
which is no problem either way, the leads are quite easy to
cut and bend.  Often the resistors will not be solid green
like the one I linked but colored based on type and striped
to indicate the value.  For example the Zalman resistor Mike
linked wasn't a great picture but appears to be striped:
Green, Blue, Black, Gold, which is a 56 Ohm 5% tolerance
resistor.  Don't worry about tolerance, this is not a
precision application requiring any particular tolernace
part.

Generally speaking, the faster the fan is at default (12V),
the lower the value of the resistor to use to reduce it's
speed.  This initially seems contrary to reason but is due
to the lower resistance of the fan's windings & circuit.  I
expect 68 ohms to not be a large enough value but I do not
have any of that particular model of fan to experiment with.


The heatshrink tubing is very handy too, maybe 3/32 (+-
1/32)" is the right size tubing depending on the fan's
wiring gauge.

If I understand this correctly, you are say these:

-  We could have gone down to 56 ohm for maximum reduction of fan speed
and (hopefully) noise.  But we may run the risk of the fan fails to
start spinning.

-  In order to reduce the chance of fan failure to start spinning while
still cutting down noise, we may want to choose a 68 ohm resistor
instread of a 56 ohm version.

This makes sense to me.  Thanks.

Yes, I have already had a pack of heatshrink tubing left over from my
previous project of re-wiring some phone-cables in my basement.
Therefore, I don't need to buy any more of this.  Your reminder will be
helpful for people who may jump into the middle of this thread without
seeing prior discussion of using heatshrink tubing though.

Have a happy St. Patrick Day!

Jay Chan

<jaykc***@hotmail.com> wrote in message
Show quoteHide quoteYou've got it backwards. Lower value resistor means faster spinning fan.
Having no resistor is the same a 0 ohms.
Adding a 68 ohm resistor will slow the fan more than the 56 ohms.

On 17 Mar 2005 06:35:54 -0800, jaykc***@hotmail.com wrote:

Show quoteHide quoteNo, it varies based on what fan is used.  Per the same fan,
the higher resistance would reduce speed more and have
higher chance of failure to spin up.  Per a different fan
that starts out (with no resistor) having higher RPM,
current, a lower valued resistor would be used than for the
(unmodified) slower fan.


No, the chance of fan failure (to spin up) would be higher
with 68 Ohm, BUT as I mentioned previously, the value
(range) applicable to a particular fan is different than for
a (different) fan.  It cannot be assumed that 68 or 56 Ohm
will have same result with one fan as another.  For your fan
the threshold might be much higher, like around 160 Ohm.  I
cannot predict this as I don't have your fan.  Similar (but
only slightly faster) fans I've had can use over 120 Ohm,
but there was not much point beyond 120 Ohm as the fan was
already inaudible.  

I presume Zalman chose the 56 Ohm value because the most
typical, common fan has amperage spec around .2A, rather
than your .05A (IIRC you wrote .6W ?).  The majority of
people seeking fan noise reduction started out with a much
louder fan that "usually, typically" needs a lower resistor
value than you should, but again I don't have that specific
model of fan so I can only speculate based on experience
with other fans.

Thanks for the correction that you and Noozer have pointed out. I got
it backward.  The higher ohm rating should mean higher resistance and
the fan will run slower and supposedly quieter.

Seem like this is one more reason to get a range of resistors instead
of just one.  Talking about getting a range of resistors. I cannot find
any suitable resistors from RadioShack. Their selection in term of watt
is very limited (1/2, 1, 5, 10 watts). I will have to order online and
pay shipping. Oh well...

Jay Chan

jaykc***@hotmail.com wrote:

Show quoteHide quoteIf you're going to order, might as well get a 100 ohm potentiometer, and
adjust for the noise/cooling of your liking. Let it go full speed when
you are out of the room, and much slower when you are around it.

Too late for that, I have already ordered an assorted number of
resistors from eBay. And I don't understand potentionmeter enough to
know what to do with it. Thanks anyway. I hope someone else who reads
this message may find this useful in keeping the noise of his PC down.

I was quite surprised that I could only find one place (eBay) that has
more than one type of 2-watt resistors that are in the range that I am
looking for.  Many places don't sell 2-watt any resistor.  Those place
that sell 2-watt resistors only sell 2-watt resistors in one ohm rating
(but I need resistors in a range of ohm ratings).

Jay Chan

On 18 Mar 2005 05:39:46 -0800, jaykc***@hotmail.com wrote:

Show quoteHide quoteGenerally the parts would be found at electronics supply
houses, online at places like Digikey, Mouser, Newark,
Allied Electronics, to name a few.  The link I previously
gave for the 200 pieces was just one of myriad hit-or-miss
surplus type companies that don't always have what you need,
but when they do it's often cheaper.  Main problem with many
suppliers is minimum order requirements like $25.  Digikey
comes to mind as one that doesn't have the minimum order
requirement.

Previously a 100 Ohm POT had been mentioned.  POTs generally
do not have sufficient current capability for a fan, or
rather, when they do they're called rheostats rather than
POTs.  A POT would be better for fine-tuning something like
a LM317 based fan control, but then probably not 100 Ohm
range.  More elaborate fan controllers have their place with
those desiring more frequent changing of speed but higher
cost, time to construct, size, etc, make them less viable
for this kind of simple mod.

On Fri, 18 Mar 2005 15:48:57 GMT, kony <spam@spam.com>
wrote:

<snip>

Correction-  Digikey does have a <$25 small order fee ($5),
I was thinking of Mouser, https://www.mouser.com.

jaykc***@hotmail.com wrote:

Why 2-watt? Even at a 5 volt drop, current would have to be 0.4a to
equal 2 watts. At 7v (a 5 volt drop), most 12v fans draw in the 0.05-0.1a
range. I've used 1w carbon resistors (much smaller) and they hardly get
warm.

On Fri, 18 Mar 2005 21:49:53 -0800, ric <nospam@home.com>
wrote:

True, either 1 or 2W could work for this specific
application, but the 2W seems potentially less problematic.
At the time I didn't realize OP would have problems finding
them or go to ebay for these.

Why not 2W?  Typically one does not try to match up watt per
watt on power resistors, rather using what the space and
budget allow, overengineering them.  It's not a PCB space
issue and cost is maybe a few cents, or maybe even cheaper
per piece since this is spot-pricing on items not often
(percentagewise) sold 1 at a time.  Yes, 1W would usually
work, but IMO, 2W is as good or better choice if someone has
nothing (yet).

Also, "barely" warm is a relative term.  They can and do get
warm, even 2W,  and some components in a PC/router/etc
shouldn't be placed in contact with a "warm" power resistor.
Granted in this one instance it shouldn't get as warm as
others.   The installer could isolate it, but then has to
consider that method and typical nylon wire-ties aren't so
great if the resistor did get more than warm.  The router
itself will already be warm inside, hence the need for the
fan.  Metal, well, sure if challenged we could all think up
some way to do it but ultimately choosing a larger resistor
also results in lower thermal density, it's not going to be
as hot against (whatever it's touching if it were touching
anything).

kony wrote:

Only because I found a good assortment of 1W carbon resistors at
my local Radio Shack. 2W? No such luck. 1W is plenty.

On Sat, 19 Mar 2005 12:12:07 -0800, ric <nospam@home.com>
wrote:

Show quoteHide quoteDo they have one of those packs with several assorted values
inside or is one forced to buy the individual smaller packs?
I suppose it could depend on the store too, those I've been
to in the past couple years seem to have scaled back their
stock of these types of components.   The individual packs
from Radio Shack always seem wildly overpriced to me, though
it could be a good idea for only a few values.

kony wrote:

Man, this was years ago. Before I started using potentiometers. As I
remember, each package had about 5 resistors of the same value.
That store isn't even there any more. Progress!

http://www.goldmine-elec-products.com/prodinfo.asp?number=G14511&variation=&aitem=2&mitem=11

Yesterday, I browsed around to look for 2-watt resistors, and I was
surpised that I could only find one vendor (in eBay) that offered
resistors in a range of ohm ratings.  Most places don't sell any 2-watt
resistors.  Now, I _really_ appreciate the fact that you had spent the
time in digging up a link for 2-watt resistors.  That had to be quite
hard.

Jay Chan

On 18 Mar 2005 05:43:57 -0800, jaykc***@hotmail.com wrote:

Show quoteHide quoteWell I've bought stuff from them before... plus I've heard
of others calling, asking them to send very small items in
an envelope for cheapest shipping/total-cost.  In another
post I listed a few more full-featured electronics companies
with online stores... generally such parts would come from
electronics-oriented companies rather than computer focused.

I finally use the following power resistor from eBay to reduce the
speed to the point that the noisy gigabit switch is now very quiet:

   100 ohm 2W 5% carbon composite

I have a feeling that I still have more air flow from the new 60mm fan
with reduced speed than what the old noisy 40mm fan can produce in full
speed. This means I could have reduced the speed further. Obviusly, I
am not going to mess with something that is working so well (plus the
fact that this is so difficult to find that type of power resistors).
But just for the purpose of curiosity, which rating of power resistor
would you recommend if someone in my situation wanted to further reduce
the fan speed slightly.

Thanks for your help and any other info that you may provide in the
future.

Jay Chan

On 4 Apr 2005 08:20:11 -0700, jaykc***@hotmail.com wrote:

Show quoteHide quoteI dont' know how much further that particular fan will
tolerate a voltage drop so I can't help much.  Past a
certain point a larger resistor will prevent it from
spinning up which obviously should be avoided.   I doubt you
could use 220 Ohm but slightly higher than 100 Ohm might be
possible.  If would be easiest to use a rheostat (variable
power resistor) to determine the maximum resistance value
then keep the (more expensive and large) rheostat for some
future use (like more fan resistor sizing?) and put in the
appropriate value of resistor in it's place.

kony wrote:

Placing a cap in parallel with the resistor seems to cure startup
problems.

On Tue, 05 Apr 2005 23:04:54 -0700, ric <nospam@home.com>
wrote:

"Cure"?
To what extent?
I mean, does it actually allow a lower working RPM then or
just offset the overly high resistance?

This is an interesting idea, do you have an example of what
gain there will be for any particular fan?

kony wrote:
Show quoteHide quoteThis is probably just plain dangerous to the rest of your system.
It would depend on "giving a kick" at turn-on, provided that the
power supply had a rapid rise time.  If the fan ever gets stalled
in normal operation the kick is not there, and the thing just sits
and encourages the rest of the system to cook.

On Wed, 06 Apr 2005 14:31:52 GMT, CBFalconer
<cbfalco***@yahoo.com> wrote:

Show quoteHide quoteWhat I'm suspecting is that if a cap were in parallel, a
bypass for the power to the fan, it would just make matters
worse because whatever the turn-on delay was already, adding
more capacitance will only lengthen that delay.

Ric wrote about it being parallel with the resistor though,
so i presume actually paralleled to it, bypassing it on the
same power lead.  While that can be a crude but effective
filter for some noise reduction, I'd not heard of anyone
employing anything like this for fans.

Sometimes it's not even, merely a matter of getting the fan
to spin-up though, with some fans there will be pulsation
below a certain RPM, actually making the fan louder than if
it were running slightly faster-enough to eliminate this
pulsation.  Offhand I'd speculate that it pulsating occurs
due to stronger (permanent) magnetic field in some fans than
others.  IE- It seems more common in >= 120 x 32mm, large
thick fans.

kony wrote:
Show quoteHide quoteNo suspicion needed.  It forms a simple high pass filter.  Very
straight forward.  Once the cap is charged it has no further
effect.  The charging current, as the power comes on, supplies the
initial start-up boost.

kony wrote:

For fans that will spin at, for example, 5 volts but have startup
problems, the cap acts as a short at startup, allowing the full
12v to reach the fan, then charging to the resistor's dropping voltage.
I normally use a 10uF, 15v electrolytic. [Get the polarity right!]

Not off hand. No gain in operational function. But it sure cures the
"I have to help my fan start" blues.

Seem like what you have suggested previously probably is a better
choice than what I end up buying.  Yours is 120 ohm, and what I get is
100 ohm.  Your probably can reduce the fan speed slightly better than
what I get can do.  Moreover, yours is _much_ cheaper than what I have
paid for ($20). Oh well...

Jay Chan

Motherboard died need replacement help.
Building a File Server
Hard drive or motherboard? Strange boot problems...
Is my monitor dying?
Unusual hard drive noises
PC 3200 Memory Question
new system....HELP!!!!
Problem trying to find RAM for motherboard
Firewire Question
Motherboard & Chipset