I'm trying to drive about 20 5meter RGB led strips with my arduino. I figured I need MOSFETs too switch my 4x10A 12v. According to the datasheet 5V (from PWM) should be enough to turn on/off the MOSFET. The thing is it doesn't work. Digging deeper in the datasheet it seems that the PWM frequency of the arduino is too slow. I need pulses of 20µs so 50KHz if I'm correct?
I'm trying to drive IRFZ44N MOSFETs http://www.irf.com/product-info/datasheets/data/irfz44n.pdf (pdf)
when following these instructions on the forum http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1152547089 My voltage out from the PWM drops from 0-5 to 0-1.8 and now back to 0-5. I think I fried the MOSFET again.
I'm missing something here, can anyone get me on the right track? The LEDstrips have been bought, as well as 4x10a 12v DC power supplys... only part missing is my RGB controller. The commercial solutions are always too light and I want to control it with arduino.
What bit says you need 50kHz? The LED datasheet?
I've used power MOSFETs down at two or three hundred Hertz (s/w PWM on a PIC) to drive strips of LEDs with no problems.
What are you using to measure the voltage - a scope, I hope?
Digging deeper in the datasheet it seems that the PWM frequency of the arduino is too slow. I need pulses of 20µs so 50KHz if I'm correct?
You are not correct, FETs will switch at DC, there is no such thing as too slow for a FET.
My voltage out from the PWM drops from 0-5 to 0-1.8
How are you measuring this, you can't measure it with a voltmeter only an oscilloscope.
Edit AWOL beat me to it. :)
I belive the problem is that the MOSFET you are using is not a logic level device, that is it cannot fully turn on at the voltage level that an Arduino I/O pin can supply. Threshold voltage spec is not the same as full current gate voltage. It's possible the device operated too much in it's non saturated region and got to hot to survive.
Most data sheets for logic level MOSFET transistors will clearly state they are compatible with logic level gate drive voltage.
Yes the clue here is in the data sheet where it says:-
RDS(on) Static Drain-to-Source On-Resistance ––– ––– 17.5 mW VGS = 10V, ID = 25A [ch132]
That means it is fully turned on with a gate source voltage (VGS) of 10V.
Any less you are in the liner region and therefore dissipating heat as there is some extra voltage being dropped across the FET.
You should pull up the gate with a transistor to the 12V rail. Or get another FET.
Hey thanks for all the info. I guess I cannot hide my noobness... I used a voltmeter.
I will try and get hold of the logic-level version of the fet. The thing is I explicitly asked the shopowner for a logic-level mosfet.
So I think I'm gonna use the same mosfet but the logic version, that's less parts to deal with... it's this one http://www.digchip.com/datasheets/parts/datasheet/232/IRLZ44N.php, I should get it today
would doubling the PWM voltage like this work?
or is Mikes method to pull up the gate with a transistor to the 12V rail the preferred way? I hate to ask, but if so could someone translate mikes sentence to a diagram? I think it looks like this (first time I draw this, be gentle):
the irlz44n I ordered won't be in before tomorrow morning, and I wanted to leave it on the whole weekend for stress testing. What kind of transistor would I need? would I need to add safety measures?
Any type of small signal general purpose transistor (NPN) will work like a BC183.
Sorry for the crude drawing but I only have a very early photoshop here.
thanks! trying it this afternoon
Well not much success with the transistor, although I'm not sure if my setup was good. I think I blew up my only transistor.
AWOL: good Ohm or bad Ohm?
I'll have to wait for the parts till monday...meanwhile I'll get some extra transistors.
Keep in mind that adding a NPN to switch the MOSFET, as shown above by Micke, will add an inversion to the output such that 0% duty cycle will be 100% current on for the LEDs and 100% duty cycle will be full off. You just have to keep track of that in your sketch.
Well not much success with the transistor, although I'm not sure if my setup was good.
It is defiantly your set up, you can't blow up the transistor with that schematic unless you miss identified the legs of your transistor or other components.
So I finally figured everything out. I learned a lot. Had to acquire an oscilloscope in the process to find the broken parts though. In the end, I think it's the easiest way to use the logic level version of the MOSFET so it's directly compatible with the arduino PWM-out.
The only thing I experience now is that the power supply makes a buzzing sound if one or more of the three PWM signals is not 0 or not 255
I'm hoping this won't hurt anyone...
Everyone thanks for the help.