Hi everyone,
I'm testing out a BTS3080 MOSFET [datasheet] on a breadboard and am experiencing some heat. I had previously read that driving MOSFETs with PWM from a microcontroller should result in little to no heat.
Currently, I'm driving it with 3.3V (which is on the lower side of VinNOR, 3V-5.5V) PWM set at 500Hz. The load current is maximum 1.6A (well below the 3A rating) and is resistive. I would ideally like to drive about 2.2A. I do not know if the frequency is ideal, as the datasheet has no mention.
I am fully aware that soldering legs onto this MOSFET to make it breadboardable is detrimental to heat dissipation, but I am trying to create a worst case scenario before putting this on a PCB. That being said, the thermal shutoff does not engage and the heat is "uncomfortable to the touch", but not melting anything (thermal shutoff would kick in sooner). I may try adding a temperature probe onto it.
Is this to be expected or could this be due to driving at the low-end of VinNOR? I suspect I will end up driving this with a 5V PWM output anyways, but do not have a way to try that out right now (no MCU with 5V logic).

 

As I read the datasheet you are barely turning the device on at 3.3V. If the rds(on) is 160 mΩ, then with 2.2 A of current through the channel, you will dissipate ≈ 0.775 watts. That's enough to make a device on a breadboard without a heatsink get warm. If I were you, I would employ a 3.3V to 5V level shifter to make sure the device is turned on hard.

 

Thanks for confirming. I'll try it out.

 

Also make sure Rin is small enough to reduce any voltage drop. Device only takes about 0.1ma gate input under normal use, so you might not even need that resistor. However, it might protect the I/O pin if something goes wrong with the device.

 

Reduce the resistance value of the Rin resistor and increase the gate current of the MOS transistor, so that it can conduct quickly and reduce conduction loss (heat).

 

The Mosfet resistance is 0.16 ohms when the gate-source is 5V and the resistance is much higher when the gate gets only 3.3V.
The gate has a high capacitance that draws a high current each time it switches so a high input driving resistance and a high PWM frequency causes a high input current and slow switching. The slow switching allows the Mosfet to be at a higher than switched resistance causing high heating.

 

That chip has a built-in gate drive and other internal systems, so I'm not too sure treating it like a discrete transistor is such a good thing.

I definitely wouldn't sweat the 3.3 volt drive voltage, but I would make sure it's actually 3.3.

Normally a chip like that would use the PC board as the heat sink, considering that the thermal protection hasn't kicked in with the chip not mounted, I would think it will probably be fine when properly mounted.

And 500Hz is in no way a PWM frequency to be concerned about.

Figure 5.1 shows that the resistance doesn't reach its maximum until over 140c at 3 volts.

 

Also consider the temperature that the body senses as pain, is far below the temperature that chip can handle.

 

The chip is smart and shuts down when it gets too hot but it does not say why.

 

Back when logic level MOSFETs were new I used some in a motor drive circuit. There was a problem with the motor that caused the MOSFETs to get so hot they unsoldered themselves and fell off the PCB. I soldered them back in and when the problem with the motor was corrected they ran it without a problem.

 

I didn't look up the part, but generally speaking you want your PWM signal to be a really sharp square wave, so the transistor spends as little time as possible partially-on. If the signal is more of a saw tooth, sine or rounded square wave then it's going to spend more time in the active region, which is where it's going to generate the most heat. You want the transistor either fully-on or fully-off as quickly as possible.

 

What is the Battery-Voltage ?
.
.
.

 

Reduce the resistance value of the Rin resistor and increase the gate current of the MOS transistor, so that it can conduct quickly and reduce conduction loss (heat).

I'm not using Rin as the datasheet recommends it for inductive loads.

I didn't look up the part, but generally speaking you want your PWM signal to be a really sharp square wave, so the transistor spends as little time as possible partially-on. If the signal is more of a saw tooth, sine or rounded square wave then it's going to spend more time in the active region, which is where it's going to generate the most heat. You want the transistor either fully-on or fully-off as quickly as possible.

I'll check the signal on my scope.

What is the Battery-Voltage ?

12V

Thanks everyone for your input.

 

Use a FET-Driver, extremely easy to implement.
2.5V for turn on, 0.3V Hysteresis, CMOS Input.
0.1-Ohm On-Resistance.
SMD Package, or TO-220 Package.
No other Components needed.
.
.
.

 

The datasheet shows that it is much more than a Mosfet and it even has a gate driving unit in it so it doesn't need another one:

 

Not "another one" ..........
JUST the FET-Driver.
.
.
.

 

Where does that chip get its VDD for those internals?

The load?

Edit:

Oh yea, I see from the image in the OP that it is the load.

I don't know, sounds like that could be problematic in some cases.

 

Not "another one" ..........
JUST the FET-Driver.

Apart from being a smaller package (and logically worse heat dissipation), the Smart MOSFET is in my opinion superior. I like the smart functions, especially the over-voltage protection.
I'm pretty sure that with some optimization (software, pcb sink), I should be able to cut heat to a minimum without brute force (larger package or additional heatsink).

Also, as @Audioguru again mentioned, that driver seems to work in symbiosis with (to drive) MOSFETs, not standalone, unless I'm misderstanding something.