Do I need a rear fan?

[Mayayana]

| Mayayana wrote:
| > | Just a thought, but if I have to buy a fan, I'd rather just buy a 120v
| > | AC fan. wire it right to the power cord, and run it direct off the
house
| > | electricity. Why suck power from the power supply, which is needed to
| > | run the computer itself, rather than waste it to run a fan. Not to
| > | mention that the more power needed from the power supply, means the
| > | power supply will run that much hotter. (which defeats the purpose).
| > |
| > | And most OEM computer power supplies are rated too low in wattage
| > | anyhow.
| >
| > I did that once, with an eMachines. I think the power
| > supply was something like 240 watts and I had two
| > hard disks in it. So I got a nice fan at Radio Shack, wired
| > it up, and just plugged it in when the computer was on.
|
| I hope you meant " when the computer was *off* " there
| in your last sentence.
|

No. I was talking about an external fan to avoid draw
on the limited power supply. I bought an AC fan and wired
it up, then plugged into a standard 110AC outlet. It was
very efficient and reasonably quiet. I later gave it to my
chiropractor, who had set up a computer inside a cabinet
and was worried about lack of ventilation. The fan was
no bigger than a small PC fan, but ran on normal AC power.

 

[Paul]

| I used to have another article with the states explained.
| This one will have to do as a substitute.
|
| http://arstechnica.com/civis/viewtopic.php?t=354552
|

This is more complex than I realized. Though S1 and S3
don't seem to be very different. If I call up the
shutdown window I have a choice between standy,
shut down and restart (assuming I haven't enabled
hibernate). That standby is what I'm referring to as
standby. (I've never seen anything else called standby.
The option to turn off monitor or disks after a period of
no use is referred to as a "power scheme" in the Power
Options applet. Maybe that's S1?)
In my experience standby has always been the same
thing on any machine: The system shuts down, the
power light flashes, the RAM content is maintained. BUT,
fans plugged directly into a power supply socket keep
running. I wouldn't think something like conserving
CPU cache would affect that.

| Users typically end up with the S1 flavor of standby, after
| the BIOS is misconfigured (sometimes after a CMOS reset), the OS
| is told S3 is not supported, and it reverts to S1. Using dumppo,
| you can do an adminstrative override, to return S3 capability to the
| system. You start by verifying the BIOS has S3 enabled. Then boot
| the OS and use dumppo.exe from ftp.microsoft.com .
|
|
ftp://ftp.microsoft.com/products/Oemtest/v1.1/WOSTest/Tools/Acpi/dumppo.exe
|
| http://forums.pcper.com/showthread.php?p=1825058&postcount=31
|

This is an interesting topic I was completely
unaware of. The MS link doesn't seem to work,
but I found this recommended in a sysinternals forum:

http://www.bellamyjc.org/fr/windows2000.html#veille

There's a download there. I ran it and got the following:

c:\windows\desktop\dumppo.exe admin cap
Min sleep state......: S1
Max sleep state......: S4 - hibernate
Min video timeout....: 0
Max video timeout....: -1
Min spindown timeout.: 0
Max spindown timeout.: -1
power capabilities
System power capabilities
Power Button Present....: TRUE
Sleep Button Present....: FALSE
Lid Present.............: FALSE
System states supported.: S3 S4 S5

That seems to be ambiguous. But it reset to min. level
3 without problem. I don't see any settings in my 3/2013
BIOS. I don't even see an "Enable ACPI" option.

I'm curious how dumppo works. It must be just applying
some sort of API call and/or Registry setting.

I don't have my older PC running -- the one with the
fan that runs all the time. If I get it running I'll have to
try dumppo to see if it stops the "hotwired" fan running
during standy. (I haven't put a rear fan into my latest
machine. The reason for that is partly the running-during-
standy problem, but also because CPUs are much improved
these days. With the ability to shut off for micro intervals
while running, I find that even in a heat wave, the AMD CPUs
I buy generally only run a bit higher than room temperature.)




| If the results looked like this:
|
| dumppo.exe admin cap
|
| Min sleep state......: S3
| Max sleep state......: S4
| System states supported.: S1 S3 S4 S5
|
| it would likely be impossible to go into S1. That's because
| the min sleep state is S3. And S1 can't be accessed by a
| request to sleep.
|
| Paul

The ftp.microsoft.com link works.

It's possible this is the server that uses PASV.
In command prompt it might be

ftp ftp.microsoft.com
user (anonymous)
pass ([email protected])
PASV
CD /products/Oemtest/v1.1/WOSTest/Tools/Acpi
binary
get dumppo.exe

You can see my browser doing the necessary PASV
on its own here. The "TYPE I" is the equivalent
of the "binary" in my suggested FTP session.

http://i60.tinypic.com/51bna.gif

"If at first you don't succeed, try a PASV"

It changes the ports used for an FTP transfer,
and is otherwise known as "passive".

HTH,
Paul

 

[Paul]

| Mayayana wrote:
| > | Just a thought, but if I have to buy a fan, I'd rather just buy a 120v
| > | AC fan. wire it right to the power cord, and run it direct off the
house
| > | electricity. Why suck power from the power supply, which is needed to
| > | run the computer itself, rather than waste it to run a fan. Not to
| > | mention that the more power needed from the power supply, means the
| > | power supply will run that much hotter. (which defeats the purpose).
| > |
| > | And most OEM computer power supplies are rated too low in wattage
| > | anyhow.
| >
| > I did that once, with an eMachines. I think the power
| > supply was something like 240 watts and I had two
| > hard disks in it. So I got a nice fan at Radio Shack, wired
| > it up, and just plugged it in when the computer was on.
|
| I hope you meant " when the computer was *off* " there
| in your last sentence.
|

No. I was talking about an external fan to avoid draw
on the limited power supply. I bought an AC fan and wired
it up, then plugged into a standard 110AC outlet. It was
very efficient and reasonably quiet. I later gave it to my
chiropractor, who had set up a computer inside a cabinet
and was worried about lack of ventilation. The fan was
no bigger than a small PC fan, but ran on normal AC power.

Yeah, I tried that just the one time. Problem was,
my AC fan turned out to be quite noisy. It's a fairly
small fan (80mm) and impedance protected (jam a pencil in
it and it doesn't catch fire). I won't be buying a second one.

Paul

 

[casey.o]

Yeah, I tried that just the one time. Problem was,
my AC fan turned out to be quite noisy. It's a fairly
small fan (80mm) and impedance protected (jam a pencil in
it and it doesn't catch fire). I won't be buying a second one.

Paul

I was thinking that it wouldn't take much to use one of those 12V "wall
worts" to power a fan. Or someone handy with electronics could even
build a separate power supply just for the fan(s). The one thing
lacking is that the old computers had an AC outlet on the rear for
monitors, but they no longer do that. But if you're lime me, I keep a
power strip for all my computer stuff, and just turn that on and off.
Otherwise I have too much stuff to shut off (computer, monitor, ext.
modem, Speakers, and other stuff.) Either way, it seems rather stupid
to use the computer power supply to run fans. Fans dont need "clean
power" like the computer needs. Meaning it dont need to be a clean DC
with lots of filter capacitors. Any 12v transformer and a rectifier
will do the job. Even a 12V automotive battery charger will work, but
that would look pretty stupid on a computer.....

 

[Paul]

I was thinking that it wouldn't take much to use one of those 12V "wall
worts" to power a fan. Or someone handy with electronics could even
build a separate power supply just for the fan(s). The one thing
lacking is that the old computers had an AC outlet on the rear for
monitors, but they no longer do that. But if you're lime me, I keep a
power strip for all my computer stuff, and just turn that on and off.
Otherwise I have too much stuff to shut off (computer, monitor, ext.
modem, Speakers, and other stuff.) Either way, it seems rather stupid
to use the computer power supply to run fans. Fans dont need "clean
power" like the computer needs. Meaning it dont need to be a clean DC
with lots of filter capacitors. Any 12v transformer and a rectifier
will do the job. Even a 12V automotive battery charger will work, but
that would look pretty stupid on a computer.....

I'm already doing that. My networking equipment is
cooled by a brushless DC fan, powered by a 12V wall wart.

The reason for feeding clean power to a DC fan, is
to prevent audible side effects (hum). If the fan
is super powerful, maybe the noise from the fan, drowns
out stuff like that. I wouldn't waste my time doing
a half-ass power supply for one. If I'm going to spend
the time doing power conversion, I make the item reusable
for all sorts of projects. Just regulate it and do it
properly, and your investment can be recycled later.

Paul

 

[J. P. Gilliver (John)]

In message <[email protected]>, Mayayana

So are you saying that some PCs with fans wired
straight to the power supply will shut off in standby?
I've never heard of different kinds of power supplies
in that regard.... or in any other, actually, other
than wattage.

Most power supplies have a standby section that runs all the time, even
when the main parts - including the 12V that feeds the drive motors (and
I assume that's where you're thinking of when you say "straight to the
power supply") - are off.

 

[J. P. Gilliver (John)]

In message <[email protected]>, Paul <[email protected]>
writes:
[]

The ATX PSU has two switching converters inside. The +5VSB has its
own switcher. That switcher is expected to make +5VSB, while
there is no fan running inside the PSU to keep it cool.

[]
It probably is these days, but I think I've seen cases where the 5V
standby isn't a switcher, but has an old-fashioned (50 or 60 Hz) mains
transformer, diodes, series regulator type circuit - efficiency not
being so important for such a low-power part of the thing.

 

[J. P. Gilliver (John)]

In message <[email protected]>,

Just a thought, but if I have to buy a fan, I'd rather just buy a 120v
AC fan. wire it right to the power cord, and run it direct off the house
electricity. Why suck power from the power supply, which is needed to
run the computer itself, rather than waste it to run a fan. Not to
mention that the more power needed from the power supply, means the
power supply will run that much hotter. (which defeats the purpose).

I think the extra cooling provided would exceed the extra heat produced
- modern PSUs are pretty efficient. Plus there's the matter of safety,
and also supply voltage universality (though you're probably not
bothered about that).
[]

find Win98 more stable than XP.

In everyday use, I find XP more stable than I found 98SElite in everyday
use - once I'd rolled back the video driver. Now that I only use my 98
machine for one particular purpose, it is very stable, but that's not a
fair comparison.

 

[J. P. Gilliver (John)]

In message <[email protected]>, Paul <[email protected]>
writes:
[]

The other question you should be asking, is why we cool
the computer case. Primarily, it's to protect the hard drive.

[]
Surely the silicon too - or are you saying that all the important bits
of that (CPU, some graphics cards) have their own fan(s) anyway? (If you
are saying that, then surely hard drive fans - which do exist - rather
than a case fan - would be the way to go?)

 

[J. P. Gilliver (John)]

(e-mail address removed) wrote: []

Seriously, can I just yank an old fan out of a dead power supply and
stick it on the back of the case? I got at least one dead ps laying
around? I hate buying stuff, when I got lots of junk around to salvage

[]
As long as it's a 12V fan, and has a standard connector pinout,
knock yourself out

And the right shape and size: though many PSUs have a big (presumably
sucking) fan on the large size facing inside the PC, some have a small
(though sometimes deep) one on the back cover.

The one I replaced, seemed to meet that description, but read the
label on the fan hub, for hints if it's not standard. (There are
no "user replaceable" parts inside a PSU, so they can use any old
fan they feel like.)

For the same reason (they don't expect the usual user to go inside
there), watch out for sharp edges and the like: I think the only time
I've cut myself working with PCs is when opening a power supply.

Check your motherboard downloadable PDF manual, for the
fan header pinout. Compare to the wire colors on the cable.
Black is GND (ground). Red is usually a positive voltage, like 12V.
An optional third wire is the tacho output, which is not present on
a lot of cheap fans. I think all my case cooling fans, lack
the third wire.

Though again, the colours may not be standard, for the reasons already
expressed.

Just try to make sure the power wires on the fan, get connected
to matching power connections on the fan header. So nothing
gets blown. Motherboards have no fuse in the fan power path!

Ideally, check it on an external power supply - if only to check the
polarity! (And direction! if it sucks/blows the wrong way, you _may_ be
able to reverse it by reversing the polarity, but only if there aren't
actual electronics in the fan hub itself; if there are, then reversing
the polarity may just fry the fan.)

Paul

John

 

[J. P. Gilliver (John)]

wrote in message news:[email protected]...

(You snipped who! It's casey.o, I think.)
[]

If you don't have the case cover on, it really won't make much
difference, if any.

I was wondering when anyone else would notice that he said that! IMO,
you're right, it won't make much if any difference if the cover's off -
possibly even just add heat, due to the (admittedly small amount of)
power it takes. (Whether from the power supply or the mains.)

Even when the case cover is off, you may need a small room fan blowing
lightly across the innards.

Not a bad idea. Probably one of those utilities that have been mentioned
(SpeedFan, for example) is worth playing with, to see if anything's
warmer than it should be. (There are usually sensors in the CPU, at
various points on the mobo, and in some cases in the hard drives.)

A proper case with covers on directs the air flow better.
If you use a free temp monitoring program, such as CPUIDHWMonitor , you
will be able to see several temps, etc.

Oops - I should read on (-:

 

[Paul]

The ATX PSU has two switching converters inside. The +5VSB has its
own switcher. That switcher is expected to make +5VSB, while
there is no fan running inside the PSU to keep it cool.

[]
It probably is these days, but I think I've seen cases where the 5V
standby isn't a switcher, but has an old-fashioned (50 or 60 Hz) mains
transformer, diodes, series regulator type circuit - efficiency not
being so important for such a low-power part of the thing.[/QUOTE]

It's true, that historically +5VSB has been a pig and
is not very efficient. (Less than 50% in some cases.)

Would there be room for a separate mains transformer for +5VSB ?

I presume if such was present, a user might hear a gentle "hum"
coming from the sleeping computer. I have computer speakers,
where when everything is off here, I can hear a hum coming from
the transformer inside the speaker with the amp inside it.

Paul

 

[Paul]

Ideally, check it on an external power supply - if only to check the
polarity! (And direction! if it sucks/blows the wrong way, you _may_ be
able to reverse it by reversing the polarity, but only if there aren't
actual electronics in the fan hub itself; if there are, then reversing
the polarity may just fry the fan.)

John

There are electronics in the fan hub. It's polarized.
See the first picture here.

http://tipperlinne.com/fan-tach.htm

The large three pin device, could be a Hall sensor.
It picks up the magnetic field as the thing rotates.
And sends some sort of signal to the two transistors
controlling the coils.

There's an example of Hall sensors here. Apparently
the three pins are power, ground and an open-collector
status. Somehow, that sensor provides control signals
for the two other transistors. Sensing the rotational
position, is so commutation happens at the appropriate
moment (switch on other coil, pull rotor forward).

http://sensing.honeywell.com/honeyw...sors-productsheet-005947-3-en.pdf?name=SS460S

I've also seen a four pin IC for the job, but considering
how cheaply fans are made, that's not likely to be used.
They're probably still using multiple surface
mount components to make a controller for it.

Paul

 

[Paul]

The other question you should be asking, is why we cool
the computer case. Primarily, it's to protect the hard drive.

[]
Surely the silicon too - or are you saying that all the important bits
of that (CPU, some graphics cards) have their own fan(s) anyway? (If you
are saying that, then surely hard drive fans - which do exist - rather
than a case fan - would be the way to go?)[/QUOTE]

Temperature wise, I consider the hard drive to be the first one
affected by high temperatures. The other components will eventually
be affected, but at a higher temperature. The CPU has throttling, as
a means to reduce heat output.

The Vcore circuit around the CPU, some of those have continued
to run properly, up until the point that the solder melted
on the MOSFETs I thought that was pretty funny. The MOSFETs
can actually go into thermal runaway (channel ON resistance increases
with increasing temperature, which makes the part run hotter).

The chipset chips might be rated for 99C in the datasheet.
Hotter than the 50 or 60C max for a hard drive.

The guy at our fab (head of the cell library development), said
our ICs would be damaged at around 135C long term (parametric shift).
And our designers simulated to 105C or 110C for operational details
(circuit still operates at 105C digitally), while the finished parts
had a 70C commercial Tcase rating.

Parts designed for higher temperature operation than that
(automotive industry), could be using tweaked materials
as a starting point. Perhaps ceramic packaging and
silicon carbide substrates. Never worked in that industry,
so don't know how they do it.

The reason I'm mentioning the hard drive, is I've seen
a few computer cases, that leave the hard drive in
a thermal shadow (no air circulation at all), so who
knows how hot those are actually getting. You can use
a SMART utility, to check the hard drive temperature, just
to be sure about it.

Modern hard drives, the operation of the data writing,
is temperature sensitive. And the drive likely has
temperature compensation for that (head current flow
adjustment versus temperature). Some drive now,
on a write, the head is actually forced to get
nearer to the platter, and then backs off during
idle or during reads. And that's actually done
with a thermal element of some sort (something
swells when it gets hot). An electric current,
sent down the actuator arm, tells the head to get
closer or move further from the platter. So a different
flying height is desired on write versus read.

Paul

 

[Paul]

wrote in message news:[email protected]...

(You snipped who! It's casey.o, I think.)
[]

If you don't have the case cover on, it really won't make much
difference, if any.

I was wondering when anyone else would notice that he said that! IMO,
you're right, it won't make much if any difference if the cover's off -
possibly even just add heat, due to the (admittedly small amount of)
power it takes. (Whether from the power supply or the mains.)

Even when the case cover is off, you may need a small room fan blowing
lightly across the innards.

Not a bad idea. Probably one of those utilities that have been mentioned
(SpeedFan, for example) is worth playing with, to see if anything's
warmer than it should be. (There are usually sensors in the CPU, at
various points on the mobo, and in some cases in the hard drives.)

A proper case with covers on directs the air flow better.
If you use a free temp monitoring program, such as CPUIDHWMonitor ,
you will be able to see several temps, etc.

Oops - I should read on (-:

Closing the side, and using case fans for forced air cooling,
can do a better job than just leaving the side off. It depends
on the CFM you use, as to how much better. I don't have any
numbers handy, to suggest just how much convective effect you
get with the side off.

And you're perfectly free to do these experiments for yourself.
The motherboard has the sensors. Even the GPU temperature can be
sensed with GPU-Z program from Techspot.

Paul

 

[J. P. Gilliver (John)]

J. P. Gilliver (John) wrote: []

It probably is these days, but I think I've seen cases where the 5V
standby isn't a switcher, but has an old-fashioned (50 or 60 Hz) mains
transformer, diodes, series regulator type circuit - efficiency not
being so important for such a low-power part of the thing.

It's true, that historically +5VSB has been a pig and
is not very efficient. (Less than 50% in some cases.)

Would there be room for a separate mains transformer for +5VSB ?

I think so. It's probably a dying art now, but I've seen very small
mains transformers back in the day - certainly under an inch cube, I
think even a half inch.

I presume if such was present, a user might hear a gentle "hum"
coming from the sleeping computer. I have computer speakers,

That's rather an endearing picture!

where when everything is off here, I can hear a hum coming from
the transformer inside the speaker with the amp inside it.

Paul

Actually, in theory, a transformer can be capable of not humming
(especially if drawing hardly any power), if it's securely fixed
together and clamped and shielded and so on, though you're right that
they usually do.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Bread is lovely, don't get me wrong. But it's not cake. Or it's rubbish cake.
I always thought that bread needed more sugar and some icing. - Sarah Millican
(Radio Times 11-17 May 2013)

 

[Bob F]

Ideally, check it on an external power supply - if only to check the
polarity! (And direction! if it sucks/blows the wrong way, you _may_
be able to reverse it by reversing the polarity, but only if there
aren't actual electronics in the fan hub itself; if there are, then
reversing the polarity may just fry the fan.)

If you look at the blades of most fans, they are curved to "scoop" up the air
and accelerate it through the fan. Even if they did work backwards, they would
not work very well.

 

[Bob F]

Only you can determine the number of fans and placement.
Does anything feel hot when running?

This machine that I am using at the moment has 8 fans
and still runs hot:
pwr supply 6"
cpu 4"
chipset 4"
ram 4"
gpu front side 4"
gpu back side 4"
harddrives 4"
comp side panel 12" a/c line powered fan

That's a lot of fan.

I've found in my experimentation that most cases are horribly designed for
cooling. They usually have all kinds of fans blowing out, and hardly any
reasonable path for air to enter the case. Additionally, neither the GPU or CPU
fans are usually designed to exhaust the hot air from these hottest heat
producers directly out of the case, so it gets re-cycled over and over, and
everything in the case gets hotter.

My latest PC now has a 5" fan in the front 3 5-1/4" drive bays blowing air into
the box through a foam filter. Shrouding around the processor fan directs air
from the front of the box through the fan, then directs the warmed output air
towards the power supply and back case fan. The HD4850 X2 GPU has a shroud
around it that directs the hot air directly out the side case cover through a
1"x12" slot cut in the cover.

Q8200, Asus P5K Premium, Saphire HD4850X2 with origional fans/shroud replaced, 1
SSD and 2 SATA HD.

Current temps as I work. (69 F room) using Speedfan
Core 0 49C
core 1 39C
core 2 43C
core 3 37C
GPU1 49C
GPU2 43C
HD0 24C
HD1 30C
HD2 33C

Back panel fan 1350 RPM
CPU Fan 987 RPM 32%
GPU1&2 fan (not stock) 1250 RPM 38%

The front panel fan is slowed down with a manual control to maybe 50% - silent

At idle, the machine is vertually silent.

Running Furmark for burn-in, GPU 1 gets into the low 70C range.

Running Prime95 at the highest hest setting on all 4 Q8200 cores
Core 0 58C
core 1 49C
core 2 53C
core 3 49C
with the CPU fan still at 987 RPM

Adding Furmark to that takes the CPU and GPU temps to
Core 0 59C
core 1 52C
core 2 56C
core 3 51C
GPU1 72C
GPU2 44C

with the GPU fans going to 100% 2600PM
and no change in the CPU fan speed

My Advise: Get the hot air out of the case immediately. It makes a huge
difference.

With the cover off, a case back fan will make almost no difference.
Additionally, the holes-in-sheet-metal venting most cases have for the back fan
really restricts the air flow. cutting out the sheet metal and replacing it with
a wire grid finger protector will increase that fans effectiveness if the case
is closed.