- Joined
- May 23, 2008
- Location
- London, England
-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Home made fan controller
- Thread starter deadlysyn
- Start date
- Joined
- Feb 21, 2005
- Location
- Boston, MA
HERE is a link to the full size pic of the mockup. The top left variable potentiometer regulates 0-100% output with a 13Kohm-50Kohm range. (A thermistor with a 10-100k resistance in your temp range would would be a good candidate) The bottom left variable is set at 33Kohm, and can be used to adjust minimum speed of fan. The top right variable is used between the npn transistor and the MIC502 output. Its value is 270ohm and can be replaced with a fixed resistor. The circuit could still benefit from some smoothing capacitors here and there, but it does work.
As far as powering fans, you are limited by the transistor used at the output. As an example, I used a TIP122 from Fairchild Semiconductor. It is rated at 5amps continuous.
It had no problem powering 3 of my delta fans. Without active cooling, the transistor does get a little hot (50-70*C), but that is well withing its operating temp. In addition its suggested to have a small heatsink on it, which I didn't put on. The breeze from the 3 fans in the picture, was enough to keep it at 30*C.
Here are some shots with 25%, 50%, 75%, and 100% duty cycle output(duty cycle output is the percent of time a signal is ON at any frequency.)
The temperature stayed within 15 degrees from room temp at full power with 3 fans blowing on the transistor. Without any cooling , the temp shot up to about 70*C in a 20*C room. I also have another candidate to power some fans.
The one on the right, is what I am currently using. It is rated at 5Amps (60Watts). The one on the left is a MJL3281A by ON semiconductor. It is rated at 15A (200Watts)
I will keep messing around with it.
As far as powering fans, you are limited by the transistor used at the output. As an example, I used a TIP122 from Fairchild Semiconductor. It is rated at 5amps continuous.
It had no problem powering 3 of my delta fans. Without active cooling, the transistor does get a little hot (50-70*C), but that is well withing its operating temp. In addition its suggested to have a small heatsink on it, which I didn't put on. The breeze from the 3 fans in the picture, was enough to keep it at 30*C.
Here are some shots with 25%, 50%, 75%, and 100% duty cycle output(duty cycle output is the percent of time a signal is ON at any frequency.)
The temperature stayed within 15 degrees from room temp at full power with 3 fans blowing on the transistor. Without any cooling , the temp shot up to about 70*C in a 20*C room. I also have another candidate to power some fans.
The one on the right, is what I am currently using. It is rated at 5Amps (60Watts). The one on the left is a MJL3281A by ON semiconductor. It is rated at 15A (200Watts)
I will keep messing around with it.
- Joined
- Feb 21, 2005
- Location
- Boston, MA
I have used the LM335 numerous times and it is a solid idea. The circuit is very simple to make, and definetly cheaper, but the fans might suffer from startup issues with low voltage output at lower temperatures.
OP
- Joined
- Mar 31, 2005
- Location
- Stealing your megahurtz at night
- Thread Starter
- #44
HERE is a link to the full size pic of the mockup. The top left variable potentiometer regulates 0-100% output with a 13Kohm-50Kohm range. (A thermistor with a 10-100k resistance in your temp range would would be a good candidate) The bottom left variable is set at 33Kohm, and can be used to adjust minimum speed of fan. The top right variable is used between the npn transistor and the MIC502 output. Its value is 270ohm and can be replaced with a fixed resistor. The circuit could still benefit from some smoothing capacitors here and there, but it does work.
As far as powering fans, you are limited by the transistor used at the output. As an example, I used a TIP122 from Fairchild Semiconductor. It is rated at 5amps continuous.
It had no problem powering 3 of my delta fans. Without active cooling, the transistor does get a little hot (50-70*C), but that is well withing its operating temp. In addition its suggested to have a small heatsink on it, which I didn't put on. The breeze from the 3 fans in the picture, was enough to keep it at 30*C.
Here are some shots with 25%, 50%, 75%, and 100% duty cycle output(duty cycle output is the percent of time a signal is ON at any frequency.)
The temperature stayed within 15 degrees from room temp at full power with 3 fans blowing on the transistor. Without any cooling , the temp shot up to about 70*C in a 20*C room. I also have another candidate to power some fans.
The one on the right, is what I am currently using. It is rated at 5Amps (60Watts). The one on the left is a MJL3281A by ON semiconductor. It is rated at 15A (200Watts)
I will keep messing around with it.
Good information here.
Is there any way I can talk you out of a list of parts that you used to make this controller? Also, how hard would it be to add three 3 pin connectors to use with the fans, instead of soldering them to the board?- Joined
- May 23, 2008
- Location
- London, England
- Joined
- Feb 21, 2005
- Location
- Boston, MA
Good information here.
MIC502
100K variable potentiometer (2x)
500ohm variable potentiometer
Rectifier diode
330uF Capacitor
.1uF Capacitor
NPN transistor 5Amp(60W)
3 pin Molex connector
You could substitute fixed resistors for the variable ones, but you will lose ability to fine adjust range. Also, I haven't received the thermistors yet, but once the range of resistances is known, then a suitable thermistor can be put into the circuit to complement R1. When they arrive, I will give you a link to the one that worked best for me.
- Joined
- Jan 6, 2003
- Location
- I can see walmart, 44906
I support the suggestion of using a microcontroller.
Get one with A/D inputs, for each temp probe you want to use, pwm D/A output for each channel you want. higher resolution ADCs and DACs will affect how fine you can adjust the speeds to temps.
you'd connect your thermistors to the ADC input with a pull up/pull down resistor network from a regulated reference voltage. you could probably figure out temp to resistance info from the thermistor's datasheet, and use that resistance to choose the resistor to pull the thermistor up or down depending on how you choose to arrange it.
The pwn outputs could be connected to some good sized switcher transistors, but you might need to use another smaller switcher between to drive it, if the UC cant sink or source the current to switch fast enough, the transistor spends more time in a resistive state than fully switched on, so it dissipates the voltage as heat. that's why pwm is so much more efficient than linear regulators, pwm reduces the ammount of time the voltage is flowing, while the linear takes some of the voltage and does something with it, turn it to heat. just look at the current needed for the fans you'll be using, get transistors to handle however many fans will be on each channel and maybe oversize it by 50 or 100% for room for expansion, then make sure the microcontroller's outputs can drive it. if not you'll need some kind of buffer.
after you have the thermistors and pwms set up, you just have to set up the software. it can be as easy the program running a loop checking if input is 1, if yes go here, if no go next. next is check if 2... it keeps counting up till one hits yes, then each yes will have a different goto which would be a line or 2 to set the associated pwm channel's output. you could use any number of fan speeds, depending on how many you want to write, could go in 5 or 10% increments or all the way down to 1 or even less than 1% increments depending on the resolution of the PWM, so you could program a scale that should transition the fan speed pretty well with temp. it might be annoying to hear the fans speeds go into an up-down oscillation from sitting at just the right spot to cool down/heat up again when the heat load stays at a constant. I know it's annoying to hear the speed go from silent to blastin just liek that, if its speed creeped up as the temperature did, it'd be less noticable.
then you can serial or USB interface to the computer, add LCD screens, flash memory for temp logging... do alot with microcontrollers.
If you didnt want to do all that you could put the thermistor resistor network to an op amp, feed that output to the pwm circuit initially talked about and you should be set, use the gain of the op amp to change temp transition range, should be able to get a decent transition.
Get one with A/D inputs, for each temp probe you want to use, pwm D/A output for each channel you want. higher resolution ADCs and DACs will affect how fine you can adjust the speeds to temps.
you'd connect your thermistors to the ADC input with a pull up/pull down resistor network from a regulated reference voltage. you could probably figure out temp to resistance info from the thermistor's datasheet, and use that resistance to choose the resistor to pull the thermistor up or down depending on how you choose to arrange it.
The pwn outputs could be connected to some good sized switcher transistors, but you might need to use another smaller switcher between to drive it, if the UC cant sink or source the current to switch fast enough, the transistor spends more time in a resistive state than fully switched on, so it dissipates the voltage as heat. that's why pwm is so much more efficient than linear regulators, pwm reduces the ammount of time the voltage is flowing, while the linear takes some of the voltage and does something with it, turn it to heat. just look at the current needed for the fans you'll be using, get transistors to handle however many fans will be on each channel and maybe oversize it by 50 or 100% for room for expansion, then make sure the microcontroller's outputs can drive it. if not you'll need some kind of buffer.
after you have the thermistors and pwms set up, you just have to set up the software. it can be as easy the program running a loop checking if input is 1, if yes go here, if no go next. next is check if 2... it keeps counting up till one hits yes, then each yes will have a different goto which would be a line or 2 to set the associated pwm channel's output. you could use any number of fan speeds, depending on how many you want to write, could go in 5 or 10% increments or all the way down to 1 or even less than 1% increments depending on the resolution of the PWM, so you could program a scale that should transition the fan speed pretty well with temp. it might be annoying to hear the fans speeds go into an up-down oscillation from sitting at just the right spot to cool down/heat up again when the heat load stays at a constant. I know it's annoying to hear the speed go from silent to blastin just liek that, if its speed creeped up as the temperature did, it'd be less noticable.
then you can serial or USB interface to the computer, add LCD screens, flash memory for temp logging... do alot with microcontrollers.
If you didnt want to do all that you could put the thermistor resistor network to an op amp, feed that output to the pwm circuit initially talked about and you should be set, use the gain of the op amp to change temp transition range, should be able to get a decent transition.
OP
- Joined
- Mar 31, 2005
- Location
- Stealing your megahurtz at night
- Thread Starter
- #48
I just wanted to post up, and let everyone know I didn't forget about this. I have been a little busy. Especially since I have been trying to hunt down a good price on a new front axle for my wifes Honda Passport. I broke it when I decided I wanted to go play around in 4WD in the snow. Needless to say, if you are on ice, and the speedometer says 40, you want to be sure the front end doesn't end up getting traction on something. It was fun til it broke, and now because of it, I may end up broke.
- Joined
- Feb 21, 2005
- Location
- Boston, MA
I have this on my backburner as well DS. Life has me cornered right now, but I do intend to finish this project.
- Joined
- Jun 26, 2006
Hey guys, how is this project coming along? I'm thinking about making my own to control 6 delta fans (about 50 watts). and dunking the thermistor into my reservoir. What I don't understand is what type of thermistor or if I need an extra resistor in parallel with it???
OP
- Joined
- Mar 31, 2005
- Location
- Stealing your megahurtz at night
- Thread Starter
- #51
Right now, life has been throwing me so many curve balls, that I have had to put it on the back burner. Right now, I have my fans on a Rheobus, and it seems to be doing the job. I will be sure to post up and update this thread when I actually have the time to build my own.
- Joined
- Feb 22, 2009
Check out the link in my sig for my PWM Fan controller. its about as big as a pack of matches. Cost of components is not expensive, assuming you have a soldering Iron, Flux, 0.75mm Solder, tip cleaner, cutters, strippers, stock 22-24 guage wire in different colors, a breadboard to practice on. I did the circuit on both a proto board and also just soldered the components direct to the IC chip. (I would recommend practicing on the breadboard first). It was a fun project that took me about a week, but it works like a charm.
- Joined
- Sep 28, 2004
I was wondering the other day; "Why doesn't OCF have an official fan controller circuit?" Enough questions and ideas have come through this section that someone could post a front-page article for an inexpensive fan-controller project and make it a sticky. If enough interest was generated, someone could fab the pcbs and put together a kit. That may raise some market questions on the forum, but I think theres a lot of room to run with this. Thoughts?
Home made fan controllers have been brought up many, many times. In fact, I've been thinking about bringing the topic up again myself. Because of the popularity of the idea, I would tend to agree it might be pertinent for someone with experience addressing these threads and a good amount of knowledge in the area to compose a front page article on the topic. Personally I'd be very interested in a front page article on fan controller circuits.
The issue of pcbs I think would be more like the way people build each other LN2 pots and that sort of thing, it doesn't necessarily (and probably shouldn't be) an official OCF circuit.
The issue of pcbs I think would be more like the way people build each other LN2 pots and that sort of thing, it doesn't necessarily (and probably shouldn't be) an official OCF circuit.
- Joined
- Nov 12, 2002
- Location
- Rootstown, OH
Not sure how relevant this is, but it's maybe half the battle:
http://www.overclockers.com/forums/showthread.php?t=641111
http://www.overclockers.com/forums/showthread.php?t=641111
- Joined
- Sep 28, 2004
EarthDog; sure theres lots of prebuilt models out there, but half the fun sometimes is making it yourself. I imagine for those of us who already have the tools for soldering, the cost to make a decent one is pretty darn low.
I.M.O.G.; A prime example! I thought it would be a neat idea to have a dedicated [Official] example. But its just an idea.
Just trying to enrich the forum.
I.M.O.G.; A prime example! I thought it would be a neat idea to have a dedicated [Official] example. But its just an idea.
Just trying to enrich the forum.
- Joined
- Nov 12, 2002
- Location
- Rootstown, OH
Best way to get a dedicated official example is for someone to write it up as a guide for a sticky or the frontpage. If you write it, we will publish.
