I'm at a total loss. I've had so many bad experiences with mosfets not working, handling the current they say they will, etc.
My project has just a couple buttons, and 1 output which is a heat tape that draws 12VDC, 2.5 amps.
I have an ATtiny85, running a 1mhz, 5VDC.
So it sounds so simple but I've spent hours trying to get the right MOSFET, only for it not to work. Does anyone have/use a MOSFET that will work at 5 volt, and PWM 2.5 amps (say 3 amps without heating, to be safe)
If needs to be surface mount, and switch with 5 volt levels.
Someone said this, "Look at all the MOSFETs here with 2.7V and lower Vgs "FET Feature"
which brings the results down to just over 2000 if I filter out 2.7V and under.
In the thousands of MOSFETS at DigiKey I tried to come up with one that would work for 3.3V & 5.0V (like the UNO, or the ATtiny85) applications, to PWM a moderate current 1-5 amps, at voltages up to 12V. I found this:
You shouldn't even be looking at the current specification of a MOSFET, that's not relevant to practical
use. The on-resistance is what you need to go by - calculate the dissipation by I-squared-R and see if
its reasonable, if not pick a device with a lower on-resistance.
The max continuous current rating is usually a maximum power rating in disguise - at maximum power
a device may need to be water-cooled - its not practical, no-one does it.
For instance if a device says 80A, 0.02 ohms, it will be good for about about 15A with a moderate sized
heatsink (5W dissipation).
If you try to use it at 80 amps it will drop 1.6V (meaning the gate voltage has to be 1.6V higher than
its nominal value), and dissipate about 130W. That would mean a large CPU cooler or similar to keep it
from frying. Pick a better device or parallel them up.
Also you have to watch for datasheets that quote the theoretical max current (ignoring the limit
for the package and bond-wires). IR seem to like specifying a TO220 MOSFET at 290A or something
crazy.
So in summary, check the drain/source voltage, the gate drive voltage (ie logic level or not), and the on-resistance. That's all you really need to consider for reliable on-behaviour.
Note that BJT's max current is different, it is a current limit, not a power limit, so if you are used to
BJTs MOSFETs can be confusing.
The IRFR3708 has a worst-case on-resistance of 0.03 ohms at Vgs=2.8, so at about 5A it will dissipate 0.75W
which is feasible for a large surface mount package. So yes, a good choice.
Note that the crazy 61A figure is only possible as a pulse rating if the gate drive is at least 5V - its beyond
anything the package or a PCB trace could handle continuously, and probably will cause premature failure
of the bond wires even as low-duty-cycle pulses. IR at their marketing games again.
"Note that the crazy 61A figure " -- yes, I've learned that amp rating means absolutely nothing!
"The IRFR3708 has a worst-case on-resistance of 0.03 ohms at Vgs=2.8, so at about 5A it will dissipate 0.75W
which is feasible for a large surface mount package. So yes, a good choice."
Great! And I can have a pull down resistor to keep it off during the MCU's startup? 10K??
The actual VGS, TH is one indication, but not more than a pre-selection parameter.
A more important indication of whether a MOSFET will work is whether the required gate levels are actually mentioned in the data sheet - it'ss usually at the top. So if you want to switch using 5V signals, make sure the datasheet explicitly lists key data at 4.5V. For switching with 3.3V levels, look for 2.8-3V voltages.
In many MOSFET data sheets you will only see 10V gate voltage listed, that are not logic level MOSFETs even if VGS, TH is below 2.8V.
I am currently looking for a mosfet to replace TIP120s. I am using IRL540s (through hole) but believe these are end of life (please correct me if I am wrong). Is there a good but simply to understand guide on selecting mosfets or specifically logic level mosfets. I have found a lot of information online but nothing seems to, either be clear to me, or tell me the whole story.
Here is a chart I am putting togerher but I don't really under everything or even if am looking at the right specs. Note that the file is an excel .xls file. Download only if you are happy to.
There are hundreds if not thousands of logic level MOSFETs out there.
The IRLZ44n is mentioned here a lot. Looks like a good replacement for the IRL540 (at least it has a much lower on resistance). Too bad it still needs >4V on the gate, so still not future proof (more and more sensors and processors are 3.3V).
That's a pretty expensive one, and total overkill for most applications (61A continuous drain). The >2400 pF gate capacitance is also not nice. On resistance is very good, though.
Using the IRLML2502TRPBF myself, can switch 4A which is enough for most applications, buying them for just RMB 0.18 a piece (about USD 0.03 each). Even at 45mΩ on resistance no problem with heat for lower currents (2-3A). Also great success with the PMV16UN (5.8A, 22mΩ), cost a little more at RMB 0.22 a piece. Main downside of these is the SOT-23 package.
Through-hole MOSFET for 3.3V gate levels are rare. On Digikey there are a few in TO92 package, but they max out at about 0.5A, except the IRF3708PBF which comes in TO-220 package and can do 61A.
wvmarle:
That's a pretty expensive one, and total overkill for most applications (61A continuous drain). The >2400 pF gate capacitance is also not nice. On resistance is very good, though.
What effect does the gate capacitance have. Does it simply slow down how quickly the mosfet turns on/off?
If I want a fast on/off what should I be looking for?
wvmarle:
There are hundreds if not thousands of logic level MOSFETs out there.
The IRLZ44n is mentioned here a lot. Looks like a good replacement for the IRL540 (at least it has a much lower on resistance). Too bad it still needs >4V on the gate, so still not future proof (more and more sensors and processors are 3.3V).
Thanks. This looks to be good for what I want (5V Arduino) and I can get them for under 1 RMB each.
Yes, it slows you down (but you're usually in the microseconds).
It's also why you need a current limiting resistor between the gate and the Arduino pin: it's as if you're charging a capacitor.
It is mostly an issue when doing PWM as the MOSFET spends some time partly on/off every cycle and that's when excess heat gets produced. The slower the on/off part the more time spent in that region, and the more heat is produced. So a high PWM frequency and slow on/off means the MOSFET is most of the time in the partial on region, and can get really hot really fast.
It is quite complex to calculate the time spent partly open (if you're interested, you'll have to do some searching for it yourself) - it's related to both the gate charge and the gate capacitance, plus of course the drive voltage, the gate resistance and the external resistor.
As a rule of thumb: the bigger current carrying capacity of the MOSFET and the smaller the on resistance the larger the silicon area of the gate, and the larger the gate capacitance.
