- Joined
- Jul 12, 2002
-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Making high power fan controllers - guide
- Thread starter Ven0m
- Start date
THANKS LARVA!!!!!! The table that Ven0m had posted above had me worried. I didn't like the idea that the range of adjustment changed with the fan wattage. Maybe Ven0m will run this circuit through his simulator and find out if it works that way.
I also found an alternate circuit, but found that the parts through Mouser could get expensive for a multiple channel controller. This one is PWM... has certain adantages and disadvantages.
PWM circuit
I also found an alternate circuit, but found that the parts through Mouser could get expensive for a multiple channel controller. This one is PWM... has certain adantages and disadvantages.
PWM circuit
- Joined
- Jul 12, 2002
Cool, I am interested in a PWM circuit. You can buy a really nice PWM 4 pin molex inline one from coolerguys coolergarage for less than $4. They have a thermister, providing temperature control. They acually work pretty well as the thermister is actually calibrated to useful values.
I bought one to check it out but fried it by accidentally letting its output wires short together. I need to order a bunch of them, they are nice. They run the fan WFO to start it and then slow it via the PWM to match the thermister's decree. If you press the thermister between your fingertips just the heat of your body speeds the fan up to near-full rpm. Turning the fan's output on the thermister slows it to very low speeds. Neato.
http://www.coolerguys.com/840556050001.html
I bought one to check it out but fried it by accidentally letting its output wires short together. I need to order a bunch of them, they are nice. They run the fan WFO to start it and then slow it via the PWM to match the thermister's decree. If you press the thermister between your fingertips just the heat of your body speeds the fan up to near-full rpm. Turning the fan's output on the thermister slows it to very low speeds. Neato.
http://www.coolerguys.com/840556050001.html
- Joined
- Jul 12, 2002
Yeah, I was thinking of putting a pot on one when I get some more. They were only $1.99 when I bought the first one, I guess I should have got more.
What's nice about them is the 2 second 12V startup mode. I'll definitely buy more, and my goal is to learn to build my own version of this circuit. It's the best I've seen.
What's nice about them is the 2 second 12V startup mode. I'll definitely buy more, and my goal is to learn to build my own version of this circuit. It's the best I've seen.
WOO HOO! The wife is letting me order some stuff! I'm going to be building the circuit that Larva posted the link to that utilizes 2 transistors except I'm changing R1 to be a 2k pot so that I can vary the voltage to the lowest possible starting voltage. It should handle 15+ watts per channel (would be more if the heatsink were better.)
Here is my mouser order: You'll notice odd numbers of things. Please ignore that. I'm afraid I might ruin something and have to replace it.
532-7136D 4 4 Ships Immediately 0.660 2.640
317-2091-1K 4 4 Ships Immediately 0.980 3.920
511-TIP30A 5 5 Ships Immediately 0.540 2.700
625-2N3904 10 10 Ships Immediately 0.050 0.500
652-3318P-1-202 5 5 Ships Immediately 0.260 1.300
I've also got some adaptors coming from SVC that I will use to connect my fans. They were cheap and they offer cheap shipping on light orders through first class mail.
I will get some perfboard from radio shack and HOPEFULLY have it built next weekend (at least 1 channel).
Here is my mouser order: You'll notice odd numbers of things. Please ignore that. I'm afraid I might ruin something and have to replace it.
532-7136D 4 4 Ships Immediately 0.660 2.640
317-2091-1K 4 4 Ships Immediately 0.980 3.920
511-TIP30A 5 5 Ships Immediately 0.540 2.700
625-2N3904 10 10 Ships Immediately 0.050 0.500
652-3318P-1-202 5 5 Ships Immediately 0.260 1.300
I've also got some adaptors coming from SVC that I will use to connect my fans. They were cheap and they offer cheap shipping on light orders through first class mail.
I will get some perfboard from radio shack and HOPEFULLY have it built next weekend (at least 1 channel).
Last edited:
OP
- Joined
- Sep 7, 2003
- Location
- Warsaw, Poland
- Thread Starter
- #28
The cost parts for 4-line controller I presented with nice pots was around $5-6, at least for me. Today I've tested on simulator a version that is more stable and allows min voltage of 0v or eventually choosing min voltage on separate pot (these simple pots for PCBs arearound $0.20 so it's no big deal). Just now it's improved.
Didn't mean to thread crap. I fully intend to try your circuit eventually. It's pretty elegant in it's simplicity. I want to try the PWM circuit too.
Are you pricing from the components that allow you to purchase only 1 item and not 1000? I did quite a bit of shopping and couldn't find examples as cheap as you say. Just wondering... not insulting.
Edit: As I understand the circuit, no matter how you bias that transistor... The range of useful motion on the pot will vary with load. You will have to recalibrate every time you change fans. Again... As I understand (which is just barely).
Are you pricing from the components that allow you to purchase only 1 item and not 1000? I did quite a bit of shopping and couldn't find examples as cheap as you say. Just wondering... not insulting.
Edit: As I understand the circuit, no matter how you bias that transistor... The range of useful motion on the pot will vary with load. You will have to recalibrate every time you change fans. Again... As I understand (which is just barely).
OP
- Joined
- Sep 7, 2003
- Location
- Warsaw, Poland
- Thread Starter
- #30
PWM circuits seem to have more flexibility and will be more accurate, with more expectable voltages no matter how powerful fan you use. I also don't want to force you to use specific controller type.
Weird price issue - I don't use prototype boards, only PCB from old electronics stuff with removed metal parts, in fact it can be anything that doesn't conduct electricity. Rest of parts, you can usually get for nice price "in local electronics shop". Here in Warsaw, there are 2 underground passages close to each other with many computer and electronics shops. Most offer good prices for parts and these prices drop when you buy more (first price drop usually with 5 pieces of same kind). Of course if I bought 1000 parts from reseller, I'd pay less for each, but I don't need that much. I think that you can find some electronics shops in your city that will offer everything you need with acceptable price.
In fact price was the main reason, why I decided to post this guide. Here, nice controllers from shops are usually $50-70, sometimes much more. And cost of my first ghetto controller (around $0.30 I think) inspired me. It was simple "cut wire, place transistor and simple pot, simple rad and ghetto enclosure", but it worked.
If you want really good controller, I found some DIY USB-driven yesterday on forums here. I think that after simple transistor controller issue is explained somewhere infuture, we'll be able to focus on advanced controllers with USB interface and temp sensors. Target seems to be controller that can have freely settable temperature-RPM curve (na matter which fan is used, unless max RPM is exceeded), settable with USB or some keys/knobs. This would rock, as I haven't seen any controller with such possibilities. My nowledge doesn't allow me to design it fast, so it will require lots of time or help of someone experienced.
Weird price issue - I don't use prototype boards, only PCB from old electronics stuff with removed metal parts, in fact it can be anything that doesn't conduct electricity. Rest of parts, you can usually get for nice price "in local electronics shop". Here in Warsaw, there are 2 underground passages close to each other with many computer and electronics shops. Most offer good prices for parts and these prices drop when you buy more (first price drop usually with 5 pieces of same kind). Of course if I bought 1000 parts from reseller, I'd pay less for each, but I don't need that much. I think that you can find some electronics shops in your city that will offer everything you need with acceptable price.
In fact price was the main reason, why I decided to post this guide. Here, nice controllers from shops are usually $50-70, sometimes much more. And cost of my first ghetto controller (around $0.30 I think) inspired me. It was simple "cut wire, place transistor and simple pot, simple rad and ghetto enclosure", but it worked.
If you want really good controller, I found some DIY USB-driven yesterday on forums here. I think that after simple transistor controller issue is explained somewhere infuture, we'll be able to focus on advanced controllers with USB interface and temp sensors. Target seems to be controller that can have freely settable temperature-RPM curve (na matter which fan is used, unless max RPM is exceeded), settable with USB or some keys/knobs. This would rock, as I haven't seen any controller with such possibilities. My nowledge doesn't allow me to design it fast, so it will require lots of time or help of someone experienced.
This circuit seems to claim that there would be a stable minimum voltage. I don't understand the engineering, but it seems like that is what it claims. If not, then I'm probably not going to be happy with my controller. Well... Happy enough, but it might make me try PWM sooner.
OP
- Joined
- Sep 7, 2003
- Location
- Warsaw, Poland
- Thread Starter
- #32
It should be stable.
An easy mod to get relatively stable 0-11v to circuit in guide would be using 1k pot, with transistor on middle leg and 0 and 12v on sides.
Anyway I'll do some simulation to that crictuit from link.
// EDIT //
After simulating that circuit from link using BD139 and BD140, I have to apologize for lack of professionalism. That circuit is VERY stable due to use of both NPN and PNP transistors. With small costs increase (~$0.20) you get far better controller.
It gives 5.3-5.4v to 11-11.2v depending on power of fan (tested 2-10W), with linear increase almost independant to power of fan. If you replace "that bottom 1k resistor" with simple on-pcb 1k pot, you get min voltage setting in range 0 - ~5.3v and 2k pot - 0-7.4v.
An easy mod to get relatively stable 0-11v to circuit in guide would be using 1k pot, with transistor on middle leg and 0 and 12v on sides.
Anyway I'll do some simulation to that crictuit from link.
// EDIT //
After simulating that circuit from link using BD139 and BD140, I have to apologize for lack of professionalism. That circuit is VERY stable due to use of both NPN and PNP transistors. With small costs increase (~$0.20) you get far better controller.
It gives 5.3-5.4v to 11-11.2v depending on power of fan (tested 2-10W), with linear increase almost independant to power of fan. If you replace "that bottom 1k resistor" with simple on-pcb 1k pot, you get min voltage setting in range 0 - ~5.3v and 2k pot - 0-7.4v.
Last edited:
- Joined
- Nov 16, 2003
- Location
- Cleveland OHIO
YES! this is awsome. only 1 question. how can i add more fans? i have a ton in my case, im switcing to watercooling, and im going to load up my rad, but im keeping the fans inside my case only to keep my mobo cool...shoul i just make 2 of these? or is there i way to wire in more?
Well, The circuit here . Already uses a transistor capable of 36 watts. The problem is finding a heatsink for it that will allow it to get there. In answer to your question, yes you just put more on your circuit board. I've ordered enough parts for a 4 channel 15 watt controller. If you look above those are www.mouser.com part numbers. I have changed the transistors to cheaper (and I hope better) ones.
Your best bet is PROBABLY not to just order the parts that I have until I have tested these exact ones. (Tested and WORKS!)
Edit: The circuit does not distinguish how many fans you are running off of 1 channel. It only cares that you do not excede Q2's maximum collector current or heat dissipation.
Your best bet is PROBABLY not to just order the parts that I have until I have tested these exact ones. (Tested and WORKS!)
Edit: The circuit does not distinguish how many fans you are running off of 1 channel. It only cares that you do not excede Q2's maximum collector current or heat dissipation.
Last edited:
- Joined
- Nov 16, 2003
- Location
- Cleveland OHIO
itll be a long time before im ready for this. ill just hit up radio shack when i go to do so
Channel 1 has been made and works wonderfully! With my little trimmer, I can take the voltage all the way to shutdown. If I were to do it again, I think I would just use the 1k resistor because I think I'll always keep the trimmer smack in the middle. On the 2 noisy fans I've tried it on... The middle setting was more than adequate to take them below the level of the power supply (loud).
I'm probably going to end up making 2 - 2 channel conrollers out of this.
It's fun when you actually make something that works. I'll post pictures sometime in the future. I gotta clean up or my wife will kill me. I gotta sleep too (work third shift).
Couldn't resist...
The top of the board.
My awful soldering.
Update: Finished product in use!
Finished top view.
Finished bottom view.
In my wife's PC.
This is the circuit that I built. I will probably add the kickstart, but don't need it. (I edited Ven0m's pic to show what transistors I used.)
I'm pleased with the final result. If I were to do it again (and I will), I would make the leads a bit longer. It was a pain in the butt to get the wires connected. I would also wire the main pot right the first time (Basically when you turned up the dial, the speed went down). I have no desoldering tools and that was a pain too. I still might go with the trimmers again. It did actually make a difference. I made it so my wife can't turn the cpu fan too low.
All in all it's been a blast. I hope others will try it. If you don't have any soldering skills... practice with some old junk stuff first. I've found the key to soldering is to use soldering flux that can be purchased at radio shack. It makes the parts heat up quicker and more evenly.
I'm probably going to end up making 2 - 2 channel conrollers out of this.
It's fun when you actually make something that works. I'll post pictures sometime in the future. I gotta clean up or my wife will kill me. I gotta sleep too (work third shift).
Couldn't resist...
The top of the board.
My awful soldering.
Update: Finished product in use!
Finished top view.
Finished bottom view.
In my wife's PC.
This is the circuit that I built. I will probably add the kickstart, but don't need it. (I edited Ven0m's pic to show what transistors I used.)
I'm pleased with the final result. If I were to do it again (and I will), I would make the leads a bit longer. It was a pain in the butt to get the wires connected. I would also wire the main pot right the first time (Basically when you turned up the dial, the speed went down). I have no desoldering tools and that was a pain too. I still might go with the trimmers again. It did actually make a difference. I made it so my wife can't turn the cpu fan too low.
All in all it's been a blast. I hope others will try it. If you don't have any soldering skills... practice with some old junk stuff first. I've found the key to soldering is to use soldering flux that can be purchased at radio shack. It makes the parts heat up quicker and more evenly.
Last edited:
OP
- Joined
- Sep 7, 2003
- Location
- Warsaw, Poland
- Thread Starter
- #38
Cool
This even doesn't look bad. You have proven that it can be done and works great.
This even doesn't look bad. You have proven that it can be done and works great.
OP
- Joined
- Sep 7, 2003
- Location
- Warsaw, Poland
- Thread Starter
- #39
So there's an update.
I planned it earlier and wanted to give circuit for uber-controller project, but it was too hard for me, so I decied to at least post one good circuit.
There was a need for good stable controller, so here it is. It requires 2 transistors, but stability is increased, so I think it's worth a few cents more. I decided to add kickstart as well, as fans usually need morevoltage to start, than run. If you don't need it - simply don't use that capacitor. There's also another potentiometer. So why to have 2 pots for each line? Answer is simple - to set minimum voltage. Some fans start at lower, some at higher voltage and this solution allows you useing of full fan range without risk of stopping it. Cost isn't increased much, as assuming that you set minimum voltage once, you can use simple cheapo installation pots, with cost of around 10-15 cents.
You can set calibration pot to zero resistance to have ability to stop fan, but it's rather better to add switch in this case.
Voltage range is 0-11v for 10W fan and 0-11.2v for 2W fan with BD139 and BD140 transistors. Minimum voltage can be set in ranges 0-7.3c for 10W fan and 0-7.4W for 2W fan. If for some reason you need higher minimum voltage, replace R2 with higher resistance pot. Maximum power output will vary, depending on transistors used. In guide there are 2 sets suggested and will give 12.5W / 36W for each line. Total cost remains low.
Unfortunately my circuit for "ultimate controller" had serious flaw that would result in terrible power consumption and will be never posted in way it was designed at first. I have to redesign it and I'll do it when only I have time and get enough data.
I also decided not to release PWM controller circuits, as without stabilization ripple is way too high, and with stabilization they aren't better than simple controllers. If someone convinces me - I'll post them. They could be theoretically controlled by computer with USB connection, but circuit would be more complicated and expensive + I haven't got enough knowledge to make it (at least now).
I planned it earlier and wanted to give circuit for uber-controller project, but it was too hard for me, so I decied to at least post one good circuit.
There was a need for good stable controller, so here it is. It requires 2 transistors, but stability is increased, so I think it's worth a few cents more. I decided to add kickstart as well, as fans usually need morevoltage to start, than run. If you don't need it - simply don't use that capacitor. There's also another potentiometer. So why to have 2 pots for each line? Answer is simple - to set minimum voltage. Some fans start at lower, some at higher voltage and this solution allows you useing of full fan range without risk of stopping it. Cost isn't increased much, as assuming that you set minimum voltage once, you can use simple cheapo installation pots, with cost of around 10-15 cents.
You can set calibration pot to zero resistance to have ability to stop fan, but it's rather better to add switch in this case.
Voltage range is 0-11v for 10W fan and 0-11.2v for 2W fan with BD139 and BD140 transistors. Minimum voltage can be set in ranges 0-7.3c for 10W fan and 0-7.4W for 2W fan. If for some reason you need higher minimum voltage, replace R2 with higher resistance pot. Maximum power output will vary, depending on transistors used. In guide there are 2 sets suggested and will give 12.5W / 36W for each line. Total cost remains low.
Unfortunately my circuit for "ultimate controller" had serious flaw that would result in terrible power consumption and will be never posted in way it was designed at first. I have to redesign it and I'll do it when only I have time and get enough data.
I also decided not to release PWM controller circuits, as without stabilization ripple is way too high, and with stabilization they aren't better than simple controllers. If someone convinces me - I'll post them. They could be theoretically controlled by computer with USB connection, but circuit would be more complicated and expensive + I haven't got enough knowledge to make it (at least now).
- Joined
- Nov 16, 2003
- Location
- Cleveland OHIO
that totally explains why my 92mm wont start half of the time
im definently doing this, in conjunction with an LCD for my watercooling setup..itd be nice to turn 11 fans to quiet!
im definently doing this, in conjunction with an LCD for my watercooling setup..itd be nice to turn 11 fans to quiet!
Similar threads
