So you want PWM control of your 3-pin fan?

[fullinfusion]

Any chance of getting them from an online supplier? Might take a little extra time, but maybe it would solve the problem.

Yeah I've looked around and same thing. Buy in bulk and still to much money.

Thats what I loved about Radio Shack when we had them in Canada. You could walk in buy a single Cap for 80 cents, now we have the Source and they sell packs of multiple Ohm and voltage parts and no pack is the same.

I swear they just bunch them up together and what ever falls into the package at the packing center you get!

I just find it easier to plug my regular fans into the motherboard hearers and set the speeds in the bios and be done.

I have a handful of temp probes I can hook into the mobo and place the sensor anywhere I want and have the bios on auto with monitoring and Bingo free PWM fans. I have one sensor plugged in and thats in the memory. Running 2666MHz with no fan needs to be watched.

The case I have, and the orientation of the fans, the case stays the same as the rooms ambient temp. Plus next to nada for dust



Just to buy the parts for me where I live, Id probably spent $60 bucks to have the right values to make 2 fans.

I also only signed up because of this thread, I've been looking for something like this for a little while because PWM fan selection is a lot more limited than 3-pin, so it would be great to be able to control a 3-pin as easily as a PWM. First of all, I wonder if you could point me towards some kind of crash course in basic electronics so I can understand what I'm doing a little better? I do have some experience working with designing and building circuits, but the last time I did anything like this was about a decade ago, so I'm beyond rusty.

Now, I am curious about your second circuit... You say it is high-side switched rather than low-side, so would it avoid the problems described by Xander? I would also like to expand it to handle more fans and be powered by molex in the way you mentioned earlier. How would you recommend going about this? Lastly, for now at least, I've noticed that commercial fan controllers and hubs almost always include some hefty caps with the reasoning described as this would allow them to handle more watts per channel. Neither of your circuits have any caps, so I'm wondering if you don't think they are needed, or if you just didn't add them because you weren't planning on running that much power through them?

Thanks.

I didn't see you just joined TPU! good for you and a huge welcome from all of us awesome and helpful members!

We hope you stick around and enjoy your stay

 

[Aviator81]

I didn't see you just joined TPU! good for you and a huge welcome from all of us awesome and helpful members!

We hope you stick around and enjoy your stay

Thanks

I take it that RadioShack.com doesn't ship outside of the US. That's really a shame considering that people who actually need this sort of stuff need specific parts in small quantities.

I did spend some time looking into fan controllers rather than trying to go this route and I think I came to the conclusion that those temp probes were more hassle than they were worth. I've heard that they aren't all that accurate, and in the case of the ones that are, you still have to mount them in places I'd prefer not to (like between the cpu and heatsink, or even under the cpu). Plus the fan control logic on the controllers seems to be pretty basic compared to the ability of software controllers to specify a non-linear fan profile based on core and case temperatures.

 

[Ahhzz]

Just to buy the parts for me where I live, Id probably spent $60 bucks to have the right values to make 2 fans.

You know with this community here, there's probably a large quantity of people willing to hit their local shack or graingers or other supply depot and hook you up....

 

[Aviator81]

You know with this community here, there's probably a large quantity of people willing to hit their local shack or graingers or other supply depot and hook you up....

I'm new here and I'm already getting that impression... Was already thinking if I can get one of these things working I'd be happy to send you one

 

[Velict]

I had to start a new thread because I couldn't update the old one.

Here's the improved mosfet version of the 4pin pwm to 3pin fan circuit. This is closer to Intel's spec which allows for up to 5.25v pull-up and 5ma on the pwm pin. This is 5.15v and 1.2ma. Any logic-level n-channel mosfet capable of a few amps will work fine.

Tested, working. Enjoy.

So what if I add 10 fans to this one circuit? Do i need this circuit per fan, or will one mod work for multiple fans at once?

 

[Shambles1980]

why dont you just use speed fan?
it can control 3 pin fan speeds just fine just set it up configure it and tell it to start minimized with windows...

 

[Lazzer408]

So what if I add 10 fans to this one circuit? Do i need this circuit per fan, or will one mod work for multiple fans at once?

A motherboards fan header is limited to how much current it can provide. With a higher current transistor and a large enough power source you can use as many fans as you want.

 

[Velict]

http://www.mouser.com/Search/Produc...PBFvirtualkey57370000virtualkey942-IRL530NPBF

Rated for 17 amps, and fans will be driven by separate power source. So, No. of fans times their amps = transistor required?

 

[Lazzer408]

Yes but de-rate about 50%.

I chose the 530 because Radio Shack stocks it.

I'm going to stop looking for part numbers. I'm rushing through this and will likely create more problems then I'd be solving. Just find a mosfet that's turned on by 5v on the gate aka logic level mosfet.

 

[Velict]

What does derate @ 50% mean? That the fans will spin at a minimum 50% from the PWM?

 

[Velict]

What does derate @ 50% mean? That the fans will spin at a minimum 50% from the PWM?

So, googling "derate factors" came up with some spreadsheets about amperage loss from temperature. Does this mean that, on the safe side, you'd want a higher amperage mosfet to meet the engineers specifications? I.e: "requires 17 amps, but has 50% degredation" so a 30 amp would be more suitable?

 

[Lazzer408]

Yes. It leaves a nice safety margin. If you have to control a 10a load, use a 20a device.

There are many variables in designing around mosfets used in switching circuits. Specs such as the turn-on turn-off delays and on resistance all add up. The manufactures use max ratings on components.

 

[Velict]

Yes sir, thank you for the information! I will use the 17 amp ratings for another project. I just purchased Koolance fans http://koolance.com/fan-120x25mm-108cfm

and the 60v 30 amp mosfets from radioshack. I'm quite excited. Also some heatsinks, and thermal adhesive.

 

[Lazzer408]

Here's the Intel specs about their PWM design if you need it. http://www.formfactors.org/developer\specs\rev1_2_public.pdf

As if I don't already have a million things on my plate... One of these days I plan on making a PCB layout for a proper buffered circuit (powered off a Molex connector) to do exactly what you want. I have a use for one myself but it seems like everything I'm trying to accomplish these days gets delayed so don't hold your breath.

 

[Velict]

Downloaded that PDF. Thanks a ton. Oh, I actually made a schematic... but I need to remake it because it was for an older version npn transistor... actually I think it was a modified schematic you made, but it uses two 10k resistors. I routed power to 10 fans, and made sure everything was in parallel, and your circuit comes before the fans. Here is a quick, horrible diagram. Pins and everything are off, but labeled so you can get the idea.

 

[Velict]

Here's the Intel specs about their PWM design if you need it. http://www.formfactors.org/developer\specs\rev1_2_public.pdf

As if I don't already have a million things on my plate... One of these days I plan on making a PCB layout for a proper buffered circuit (powered off a Molex connector) to do exactly what you want. I have a use for one myself but it seems like everything I'm trying to accomplish these days gets delayed so don't hold your breath.

I'll try to assist you, because this is a project i'm enjoying already. I'm going to keep updating this thread and try to keep it alive. Later today i'll post a cleaner schematic and people can poke at it. Most of us are adults here, we can't be online all the time! I'm currently deployed, even. hahahah

 

[Lazzer408]

Downloaded that PDF. Thanks a ton. Oh, I actually made a schematic... but I need to remake it because it was for an older version npn transistor... actually I think it was a modified schematic you made, but it uses two 10k resistors. I routed power to 10 fans, and made sure everything was in parallel, and your circuit comes before the fans. Here is a quick, horrible diagram. Pins and everything are off, but labeled so you can get the idea.

Don't parallel the tachometer signals. It won't read the RPM correctly. Leave just one fan's tach connected and the RPM reading shown would represent the rest if they're all the same model.

I was actually talking about making a circuit board layout and having a few made so one could simply solder on their own components. If I go through the trouble, the design will be PWM to voltage to PWM and fully adjustable. The circuit, as it is now, is as simple as it can be to function and so inexperienced tinkerers could assembly it easily.

Go grab yourself a copy of Express PCB and Express Schematic to aid you in your designs and drawings. http://www.expresspcb.com/expresspcbhtm/download.htm

 

[Velict]

I see where you're coming from. We can do variable resistors and swappable 556 sockets http://www.circuitspecialists.com/machine-tooled-low-profile-ic-sockets

Oh, thanks for the schematics link

 

[Velict]

Check it out

 

[Lazzer408]

You need to make some changes to your gate drive. You don't need a logic-level FET if your driving it off a 556. Replace the zener with a 10k resistor (keeps mosfet off). Since the 556 can source and sink it's output so R2 isn't needed either.

If there is any flyback off the fans, don't forget to add a diode (in parallel with the fans) to protect the mosfet.

I personally HATE the 555 and 556 timers. >.<

EDIT - I see your using 5v for IC VCC. Disregard the mods or move the zener to R1 to ground.

 

[Velict]

Yeah, it's basically two circuits turned into one. Your schematic converts 3 pin to 4pin pwm, and the 556 boosts the signal FROM the motherboard to the fans, because there are so many fans it will degrade the signal, and the fans will only run at max speed.

 

[Lazzer408]

My circuit uses the motherboard's 4-pin PWM fan control to control 3-pin fans. There's no degrading of the PWM signal.

Are you trying to control many 4-pin PWM fans?

 

[Velict]

Yes, 10 fans, at .28 amps each.

 

[Lazzer408]

And they are 4-pin fans? If so, you may need a different circuit.

 

[Velict]

They are three pin fans. They need to be converted to 4 pin.