FEATURED Building PWM Controller for 4 wires PWM fan

[inVain]

so.... the postman came with a goodies.
sorry for the crude pics, but I can't be bother to find any better cam, ATM...
well, you know.... this joyful feelings

View attachment 178878

I'm quite embarrassed to admit, that after around 4 years I just found out that the 9SG is an anti-clockwise rotating fan

View attachment 178879

should we try a push-pull?

 

[LennyRhys]

I'm glad it arrived safe and sound! Now you can enjoy its silent but powerful cooling!

 

[inVain]

Thank you very much, LennyRhys

I'm still gazing at this beauty, really love the brass on the axis of this fan.
Though already found on most of the Denkis, somehow the Nidec's looks much better.
The proportion looks so much sexier on it.

And the weight of this fan seems to be much lighter than most of the 38mm too

 

[jbernie51]

Hi,

First off, thanks for the great read. I have been following this thread with the intention of one day building a controller that would control six (yes [6]) Delta PFC1212DE fans.

So, now that I built the 0-100% circuit that bing diagrammed, I am having some trouble and can use some critique. There is no control and the single Delta PFC I hooked up for testing is running at 100%..

Here are the photos of what I built:

From the Top
View attachment 179227

Angle 1
View attachment 179228

Angle 2
View attachment 179229

 

[jbernie51]

So, in my haste, I did some troubleshooting myself and found the following issues:

1. Red Circle: Pin got bent on inserting and was not making contact.
2. Red Square: Transferring the power to the opposite side of the BB I reversed the wires.
3. My connection to the power source is wobbly. Changed out the cables for a tighter fit.

View attachment 179230

 

[LennyRhys]

Looks great - very tidy for a breadboard circuit! I assume you have it working now?

 

[jbernie51]

Yup. Got it working. Three Delta PFC1212DE's running nicely. Video will hopefully come soon...

 

[inVain]

I'm glad it arrived safe and sound! Now you can enjoy its silent but powerful cooling!

the nidec really impress me

just got him running around 900rpm on the side panel and I literally hear nothing.
previously, I had real trouble finding the right fan to fill this spot as vibrations and tight clearance on the intake became apparent.

no wonder you really like this fan, Lenny

 

[quanlop93]

Hi everyone,
I just came here for the PWM fan controller. I saw this topic at the top of google search. After reading the whole thread, 112 pages, I find this very helpful.

I am building my WC project, in which I need to control 8 PWM fans for my radiator, 3 more PWM fan for my case. Unfortunately, I can't find any aftermarket fan controller that fits my requirement.

I am not very farmiliar with electronics. Could someone please give the schematic a massive 8 or 12 PWM Fans controller?...

Can I improve it more, like attach an LCD touch panel, temperature sensor, digital flow sensor? And controll them via a PC software? Do I need a raspberry pi, arduino, PIC or MCU to do this? Do I need to put extra diode to protect the circuit from the back EMF when the PSU is turned off?

Why do I need various frequency for different fans? 25kHz for delta fans, 16kHz for San fans and 5kHz for Nidec? I even saw someone pick 33kHz for Nidec fans.

Can the board handle 8 Nidec 12V 5A fans? Is there anyway to use 24V fan with the 12V line of PSU? Because I can buy some Delta 24V fan on Taobao with a cheap price.

Sorry if i ask too much but this is much more interesting than the lab I had. I used the breadboard, scope, DMM in my electronics lab so I can do some basic. But not the soldering.

Thank you for reading this. English is not my mothertongue so please forgive my grammar mistakes.

 

[LennyRhys]

Hi quanlop93!

This thread is not very active now which is why nobody has answered your questions.

A schematic for a multi-channel fan controller will look exactly the same as for a single channel fan controller; the only difference will be that each complete circuit will share the same ground and +5v connections. Adding things like LCD touch panels and sensors etc. would be very complicated, and for PC control you would need to write code as well as have the appropriate interface (arduino etc.). IMO it's much easier and more satisfying to make a simple manual controller with no digital elements.

Fans operating frequency is specified by the manufacturer and depends on the motor in the fan; most common are 25KHz and 16KHz but it's very easy to change the frequency of the PWM output if required. I think some Nidec PWM fans require a much lower frequency of 5KHz.

As for handling large numbers of fans, the controller has no real amperage running through it so you could (in theory) connect any number of fans, but your PSU probably won't like running 8x fans at 5A each! And for 24v fans they probably won't work at 12v.

 

[jbernie51]

Hi Guys,

Complete n00b here to PCB design, and I have some issues... If someone would be so kind to take a look at my schematic and help with the PCB...

https://tools.upverter.com/eda/#designId=45e0fa73148168b8

 

[inVain]

the link took some whiles to open with my current b/w
will be better for somebody else if you just take a quick snap shot of what you're doing and attach it on your post.

just do a "double triple" check your connection when doing the schematic,
take a break after working for few hours, and recheck to avoid regret.

think few miles ahead where you gonna place your board before you line up the components will help.
this is the unique aspect and involving very heavy personal touch.

 

[getgray]

Great thread. I read a lot of it but admit I had to do some skimming. I'm not new to electronics but I'm no EE either. Decent soldering skills and I've made a few PCB's in my time. I have a problem with a dell server I want to try to solve with this controller. Th eserver ramps the fans up unnecessarily when I add a second raid card. That card is at the rear of the box and I can add a quiet fan to it if necessary. I can't listen to the server running at 5000rpm 24/7. Either fix the fans or ditch the server.

Hoping some of you guys with more knowledge than I can tell assist with a couple of questions. The server has 5 high speed fans (60mm I think). Controlled by PWM via 5 motherboard headers. As is often the case with OEM equipment like this it all fits together quite neatly. My plan is to basically intercept the PWM wire of the fans, after the plug and send the PWM signal there. Thus overriding the servers signal.

Couple of questions... I understood the PWM signal is just that, a signal. And the fan gets it's power from the +/- DC leads. But from reading here I see it mentioned people are beefing up their controllers to handle more current, or high current fans. Is the PWM part of the circuit actually supplying some current? Why does the fan current matter? Or maybe it does not, as long as it is being fed from the same headers it always was, like I plan to do.

Since I need 5 fans, and they all need to be the same speed, I had hoped to build the version that was in Post #846:
http://www.overclockers.com/forums/...ires-PWM-fan?p=6791950&viewfull=1#post6791950

Namely the one " for those who wants to hook with many-many fans or even PWM water pump"

That one adds a 556 timer IC and gets 2 additional outputs.

But as compared to the original, it loses C4 and C3's value changes. It also loses R2. Is that right? Or was that an incomplete drawing?

I need to be able to control all 5 fans, that circuit seemed to be the one to do it.

Any help?

Thanks!
Scott

 

[inVain]

essentially,
they both are identical circuit.
I think they're best explained ~15 first pages.

C4 and C3 were supposed to do the supply filtering job.
nothing critical, but it's always a good practice to have them in the circuit.

the 2nd circuit that you post used the CMOS version of the 556.
compared the "TTL" counterpart (NE556) it has lower capability to source /sink current on the output pin.

I think for your need, using a single NE556 will do the job fine (one knob to rule em' all was something I hold dear back then )
or you can just order a TLC556 + NE556 so you can expand your controller capability just like on the first circuit; tripple output, but all at the same setting.

 

[getgray]

Thanks and, Stink! . I missed the CMOS difference. Already ordered (4) of the TLC556IN going by the drawing (2 to use, 2 for spares/2nd ckt). They be here Sat if USPS makes it on time, was hoping to get this sorted over the weekend. No biggie, I can still get my circuit laid out with the part and test with one fan. I'm DEFINITELY wanting to have one knob to rule them all. And if a single chip can do it, that's icing on the cake. If a single NE556 can do it, I'll round that up. I'll go back through the 1st 15 pages more carefully to see, but if you have a pointer to the schematic and components using the non-CMOS chip I welcome it.

The first circuit was meant to be a mix of NE556 and TLC556? Just said 556 in the schematic, I made the leap both were TLC.

We used to have a couple of electronic supply shops around the city, but I think Digikey and Mouser, and alot of local growth in the downtown area put them out of business.

Thanks again very much for the help.

 

[inVain]

the TLC and NE, are both sharing the same pin mapping.
build your circuit with a socket IC and you can swap them clean and easy

I'll be surprised if the local electronic supply don't have the NE556 on strock.
on the other hand I'm not as lucky as you, to be able to get my hand on the dual CMOS version of the TLC556.
I can only have the single version of it; the TLC555.
the CMOS works better for my current fan (sanyo Denki 9SG) due to the fact that it can reach a lower RPM and eliminate the whining noise when compared to the NE counterpart.
but, if I used the CMOS as IC3 on the 0-100% (the schematic can be found on page #2), I can't reach the 100%.... or the 0% (I can't quite remember but I don't have this problem if I use the NE555).

I also tried the NE555 to control 4 of the 9SG at once with no problem.
but since I now opted to run my fans at different speed, I kept the CMOS version for my need.

in any case, it'll be handy to have both version on hand.
and if you happen to go the local supply shop, I suggest you to get the extra part to build the 0-100%.
they don't cost much and offer more stable output frequency, but the connection are little bit complicated.

as for now, I'm using the 2-98% version and quite happy with it

 

[getgray]

Thanks. I figured I wasn't going to be anywhere near 2%, I'll never turn them off, and they definitely will not run over 80% (way to loud). So I had opted for the simpler 2-98 circuit. IC's are easy to get here, if you can wait on them to mail them. Pars are cheap, shipping is the killer (min ~$9). I verified local shops are gone. Radio Shack might maybe have one, but they have all but turned into phone and toy stores, not like they used to be, so doubt it.

P.S., I'm through page 21 now, reading completely, still haven't encountered any discussion on the difference in those 2 circuits... trudging on.. It was pretty to the point until the mcu discussions erupted.

 

[inVain]

that's sad
I remember seeing radio shack's packaging during the early day of this thread.
yeah... the shipping cost is ridiculous for small qty electronic components.
but if I'm looking for a TLC, I'd go for something like mouser as well

oh...
I should mention this earlier,
we can determined the required current for the signal by measuring the resistance between PWM line and ground line on the fan.
my 9SG shows around 2.7k Ohm (I use a crappy meter, though).
calculate this value with Ohm's law: I=V/R (5/2.7k)
I get 0.00185A or 1.85 mA
and if I parallel two of this fans together, I'll get 1.35k Ohm of resistance between the PWM line and the ground
so both fan will roughly need 3.7 mA

your fan resistance value may differ, but you'll got the idea to decide which circuit to be used.
if your current requirement came close to the source /sinking capability of the chip (I'll take 10mA as the lowest value stated on the datasheet), then it'd be a good idea to use the dual TLC556 circuit.
as for the NE556, it can do up to 200mA (but always spare some headrooms)

I don't remember precisely on what page, but I can still remember Bobnova build this controller for Miahallen (pardon me if I write you guys' name incorrectly ) to control a wall of fans... was it nine or more .... it's more than a bunch for sure.
the vid was super epic

I also spotted the issue for the circuit selection on page5, post #88.
pics are gone, but the vids are still there
edit: some vids are gone as well @.@

yep...
the 2-98% is super simple to build.
keep us updated with your project

 

[getgray]

Thanks, will do. Turns out to my surprise a radio Shack in driving distance actually had 4 NE556's. Fan is supposed to arrive tomorrow (I bought a "spare" case fan for testing), I'll check it's resistance, thanks for that tip.

 

[getgray]

Still waiting on the electronic parts but the spare fan arrived today. 4 wire Red - Yellow - Blue - Black. Everything I've read says blue is the PWM line. I have a pretty nice Fluke multimeter, I checked the resistance between blue and black. 34.3k. So that works out to 5/34,300 0.000146A or 0.146 mA. Since mine's drastically different than yours I'm skeptical. These are enterprise class server fans, maybe they are just better. Dunno.

If I did my math right (5) 34.4k resistors would be equivalent to 6.86k, so all 5 would result in 0.73mA.

Right?

I'd be cleanly within the CMOS range if it can do 10mA. The NE part wouldn't know the load is there LOL.