Need a little help with old problem, fan control

charleyshf · 2011-09-29 10:23

Hello,

I've been dealing with a situation every year with of all things my multi fan controller for my pc. Almost every year or so the "low end" fan controller I have starts to lose it's ability to control my fans and just burns out or stops working. This time I decided to not buy another one ($39.95) and I want to use the prop this time. I have most of the components I believe that I will need to do this.

I am not worried about looks, I just want something that can control my 3-pin 12vdc .30amp fans.I have a total of 4 of them. I have been looking around the forums for an idea of what I need for parts and the first thing that I see is that I need some sort of mosfet(s), right now I have a bunch of mosfet's from RS (IRF510) however it looks like I should be looking for IRL510's.

The only other things I thought about doing was to also make use of the RPM cable from each of the fans just to monitor rpms and maybe use my lcd display to show the status of each fan's rpm.

Is there a better choice of mosfets to go with?

I'd really appreciate any help.
Thank You

Toby Seckshund · 2011-09-29 11:30

I built a bench PSU a while ago and as there was an AVR doing the Volts and Amps metering I thought that it might as well sense the temp of the heatsink (internal) and vary the fan speed. I went for PWM and soon found that there was some audible interaction between the PWM rate and, presumably, the "cogs" of the brushless fan. I ended up giving the AVR a faster crystal to minimize it whereas I had intended to just use its internal 8MHz.

I now wish I had gone for some form of stepped resistance, if the PC made that noise it would drive me even more nuts!

PJAllen · 2011-09-29 11:43

If you have small-signal NPNs you could use those to switch/control the IRF510 gate voltage - or you can go buy more stuff.

charleyshf · 2011-10-05 11:11

Thanks for the responces. I am looking at using as few components as possible, so I am thinking about getting the RFP30N06LE specs here: http://www.sparkfun.com/datasheets/Components/General/RFP30N06LE.pdf . I should be able to interface the prop to it without any additional electronics. I am hoping to avoid the audible noise that I have seen people running into with using PWM.

JonnyMac · 2011-10-05 15:46

I like using the Microchip TC4427 to drive MOSFETs; it serves as a level shifter between the output of the Propeller and the best gate voltage, and it's push-pull output lets it switch very fast in either direction.

Peter Jakacki · 2011-10-05 18:35

I'm wondering why all the kerfuffle over some itty-bitty fan motors. Skip the MOSFETs as they are overkill and just use NPNs which the Prop has no problem driving both in terms of voltage and speed. Do you know that you can get 4.3A NPNs in tiny SOT-23 packs? This is just an example. With extremely low Vce-sat and high gain these devices are great for jobs like this. You can PWM as fast as you want (to a point).

However, to get the fan really quiet at all speeds you should feed it DC and that's rather easy if you basically construct a DC-DC converter using an inductor, diode, capacitor, along with your PWM element.

There is of course the linear regulation method which although more inefficient you are still only talking about 300ma @12V motors.

charleyshf · 2011-10-09 10:32

Hi,

Boy you two have given me some things to think about here. I built a circuit that I found a picture of on this forum(attached in this message) on my breadboard and the RFP30N06LE Mosfet that I just got in last week. I also used a pwm object from the OBEX from here: http://obex.parallax.com/objects/216/ . With some minor changes to the demo code I am able to control the mosfet and change the voltages (watching the output on a voltmeter) however I am trying to understand a few things. First is that I should have the democode changing the period and not the duty from what I have been reading, and so far I have been able to change the voltage output of the mosfet changing the duty, so I am uncertain there.
Second should I have a diode in place to prevent feedback if I am controlling fan(s)?

Last is there better code that I could be using so I can understand this better?

@Jonnymac - The Microchip TC4427 , would I use this to deal with the noise issue I hear other talking about with the fans? I also found a few different variants of this chip, looks like it's down to delay time of 20ns or 30ns, which would be better?

@peter - I guess I am just trying to find the best way to control a few fans in my pc, after reading around on the forums I assumed that using a mosfet would be the best way.

Thanks again for the help.

mosfet_drive.gif

kwinn · 2011-10-09 10:52

charleyshf, for something as low powered as these fans you could use a ULN2803 with 2 of the transistors in parallel for each of the 4 fans. You might also consider using a low frequency (10-20Hz) or frequency modulation to control the speed of the fans. The mechanical inertia of the fans makes high frequencies unnecessary in most cases.

Peter Jakacki · 2011-10-09 18:38

Believe me Charley using a MOSFET is sometimes the best way but not always. With such low currents a BJT either as discrete transistors or the ubiquitous ULN2803 will do the job better. Unless MOSFETs are driven with sufficient voltage and high switching current they will not perform anywhere as well as they could. In comparison a BJT is easily driven from the Prop and does not require any high switching current to overcome gate capacitance for fast switching.

I'm not sure if fans are all that happy being driven at low frequency as kwinn mentions but it is easy enough to find out. Shucks, the whole thing could be breadboarded (with MOSFET or BJT) and tested in less than 15 minutes. Go on, stop asking questions and do it

charleyshf · 2011-10-10 07:58

You know I would try it, unfortunately I don't have a single ULN2803 to test with(but found a few cheap deals on EBay (10 for $6.59 free S&H), so it's going to be a few days....

PJAllen · 2011-10-10 08:17

I was trying to use what you had already bought, providing an answer within the confines of the question posed. Using an NPN, which I assume most everyone has a few of, to switch the gate voltage for the IRFs would've solved this intractable "problem" a long time ago.

I could've made a pitch to buy thus and sundry, too, but as you'd already cast your lot I figured why not work with what you've got.

PJAllen · 2011-10-10 17:44

Here's the circuit --

See the effect here!

SPIN "code" attached.

charleyshf · 2011-10-12 06:06

Thank you for posting this, it's helped me to understand some things I wasn't clear about before..

Need a little help with old problem, fan control

Comments