Heat problem with mosfet and PWM

Using Arduino General Electronics
#1

I connect a motor with a logic mosfet to control it in the simplest circuit. (+ battery → motor → mosfet → - battery) and an arduino to provide 5v to the gate of the mosfet. The battery is 12v.

The mosfet have a nive 5v gate to source and should be fully open and should be able to lets about 110 amps which I think it’s enough for my motor to go to it’s full potential without any load.

The problem is that the motor doesn’t as fast as when I connect it directely to the battery and does not accelerate as fast it should be.

I really don’t understand, the motor receives a nice 12v and I 110amps should be enough.

What I am missing out?

#2

control it in the simplest circuit. (+ battery -> motor -> mosfet -> - battery)

Is that + battery -> motor -> mosfet drain / mosfet source -> - battery -> Arduino ground ?

and should be fully open and should be able to lets about 110 amps

In your dreams, the data sheet says :- Absolute Maximum Ratings ID @ TC = 25°C Continuous Drain Current, VGS @ 10V 62A

You run into heat dissipation problems long before that anyway.

#3

Beurnii: I connect a motor with a logic mosfet to control it in the simplest circuit. (+ battery -> motor -> mosfet -> - battery) and an arduino to provide 5v to the gate of the mosfet. The battery is 12v.

The mosfet have a nive 5v gate to source and should be fully open and should be able to lets about 110 amps which I think it's enough for my motor to go to it's full potential without any load.

That device has an absolute max current rating of 62A, it will need heatsinking above about 8A (and I wouldn't be tempted to put more than 20A through it really, the abs max current rating is not a place you want to operate at (and remember the max current for 5V gate drive will be sifnificantly less)

The problem is that the motor doesn't as fast as when I connect it directely to the battery and does not accelerate as fast it should be.

I really don't understand, the motor receives a nice 12v and I 110amps should be enough.

That motor has a stall current of 100A, beyond the MOSFET's handling at 10V gate drive, and even further beyond its logic-level handling. (*)

Try parallelling 3 of this MOSFET type (remember gate resistors when paralleling), or get a higher performance mosfet.

For such high powers I'd never dream of using a logic-level MOSFET, they are much less robust. Get a normal 10V gate drive MOSFET of about 2 milliohm or less on resistance driven by something like a MIC4422 mosfet driver which will actually be able to handle driving the large gate capacitance comfortably. Add a 15 zener between gate and source to protect the gate oxide from transients too.

As an example, here's a capable device: http://uk.farnell.com/international-rectifier/auirfs8409-7p/mosfet-40v-240a-q101-d2-pak-7/dp/2352041

(*) Maximum on current is related to the thickness of the conducting channel between source and drain, which depends on the charge on the gate-source capacitance, and thus the gate-source voltage - any logic level MOSFET will carry more current at 10V than 5V (and the on-resistance is lower)

#4

When I was talking about the mosfet being able to provide 100amp to the motor, I was thinking for only a fraction of time, when the motor is starting to move. There is no way the motor need 100 amps without any load.

Anyway, I found a place in my wiring that was getting hot, so I remove it and now the motor seems to go full speed according to the sound it's making. And I don't know how many amp there is because I don't have a clamp meter.

I still have I problem, when I send pwm signal to the mosfet, the mosfet is getting really hot rapidly and I'm pretty sure that the mosfet can support this heat even with a good heatsink.

MarkT, you talk about putting mosfet in parallel, can you say more about it? I don't know what is the pwm frequency of an arduino, but I'm guessing it's quite high and I though that slowing it down would reduce the heat. What are the most common ways to reduce the heat of a mosfet?

I'm also changing the title

#5

What's your battery situation - or are you using a power supply? The PWM out of 3, 9, 10, 11 is 490 Hz; from 5, 6 it's 976 Hz.

#6

Look at the graphs. At 50Amps, the voltage across the mosfet will be 0.5volt. That's 25watt. You need a reasonable heatsink to dissipate that. And when the mosfet gets hot, losses could be double of that. Leo..

#7

And that is only the static case. You will get extra heat dissipation in the time the FET transitions from on to off. It passes through a linear control zone that has very high dissipation briefly, this is sometimes called switching loss. You minimise this by driving the FET with a much lower impedance source than the Arduino. This is normally done with a FET driver chip which is normally another FET.

you talk about putting mosfet in parallel, can you say more about it?

Just connect the drain and source to the drain and source of another FET. It splits the current and so splits the heat dissipation. The gates too need to be paralleled up, but this increases the switching loss when driven directly from an Arduino so you could use a FET driver for the pair.

#8

I don't think switching losses are significant with Arduino's ~500herz PWM, as long as you keep the gate resistor ~220ohm.

@ Grumpy_Mike Don't know if it's easy to sync Arduino PWM pins. Then it would be possible to give each fet it's own drive pin. Leo..

#9

as long as you keep the gate resistor ~220ohm.

I would put the resistor at 125R, to maximum the drive current at just under 40mA ( 38.4mA )

Don't know if it's easy to sync Arduino PWM pins.

I would have thought that two PWM signals off the same timer would be in sync.

#10

I try to use a 110 ohm resistor to the gate but it still getting really hot to the touch in about 5 to 10 seconds. Without the pwm, the mosfet remain cool forever.

You are saying that switching the mosfet with an other fet can reduce the switching time? This sounds good, how are those deveice call (good old mosfet driver?)? Do you have any articla or IC example?

#11

Beurnii: how are those deveice call (good old mosfet driver?)? Do you have any articla or IC example?

You missed good old MarkT's good old reply (#2)

#12

[quote author=Runaway Pancake link=msg=2315267 date=1436840065] You missed good old MarkT's good old reply (#2)

[/quote]

Your are right

#13

Holy crap, the AUIRFS8409-7P Can carry up to 500amps and it cost 5$?!? And what is this atrocity , there is 5 pin for source, are they all the same thing?

MarkT, is there a similar version of this in a P channel? You think that the MIC4422 should be a good driver for those mosfet? I dont thing it's an H bridge driver, should one be better?

If this is all good, I think I'am set

#14

Yes, one big drain pad, 5 smaller legs for the same current to flow thru the source.

Biggest p-channel I see that is in stock at digikey is for 170A. http://www.digikey.com/product-search/en/discrete-semiconductor-products/fets-single/1376381?k=p-channel%20%20mosfet

#15

Hi,

Can you please post a copy of your circuit, in CAD or a picture of a hand drawn circuit in jpg, png or pdf? Can you please post a picture of your project. So we can see how you are powering up the motor and arduino and the hardware you are using.

Thanks...Tom..... :)

#16

Beurnii: Holy crap, the AUIRFS8409-7P Can carry up to 500amps and it cost 5$?!? And what is this atrocity , there is 5 pin for source, are they all the same thing?

MarkT, is there a similar version of this in a P channel? You think that the MIC4422 should be a good driver for those mosfet? I dont thing it's an H bridge driver, should one be better?

If this is all good, I think I'am set

No one uses p-channel for power, they perform 3 times worse due to the poor mobility of holes v. electrons in silicon.

There are 1000's of MOSFET drivers for all sorts of configurations. There are search tools at the major electronic suppliers' websites.

The AUIRFS8409-7P cannot carry 500A or anything like it. The package will fuse at 240A, and in practice you'll find a 100A is taxing that package. If you want 100A continuous use a decent sized package like ISOTOP.

#17

Yes, remember that you are limited by the lowest limit. It says "Silicon limited 522A" and "package limited 240A", later in the spec sheet it says that the bond wire limits it to 240A, as MarkT means when he says that it'll fuse at 240A. Also remember it has resistance and therefore IsquaredR power dissipation. Maximum resistance is 0.75mOhm at 10V Vgs. Derate by 2.5W/C.

Although it claims 1200A pulsed drain current, that is probably one non-repeating single pulse.

You've got to take into account not only the resistance and current, but the time it takes to switch on and off. That is influenced heavily by how you are driving it.

The maximum current drops to a fraction with 4.5V drive voltage and the voltage rises, therefore power dissipation rises. See the graphs on page 3.

#18

IR datasheets are laughable when it comes to current ratings. 270A for a TO220!

They seem to quote the theoretical max current not taking the bonding into account, ie the simulated silicon limit.

Anyway you never look at the current rating, the on-resistance is always the important value since you need to know IR and I-squared-R to select a device. You have to subtract IR from Vgs to get the actual Vgs, and I-squared-R tells you how much current.

#19

Ok thanks guys.

Assuming I will use those mosfet for high and low side of an hdridge, I need a driver to drive them. I don't really know how should I chose a mosfet driver. What should I look for? what number in the datasheet? What does effect the switching speed of a mosfet? I know I need a different driver for the high and low side.

The MIC4422 seems to be good for this job when i'm at the description and feature. Is there a high version of this driver? or even a half or full bridge? But I'm kinda scare of those hbridge driver, their input are complicated.

Thanks

#20

Is this for some sort of vehicle that us USUALLY going forwards, and SOMETIMES backwards. With some DEAD TIME in between.

Then you could use a reversing relay. Normally closed contacts for forward motion. Leo..