"running it fully" as in turning it full on (so the MOSFET reaches its minimum Rds and dissipates very little power) and then full off (so the light is off) at a 50% duty cycle so the light appears dimmer.
If you do Not turn the MOSFET full on, its Rds will be higher, and it will dissipate a lot more power and be much hotter.
From the datasheet:
RDS(on) Static Drain-to-Source On-Resistance ––– ––– 17.5 mOhm VGS = 10V, ID = 25A
There is no data for VGS at 5V.
You can find other parts with Rds that is lower with VGS at 4.5V, for example:
Yeah thought it would be something about Rds. So irfz44n is designed to be fully closed or fully open(?) I should be looking for a fet that has low Rds(on)? Any suggestions for logic level that suits my needs?
The IRF3707ZLPBF will do.
You can browse at www.irf.com, or at digikey, filtering your way down thru the N-channel single MOSFETs and find something similar.
You want to have 12V to the halogen, to the MOSFET's drain, with MOSFET's source to ground. The Arduino driving the gate.
Power (P) = Voltage (V) * Current (I)
P = IV, or P/V = I
55W/12V = 4.58A
V = I * Resistance (R), or V=IR
Sub in IR for V and get P = I^2 * R
Power dissipated by MOSFET then => 4.58 * 4.58 * 0.012 (Rds) = 0.25W of dissipation in the MOSFET when on.
50% duty cycle = 1/8W. Not too bad. Probably won't even feel warm.
If you can find a 5 mOhm part you can get that down to 100mW, 50mW at 50%.
That one is worse!
rDS(ON) = 0.047?
We're after Lower Rds here, not higher.
Just because it says Logic Level it doesn't make it better.
The 3707 part I suggested, while not saying Logic Level, has a better Rds at 4.5v:
RDS(on) Max 4.5V (mOhms) 12.5 - just 1/4 of the resistance of the Sparkfun part.
At 4.5A of current it makes a difference.
Gate threshold voltage (logic level or not) is just part of the equation when using a power MOSFET in switch mode applications (such as with PWM power regulation). We also need to look at the gate charge (Qg) to determine its dynamic characteristics. If time required to turn the MOSFET fully on/off is significant in relation to PWM cycle time, it will quickly heat up despite a low RDSon. For example, a device with Qg of 20nC can be turned on in 20msec if 1ma is supplied to the gate or it can turn on in 20nsec if the gate current is increased to 1A.
The current sink/source capability of the AtMega is generally insufficient when it comes to driving a MOSFET effectively in switch mode for high frequencies and high loads. To improve efficiency, we need to add a more capable gate driver circuit. One such example is the IRF4428 low side driver IC capable of sinking and sourcing 1.5A in the 6-20V range. If we add this driver in between the Arduino PWM output and the MOSFET we will be able to benefit from low RDSon also in switch mode. Since this driver can operate its output in the 6V to 20V range we‘re also no longer restricted to logic level MOSFETS.
You can browse at www.irf.com, or at digikey, filtering your way down thru the N-channel single MOSFETs and find something similar.
The "...and find something similar" appears key to your suggestion as Digikey is not going to restock the IRF3707ZLPBF when they run out. Mouser and Jameco do not list them.
I actually might be satisfied with ifrz44n. I've tested some more and let the bulb on for half an hour. It gets really hot but it still works in 21c ambient. Tried to take readings with IRtempmeter and got just over 70c with it. Dunno what the real internal temp is. I will order IRF3706 to test with.
Could it be possible to run it with arduino? I mean the 5V could be enough for it.
When measuring fet's gate i get 2-3V when the bulb is as dim as i want it to be. So i reckon arduino could easily be enough for it?