FEATURED Building PWM Controller for 4 wires PWM fan

inVain · Oct 21, 2012

using the analog design with SMD,
I can reduce my sanyo SG to ~600rpm.

problem occurs on R3 & R4, it doesn't seem to get along with the smd resistor.
work okay when I'm using a pair of 0.25watt 9.1K metal film resistors.
after I changed it to couples of smd, then fan's ramping up and down like it's crazy :screwy:

I'm still in awe about how precisely the resistance in smd are.
if the label says 103, and when i measure it really 10k (well, some vary to 9.8k though

)

switching back to the regular analog 0-100%.
now my SG's spinning ~1000rpm for its minimum, and the noise is such a pain compared to the ~600-ish

will solder the 7th SMD board now :bang head

TsunamiJuan · Oct 24, 2012

you know you have a good collection of parts going when you can throw together a fan controller on a whim. Will probably be adding another two controllers to this board. Which will be relegated to the bench for fan control. Works great with my Phatty 4 amp nidecs.

View attachment 117623

View attachment 117624

inVain · Oct 24, 2012

I love perfboard.
it's simple, and you can always use it as a working unit :thup:

I finally manage to have a working unit (after spending 7 PCB

).
at least I got the frequency, polarity selector and the LED indicator working.
that practically leave me with the kick-start function missing.

I got something weird on this SMD setup, since when I leave the 1st 555's reset pin open without connecting it to the +5v rail, nor to the kick-start circuit, I still can get a fully control over this PWM controller :screwy:

(shouldn't it set the duty cycle to max?)

I wonder if got the wrong type of the timer unit (will check this out after the sun rise

)

and I finally manage to find out that the flux was cost me the trouble for this long.
maybe I used to much flux and it messed up with the tiny space on this SMD board.
hope this might be useful to someone who is intended to build something like mine.

now I'm quite happy with my current progress so far,
and even without the kickstart, the controller was able to power up the fan (sanyo SG) at around ~500rpm-ish without any issues :bday:

Bobnova · Oct 24, 2012

Many ICs have an internal pull-up resistor on the reset pin. It's a high enough resistance resistor that a strong EMF applied near the pin can sometimes pull it low though, so having an outside thing pull it high is a good idea.
Atmel MCUs are like that, it sounds like 555s are too.

(thinking and googling happened here)

Looking at the schematic I found here : http://www.electronics.dit.ie/staff/mtully/555 folder/555equiv.gif

It looks like the 555's reset pin is a PNP transistor, and when pulled low it shorts out most of the guts of the thing. If left floating it doesn't allow current flow and hence doesn't reset things.
In this setup a strong EMF could, in theory, pull it low long enough to short things out. It seems unlikely though, a 7k resistor isn't exactly weak. You'd need one hell of an abrupt field to overpower that. I don't think that field would be appreciated much in general.

Of course, this assumes that the schematic is correct.

EDIT:
I'd chalk this one up to accidental discovery on your part

Much like how the LED was invented, Viagra was discovered, and a variety of other things more or less useful than the preceding.

Jasonaaronfox · Oct 24, 2012

inVain said:
and I finally manage to find out that the flux was cost me the trouble for this long.
maybe I used to much flux and it messed up with the tiny space on this SMD board.

i only use flux to tin, then scrub the pcb with acetone afterward. its a ***** to clean up

TsunamiJuan · Oct 25, 2012

Theres no clean flux aswell that burns off clean, I'll see if i can find the part number for the good stuff. But yes alot of the fluxes out there don't burn off clean and can later cause problems if an excess is used. or not cleaned up. Burned layers of flux cause problems :\

http://www.drivestar.biz/amtech-nc560lf-leadfree-noclean-solder-paste-100g-p-1782.html

inVain · Oct 25, 2012

the poor guy only last for 2 hours

Bob, will adding a pull up resistor right before the reset pin will help solve the problem?
since this is the easiest approach, maybe I'll do this before drawing another schematic

Jasonaaronfox said:
i only use flux to tin, then scrub the pcb with acetone afterward. its a ***** to clean up

now I'm doubting what I wrote before :-/

(sorry for my misleading post earlier

)
but the fact that the controller's working, even for a very short periods of time when I only use flux to tin the pads (and not for the soldering) was really made me to jump on that conclusion.

yes, it's really such a PITA to clean the flux after the components were on the board.
but it's still fun to do (and I think it's good) when we're prepping the PCBs for the soldering stage :attn:

TsunamiJuan said:
Theres no clean flux aswell that burns off clean, I'll see if i can find the part number for the good stuff. But yes alot of the fluxes out there don't burn off clean and can later cause problems if an excess is used. or not cleaned up. Burned layers of flux cause problems :\

http://www.drivestar.biz/amtech-nc560lf-leadfree-noclean-solder-paste-100g-p-1782.html

must admit that I'm using excessive amount of flux

never had any issues with single layer boards before (it did the job terrific),
dealing with these dual layer boards, is really giving me a serious headache :bang head

:

it's really such a shame that I don't have too much choice of flux type on the local shop.

still will draw a further revision for the controller.
it's really worthed to do so since these Denkis are begging to be tamed :rofl:

TsunamiJuan · Oct 25, 2012

yeah its alot of flux, but its the correct mix number, they sell it in little syringes aswell.

inVain · Oct 25, 2012

this is the flux that I'm currently using

arpus_lotfet%20100%20GR%20toko%20robot.jpg

linky

the local web even list it for cleaning your solder tip :shock:

inVain · Oct 25, 2012

Bobnova said:
It looks like the 555's reset pin is a PNP transistor, and when pulled low it shorts out most of the guts of the thing. If left floating it doesn't allow current flow and hence doesn't reset things.
In this setup a strong EMF could, in theory, pull it low long enough to short things out. It seems unlikely though, a 7k resistor isn't exactly weak. You'd need one hell of an abrupt field to overpower that. I don't think that field would be appreciated much in general.

Of course, this assumes that the schematic is correct.

Bob, I'm little bit lost with your post :chair:

but I manage to add a 9.1K pull up resistor on the reset pin.
it seems solved the problem, and will leave it running to see if the problem return.

will try to get pics to see if I messed up with the trace :salute:

Jasonaaronfox · Oct 25, 2012

if your problem continues, try putting a layer of substrate over or under the circuit, with another layer of aluminum foil tied to ground on top of that. It will absorb the em field and eliminate parasitic capacitance, etc... proving whether or not thats your problem.

inVain · Oct 25, 2012

and so the problem return :rain:

haven't attach any cable as suggested by Jason, though :chair:

will try to solder a wire from a ground via to the PCI cover where I mount this puppy on (the reason why I must build this one on a very tiny PCB + SMD)

in any case, would you please take a look at the top (red) and bottom trace of my PCB and help figuring out, what should I fix on the next rev :salute:

View attachment 117746
View attachment 117747

U1, U2, U4 are the 555s, and the U3 is the LM311.

thx

Bobnova · Oct 25, 2012

What exactly is it doing?

bing · Oct 25, 2012

TsunamiJuan said:
you know you have a good collection of parts going when you can throw together a fan controller on a whim. Will probably be adding another two controllers to this board. Which will be relegated to the bench for fan control. Works great with my Phatty 4 amp nidecs.

Thanks for sharing this ! :thup:

inVain said:
this is the flux that I'm currently using

linky

the local web even list it for cleaning your solder tip

That thing is corrosive !! Beware and actually its a solder tip cleaner and not suitable for fine solder tip, and after a prolonged used, your solder tip might getting thinner cause that thing is "acid" based flux. :-/

Go to your local cell phone "sparepart" shop (not cell phone repair shop), ask for proper solder flux and its quite cheap and easily washable using alcohol, and some are even packaged in a syringe for easy application.

Something looks similar to this :

View attachment 117755
Randomly grabbed image from google

inVain · Oct 25, 2012

Bobnova said:
What exactly is it doing?

sometimes it makes the fan full stop, sometimes full speed, sometimes up and down, and sometimes it runs just fine.

I leave the reset pin on IC1 open ATM, and weird it still run okay and able to control the fan speed with the pot,
but sadly I won't be able to add the kick-start with this condition :shrug:

bing said:
That thing is corrosive !! Beware and actually its a solder tip cleaner and not suitable for fine solder tip, and after a prolonged used, your solder tip might getting thinner cause that thing is "acid" based flux.

Go to your local cell phone "sparepart" shop (not cell phone repair shop), ask for proper solder flux and its quite cheap and easily washable using alcohol, and some are even packaged in a syringe for easy application.

Something looks similar to this :

View attachment 117755
Randomly grabbed image from google

dayum....so that thing really is a tip cleaner
and the real flux were hidden at the cell phone supply shop :shock:

off we go to the supply shop then :salute:

svtfast · Oct 26, 2012

Will any of the PCS board in this thread work to control PWM fans with temp sensors? I might have this all wrong but....does the PWM sensor work to control the fan speed based on the temperature reading of the sensor? If so, cool, if not, that is what I want to do.

Jasonaaronfox · Oct 27, 2012

svtfast said:
Will any of the PCS board in this thread work to control PWM fans with temp sensors? I might have this all wrong but....does the PWM sensor work to control the fan speed based on the temperature reading of the sensor? If so, cool, if not, that is what I want to do.

Umm... need more information. I can a program for a microcontroller based pwm controller that will adjust the fan speed based on temperature -OR- you can replace the potentiometer in the other, analog, controllers with the temp sensor if it is resistive in nature.

For either of those options, we need to know what type of temperature sensor it is, if the output is a digital pulse, or if the output is a simple resistive change such as that in a temperature sensing diode.

Bobnova · Oct 27, 2012

Fans like the CPU fans that have a built in thermistor and regulate their own speed only do it when they aren't receiving a PWM signal. That's my understanding anyway.
Once you give them a PWM signal they should ignore their thermistor and do what you tell 'em to.

svtfast · Oct 27, 2012

Jasonaaronfox said:
Umm... need more information. I can a program for a microcontroller based pwm controller that will adjust the fan speed based on temperature -OR- you can replace the potentiometer in the other, analog, controllers with the temp sensor if it is resistive in nature.

For either of those options, we need to know what type of temperature sensor it is, if the output is a digital pulse, or if the output is a simple resistive change such as that in a temperature sensing diode.

I thought that I could power the fans (3 Noctua NF-P14 FLX 0.1 amp (each), 2 San Ace 9SG1212G101 4 amps (each) with my PSU. I would reroute the PWM wires to my radiator. I know that the San Ace fans can be controlled with a PWM non mobo circuit, this is what I want. I want the San Ace fans to cool a 240mm radiator. The sensors will be in the radiator and when the temp goes up, so do the RPMs on the fans.

Simple?

Bobnova · Oct 27, 2012

With a microcontroller, quite simple.
With passive components, less so.

FEATURED Building PWM Controller for 4 wires PWM fan

Member

Member

Member

Senior Member

Member

Member

Member

Member

Member

Member

Member

Member

Senior Member

Low Profile Senior

Member

New Member

Member

Senior Member

New Member

Senior Member

Similar threads