Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Ideas for computer fan control please

Status
Not open for further replies.

pict

New Member
Hi,

I'm building a computer case with 16 fans - 3 on each of 4 radiators, and 4 general exhaust fans. I'd like to be able to control the running speed of the fans in 5 groups - 4 groups of 3 on the radiators and a group of 4 exhaust fans. I'd hoped to use a commercial 5 channel fan controller but the one I want to use is only rated at 4W per channel and each fan has a rating up to 1.8W so I'm concerned I will overload the controller - especially with a bank of 4 fans connected in parallel. The controller will adjust fan supply voltage independently on each channel over the range 7v to 12v.

I was wondering whether I could connect just one fan from each bank to the controller and then have an auxilliary power supply detect the voltage supplied to that fan and feed the same voltage to the other fans thereby bypassing the load on the controller. The auxilliary supply could be powered by a discrete PSU if more than 12v is required to get a 12v maximum output, or it could perhaps use the computer supply 12v rail if that doesn't affect the maximum output reaching that level.

So I'm not entirely clear how to approach the problem. I do know the auxiliary voltage regulation should be reasonably efficient (I don't want to be be dumping heat into the case unnecessarily) so I'm guessing some sort of switched mode solution, but I have no idea how to begin designing this. Can anyone point me in the right direction for a solution please?

Thanks,
John
 
The first thing you should know is a computer fan is more than just a regular motor. It is a brushless motor which means it requires electronics to control the motor commutation (since it has no brushes automatically) to do this. You don't seem them on computer fans because they are all built-in. THerefore you can't simply use PWM switching on the motor to control it's speed/voltage like you can with a "regular" motor- instead of sending a square wave to directly to a motor (which would work) you would be sending it to the electronics which would then use the power to send the proper square wave to the motor. The electronics would see this input square wave as rapid powering on/off and wouldn't work properly. So you can't go for this type of switch-mode solution.

But you could go for a switching DC-DC converter, probably a buck (voltage step-down) converter. If the fan controller is just sending is a signal to the fans to control their speed that is NOT a power signal, then yes, an auxillary power in place of the regular fan supply for ALL fans. I don't know how your fan controller works.

It is not so easy for a switching converter to detect another power supply and have it's output track that. They are usually made to provide a fixed voltage relative to their internal voltage reference.
 
Last edited:
I would just use a adjustable voltage regulator like a LM317 or one of its higher powered cousins. Set up the 2 ressistor voltage control they use with a thermistor as the variable one.
If proper resistor to thermistor ratios are used you will get a near full range speed control based entirely on tempurature.
 
The OP said he didn't want to be introducing more heat into the computer case which is what an LM317 would do beng a linear regulator.
 
If he is running 16 fans to keep something cool I dont think the few more watts a LM317 dumps out in its worst case senario is going to even be noticed.
I was just looking at it as a simple, cheap and effective way of controling that many fans at once.
Far as I know 16 fans running at 1.8 watts each is
only 28.8 watts. At half their running voltage they will only draw about 1/4 of that. or about 7.2 watts. if they are standard 12 volt fans 7.2 watts at 6 volts is 1.2 amps. the LM317 would only be disipating another 7.2 watts also. 14.4 watts total sytem use at half speed. full tilt is double that!

If he needs 16 fans to cool something another 7.2 watts of heat are pointless to worry about.
But if he wants to spend his mony on a different route thats his buisness not mine!

I just with with simple cheap and practical.

So Pict, what do you need to cool with all of those fans?
 
7.2W equates to a temperature increase of 288C above ambient for an passively cooled LM317. Heatsinking with a fan could cut that in half, but that's still too hot for an LM317 to survive...about 70-100C above ambient.
 
Last edited:
Sorry for assuming most people would be able to realize a LM317 would need a heat sink in this type of application.

I figured if he has that much heat trouble in his computer he already knows a thing or two about thermal disipation devices.

If this thread is your baby just say so I will back away peacefully.

Do you need a nap? :confused:
 
16 fans in one PC? There is no way that makes sense unless your case is a 5 foot by 5 foot box. What are you trying to do, make it fly? Simply throwing more fans at a machine doesn't make it run cooler, it's all about controlling the ariflow over the parts that really need it. Not that I'm doubting your solution but what exactly is this 'case' you're putting these fans in? I'd like to see a picture. You're going to get better results with a single larger fan than you will with 4 fans running near each other, the air flow from a single fan will be much more laminar, separate fans will create some pretty chaotic airflow patterns which is not what you want when you're trying to remove heat. Also give DC blowers a shot, they're quieter than normal 'fans' and can deal with back pressure better (they're used in low pressure compressors) so they're ideal for exaust fans. Standard twisted vane fans suck with backpressure (exaust) unless you have fans blowing it and a low point and out at a high point. Even then blowers would likley work better. Most modern Video cards have blowers not fans on them for just this reason.

Generally speaking fan controllers for computer don't control the voltage, they use PWM and adjust the duty cycle to control the speed. What do you need to adjust the speed of the fans? A remote control panel or a thermal management system?
 
Heres a LM317 bases switcher type regulator.

(now I am just messing with someone!) ;)

but still keeping to the thread! :D

PS. The LM317 needs a heat sink!
 

Attachments

  • LM317 3 amp Switching regulator.doc
    48.5 KB · Views: 137
Last edited:
Thanks for the responses. To give some background the PC case is 18x18x24 inches (mountainmods.com ascension with trinity front panel). I am planning to watercool 3 graphics processors (~500W max), overclocked i7 (~120W max), peltier thermoelectric pad (~450W max) plus assorted motherboard items (chipset, voltage regulators). So perhaps in the grand scheme of things looking for a more efficient solution to a 20W extra heat issue isn't that significant. However I'm pretty sure I'm going to locate this stuff in the front drive bays underneath an optical drive, which is probably an area with poor airflow. If there isn't a more efficient (i.e. elegant) way to effect a supply solution then I can just look at locating 5 hot regulators somewhere in the main airflow, or get them in TO3 package and mount them on a heatsink (or waterblock). So I'm not completely averse to a linear regulator (I was toying with the possibilities of using an L4962 earlier today as an alternative which claims decent efficiency). The output of the fan controller is designed as a variable voltage power supply (i.e. to drive a fan), so really what I think I am asking is how do I use that variable output voltage to control a separate power supply capable of duplicating the input signal voltage range of 7-12v but delivering up to at least 0.6A at 12v (4 fans). If the solution involves additional cooling considerations then so be it. Just nice if it didn't. I haven't a clue what device I should be looking for to do this. Most adjustable supplies (if I'm reading the diagrams right) seem to feedback some of the output voltage back into the device via the controlling potentiometer. How you would adapt this fedback control to external control is beyond me.
 
Wow... I'm not even going to ask the price, cause I don't want to know. The case, and fan setup with one of the video cards probably cost more than my entire machine, and I have no trouble with modern games. What you're putting together is way past the level of 'obscene'.
 
As I pointed out earlier all 16 fans if they are all 1.8 watts each only draw a total of 28.8watts. At half speed, 6 volt input they are drawing 7.2 watts. given the Linear regulator is having to disapate 7.2 watts also thats 14.4 watts load total off the 12 volt supply. your hard drives pull more than that.

Given your total load plus the fact you are running multiple fan control circuits each regulator would only be having to disipate a couple of watts at most. They easily could be mounted to any open metal frame of the computer case. Just spread them out.
As far as control, the processor fans are typicaly run from the boards themselves so they should not need any external speed control. Or at least the few machines I know anything about use on board processor fan control. And mine runs two 2 watt fans off of the one output for the processor heatsink. Asuming your boards have fan drive outputs for each processor you would not have to control all of your fans just some.
I could be wrong on this though.

The intake and exaust should be able to run off a simple thermistor controled regulator. As the temp rises they speed up as it falls they slow down. Total power loss is not even worth worying about if the wattage numbers you are saying are even close!
If your 970 watt total is correct all fans on full at 28.8 watts are only about 3% of your total load.
I hope this makes some sense.
 
QUOTE=tcmtech;715874]
The intake and exaust should be able to run off a simple thermistor controled regulator. As the temp rises they speed up as it falls they slow down. Total power loss is not even worth worying about if the wattage numbers you are saying are even close!
If your 970 watt total is correct all fans on full at 28.8 watts are only about 3% of your total load.
I hope this makes some sense.[/QUOTE]

I'm on board with the linear regulator solution not being a heat issue. The fan controller has thermister driven sensing with five temp probes so it will run in auto mode as you describe and ramp the fans up and down to maintain the target temperature. However in addition you also get front panel display of temps for each channel sensor, and can adjust the target temperatures, and can see the fan RPMs for each channel. More functionality than I could achieve with a home built solution (**broken link removed**).

Regarding motherboard fan headers, I don't have enough to drive the entire fan load, and I will probably already have the motherboard rail (12v2) well loaded with the PCIE cards and oc'd processor.

So back to the question - how to get an auxiliary power supply to mimic an external voltage (from the fan controller). Possible? Is some sort of power amp circuit possible to adapt set to a gain of 1?

Regards,
John
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top