|
|
Heatsink/ Heat Pipe / ThermoSiphon Cooling The cat will only make the mistake of putting its paw by your HSF once. :) Also the place to discuss the new high end heat pipe goodness. |
Thread Tools |
10-31-2002, 11:34 AM | #1 |
Cooling Savant
Join Date: Nov 2001
Location: Suwanee, GA
Posts: 130
|
Pulse Width Modulation and other fan Q's
I'm h20 cooling my p4 2.5 so I can hit the wonderful 3GHz mark. It's great! My PC is silent and 100% stable. I also run it at full load 24/7 with f@H, and I encode DivX and stuff overnight all the time.
Well.. the 2 Antec 92mm Smart fans started to rumble (first one and then a few weeks later the other) and now rumble after being on for a few min. The rumble is loud and very annoying where as before they were practically inaudible next to your ear. So I pulled out 2 Panaflow 92mm FBA09A12L fans. These fans really push the air at low noise. They push a LOT more air than the Antec's did. Well, when I put them on my heater core, they become very loud (air rushing, not any whine noises). So I tried them at 7V instead of 12V. They become 100% silent, but push less air than the Antec fans. So I was wondering how I could achieve a happy medium of air flow/noise somewhere in the 12 to 7V range. it seems like PWM would be better than using a rheostat based devise, but I don't know this. Can any one share their experiences with both. pH, perhaps? Also, I was looking at PWM controllers. retail product: http://www.natriumtech.com/products.php?pid=1 DIY: http://www.cpemma.co.uk/pwm.html I could do the DIY with a little help (I'm @ a school for engineers (GATech) and my dad owns a business that manufactures surface mount stuff, but they still do some thru-hole stuff too), but I don't know my way around circuit diagrams as well as I should. The DIY link looks like it can also vary the speed of the fan like the retail (but in an analog and not digital control). Is this right? Also, if I used the DIY approach, how would I control 2 fans? would I need 2 of these? Does any one know of any other retail PWM control devices? now onto rheostat based stuff: http://www.vantecusa.com/nxp101.html that looks nice, and I'm pretty sure it's not PWM. any other good rheostat based devices? I've seen the cliff's bay bus guides too, so I can look into DIY for this as well. Thanks and I hope someone can answer this barrage of questions! -Russell
__________________
My only PC is a PII 333 with 256MB RAM a GeForce2 MX 200, and 30GB's of 7,200rpm ATA/66 goodness. And yes, I know it sucks. |
10-31-2002, 11:44 AM | #2 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
PWM wouldn't be my first choice for controlling a fan, because the fan may start behaving erratically due to the frequency. I'm going to assume that the products to which you've provided a link have addressed this issue, which is really a minor one.
A rheobus is far simpler and cheaper. It's a nice medium between the rheostat and the PWM controller. The rheostat is not energy efficient. As for fans, I have yet to pick it up, but I am going with an automotive "squirrel-cage" type blower. The advantages include lower noise per cfm, and more importantly, they are able to create good flow with much more pressure than what an axial fan can do. Check out Delta's lineup, at www.deltaww.com . |
10-31-2002, 04:28 PM | #3 |
Cooling Savant
Join Date: Nov 2001
Location: Suwanee, GA
Posts: 130
|
what is the difference between a rheostat and a rheobus?
I thought a rheobus was just a a drivebay-mounted rheostat (or several to control multiple fans)....
__________________
My only PC is a PII 333 with 256MB RAM a GeForce2 MX 200, and 30GB's of 7,200rpm ATA/66 goodness. And yes, I know it sucks. |
10-31-2002, 04:32 PM | #4 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Ah!
No, a rheobus uses a voltage regulator, that's controlled by a low power potentiometer (a rheostat is a high power pot). |
10-31-2002, 05:03 PM | #5 |
Cooling Savant
Join Date: Nov 2001
Location: Suwanee, GA
Posts: 130
|
ah.. I got ya.
So I guess the vantec thing is a rheobus? I've looked @ the "squirrel-cage" blowers too, but the area they blow across isn't that large, and I noticed that if I have my 92mm fans blowing over ALL of the veins of the heatercoer, I get MASSIVELY better results than just a higher CFM. so if they make a ~6" wide blower that is silent, I'd go for it. I looked @ the delta's and none of them had that great of CFM (great pressure I know) but were all like 40 and up dBA. I want a 20's and lower dBA PC. does anybody know where I can buy that vantec gizmo? I can't ind any shops that have it. (I also can't find the PWM device for sale or a price for it..) Also, it the vantec gizmo and the PWM gizmo are the same price, which would be better?
__________________
My only PC is a PII 333 with 256MB RAM a GeForce2 MX 200, and 30GB's of 7,200rpm ATA/66 goodness. And yes, I know it sucks. |
11-01-2002, 08:55 AM | #6 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Too bad. I didn't have any problems finding a blower that had a pressure drop of more than 1 inch H2O. Axial fans don't push any air past 1/4 inch H2O.
Shrouding is required for all rads, so the output area of the blower is not an issue. I don't know how the Vantec unit operates. I read through the manual, and it just doesn't say. |
11-01-2002, 01:24 PM | #7 |
Pro/Staff
Join Date: Oct 2001
Location: Klamath Falls, OR
Posts: 1,439
|
Somebody dissin PWM fan control? Avast maties and prepare to be boarded! PWM is MUCH better in many ways than rheostats, voltage regulators and such. The ONLY drawback to it is that it doesn't work well with some fans because they stutter or whine when PWM'ed.
Home brewed PWM works just fine on my home system, thank you. If you don't like the stutter or whine and you don't want to switch fan vendor or model, up the PWM frequency and stick a big cap on the output to turn it into a voltage controller. |
11-01-2002, 01:28 PM | #8 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Well if you're going to turn it into a voltage controler, why not just use a voltage regulator?
Otherwise you're right on: upping the freq and adding a big cap will fix it. Incidentally, what's the cost of a PWM controller, versus a rheobus? |
11-01-2002, 04:00 PM | #9 |
Cooling Savant
Join Date: Nov 2001
Location: Suwanee, GA
Posts: 130
|
i noticed the PWM gizmo @ the linked site is like $70 with $30 shipping to the states.
so ya... I'm thinking rheobus... cheaper=happy me.
__________________
My only PC is a PII 333 with 256MB RAM a GeForce2 MX 200, and 30GB's of 7,200rpm ATA/66 goodness. And yes, I know it sucks. |
11-01-2002, 05:44 PM | #10 |
Pro/Staff
Join Date: Oct 2001
Location: Klamath Falls, OR
Posts: 1,439
|
Why not use a linear voltage regulator? Because it dumps waste heat. We are all about not putting unecessary heat in our computers, right?
As for the cost... it might be a bit more expensive to do the PWM. You'd need (to duplicate my version): 3 56K resistors - $0.99 1 5K resistor - $0.99 1 10K pot linear taper - $2.49 1 0.47uF capacitor - $0.99 1 555 timer chip - $1.49 1 741 opamp chip - $0.89 1 NPN transistor - $0.69 1 IRF510 transistor - $1.99 1 diode - $2.49 1 proto board - $3.99 Total: $17 To build a linear voltage regulation system, you would need: 1 proto board - $3.99 1 LM317 - $1.99 1 resistor (I can't remember what value) - $0.99 1 10K pot linear taper - $2.49 1 heatsink - $1.59 Total: $11.05 Prices are from Radio Shack, which is more expensive than most, but very available. With the above prices, you will also be left with a number of leftover components when building the PWM circuit. As for the linear voltage regulator, not much savings was it? |
11-01-2002, 06:22 PM | #11 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
I have to admit, the PWM is by far the most elegant solution. Now if I can replicate this circuit for my RC boat... but I need a slow reverse.
|
11-05-2002, 02:05 AM | #12 |
Cooling Savant
Join Date: Aug 2002
Location: Pa - USA
Posts: 264
|
Fan "chatter", etc... from PWM can be fixed with a capacitor on the output leads.
|
11-05-2002, 03:42 AM | #13 |
Thermophile
Join Date: Aug 2002
Location: Just shut up ;) ...
Posts: 1,068
|
I've got two 92mm Panaflo's on my HC, they whine and pulse when at full power, I just rigged them both to a variable resistor, tweaked them down till they were quiet.
PWM is good for running fans at very low speed is'nt it?, it allows them to run at a lower stalling speed?... |
11-05-2002, 08:44 AM | #14 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Yes, that's right, but the difference is minimal.
The reason that you can, is because the pulse is at a higher voltage, rather than a straight out low voltage. The high pulse keeps the motor running, while the low voltage will simply cause the fan to stall. |
11-05-2002, 06:32 PM | #15 |
Pro/Staff
Join Date: Oct 2001
Location: Klamath Falls, OR
Posts: 1,439
|
Be careful, though. Too much capacitance will turn the PWM fan control into a DC-to-DC converter. You will end up with a variable output (unregulated) switching power supply. Kind of like the 12VDC rail on most cheap PC PSU's. Well, I take that back. I think it's the 12VDC rail that's unregulated. It might be the 5VDC rail.
In other words, large amounts of capacitance will smooth out the PWM output such that the pulses are gone. Instead of having peaks at 12VDC and a 25% duty cycle, you will have a smooth 3VDC feeding into the fan. The fan will not run on 3VDC, but it WOULD run on a 25% duty cycle PWM waveform. |
11-05-2002, 08:48 PM | #16 |
Cooling Savant
Join Date: Apr 2000
Location: Bremerton, WA
Posts: 514
|
Ok.. So you tried the 7v trick I see.. Not enough air flow that way.. Do you think 10v would do it? Doesn't involve buying any parts or building anything to get 10v.. The blue wire on the ATX connector going to your motherboard is the -5v line.. Use that for your fan's negative, and the standard red line from the power supply, which is 5v, run that to your fan's positive.. That will give you 10 volts.. You could also get 8.3v if you want.. -5v for the negative, and orange 3.3v line for the positive..
Hope that tip can save ya a buck or two.
__________________
The UnaClocker Watercooler Extraordinaire! Overclock till it goes BOOM! |
11-05-2002, 09:03 PM | #17 |
Cooling Savant
Join Date: Apr 2000
Location: Bremerton, WA
Posts: 514
|
Sorry about that, make that the WHITE wire.. Blue is -12v, not -5v.. You'd have 17v with the blue wire.. Which is also sometimes handy, if you want MORE flow..
__________________
The UnaClocker Watercooler Extraordinaire! Overclock till it goes BOOM! |
11-05-2002, 09:57 PM | #18 | |
Cooling Neophyte
Join Date: Aug 2002
Location: Singapore
Posts: 62
|
Quote:
|
|
11-06-2002, 02:16 AM | #19 |
Pro/Staff
Join Date: Oct 2001
Location: Klamath Falls, OR
Posts: 1,439
|
The main point of PWM is efficiency, so the pots (just variable resistance) are used in a way that very little current flows through them. So, you could use tiny wattage pots.
In a rheobus, you are dissipating multiple watts in the resistors, but in PWM, you are dissipating small fractions of a watt. |
11-06-2002, 08:27 AM | #20 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Now see, I thought that the rheobus was a voltage regulator, and that the pot didn't have lots of current running through it anyways...
|
11-06-2002, 09:43 PM | #21 |
Pro/Staff
Join Date: Oct 2001
Location: Klamath Falls, OR
Posts: 1,439
|
Well, it depends upon who you talk to. Rheobus technically stands for rheostat bus, and a rheostat is basically a high-power pot (or variable resistor). However, some people use the term rheobus for a linear voltage regulator based fan speed controller (using an LM317, for instance).
But, the LM317 is no more efficient than a rheostat. The linear regulators dump just as much waste heat as a rheostat for a given voltage, they are simply cheaper. Also, the LM317 can maintain a steady voltage no matter how the load changes, but that really does no good for a fan bus; the load never changes. It looks more high tech, though. |
11-06-2002, 11:41 PM | #22 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Well that certainly explains the heatsinks!
|
11-09-2002, 12:34 PM | #23 |
Cooling Savant
Join Date: Nov 2001
Location: Suwanee, GA
Posts: 130
|
Una,
I've never thought bout using the rails that go to the MoBo. I guess there'd be no problem tapping in to some of the -V rails on the mobo as I've done that beofre to power a comair-rotron 24V fan, but I've never had the PSU powering a MoBo @ the time. would that be safe, or will messing with those rails cause issues? I've got a PCP&C 400W silencer, so I don't the PSU will have a problem, but I'd want to make sure b4 I fry an IT7 cuz they aren't cheap. Also, how would you tap into the rail AND have the ATX headder intact? thanks. I already bought the vantec nexus, but this opens up a whole new range of fan tweaks to me as I love the 7V trick.
__________________
My only PC is a PII 333 with 256MB RAM a GeForce2 MX 200, and 30GB's of 7,200rpm ATA/66 goodness. And yes, I know it sucks. |
11-09-2002, 01:27 PM | #24 |
Pro/Staff
Join Date: Oct 2001
Location: Klamath Falls, OR
Posts: 1,439
|
I can answer that because I researched it a bit and I've used PSU's for lots of hobby experiments.
1) Using the negative voltage rails is dangerous because they are very low capacity. They can only supply a bit of current before they lose regulation (that means that the output voltage droops). Even large wattage PSU's don't rate the lesser rails at higher capacity. This is because the negative voltage rails are mostly used for ISA, RS232 and other older I/O standards. There isn't any point at making the -12VDC rail a 30amp circuit. All (or nearly all) the extra headroom in a 500W PSU goes into the 12VDC and 5VDC rails. 2) You can't tap into the ATX power "cable" without splicing into it or soldering an extra wire into the PSU interior. Be careful. That doesn't mean you can't do it, but it does mean you should think about how much current you are demanding of those lesser rails. |
11-09-2002, 01:33 PM | #25 |
Cooling Savant
Join Date: Nov 2001
Location: Suwanee, GA
Posts: 130
|
i got ya.
seeing as how I often swap things around, I don't want to splice my ATX headder only to have to un-do it.. I way atlease power the fans at the diff voltages disconnected from the mobo to see what happens tho! :->
__________________
My only PC is a PII 333 with 256MB RAM a GeForce2 MX 200, and 30GB's of 7,200rpm ATA/66 goodness. And yes, I know it sucks. |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|