I use a Microship TC1411N (non-inverting) MOSFET driver to power an N-channel MOSFET. I used to have IRF520 MOSFETs, but ran out of them and instead switched over to IRF644s. Since then my MOSFET drivers have been burning up left-and-right.

I am not too familiar with capacitance theory for MOSFETs, but would this be the likely culprit?

The IRF644 has a typical input capacitance of ~1300pF. The IRF520 is rated at 360pF. The TC1411N is supposed to only handle 1000pF gate capacitance.

Is this a reason why my the drivers are burning up (literally, they get up to 100C!)? Thanks everyone.

 

Yes. Use gate resistors to reduce the gate charge current, which will in turn increase the transition time, or use a higher current gate driver.

Personally, I ignore gate capacitance and pay attention to gate charge instead. 22nC vs 68nC. It seems that capacitance can be more misleading and varying than gate charge.

__________________
Tanaka Sensei (avatar) says: Please spell it "ridiculous" correctly! Not "rediculous". ^^

 

Thanks for the info! I will probably switch back to the IRF520s which seemed to have been well matched.

In the mean time, is there a way to figure out what value gate resistor I need?

 

The value you want depends on how fast you want the transistor to turn on or the gate current you actually want to pull from the gate driver.

I use

Q = It
V=IR

where Q is the total gate charge, I is the driver current, and t is the turn on/off time. V is the gate driver supply (or the peak voltage on the output of the gate driver if you want to account for the driver's losses) and R is the base resistance.

__________________
Tanaka Sensei (avatar) says: Please spell it "ridiculous" correctly! Not "rediculous". ^^

 

Thank you so much again, I feel bad asking for all this help knowing I can't help people as much. But I am learning (it's just a hobby I've had since I was a young kid).

I will see if this works, hope that helps out.

Now here's another thought. The MOSFET driver is being controlled by a PWM output of an AVR (ATTiny13). I used to have ATTiny12s where I "faked" the PWM by turning on and off one of the outputs using a pair of for loops and it worked great. Now I have the Tiny13s which use real PWM. What's interesting is that I can hear humming in near by speakers (on the same power rail) when the PWM is working (the hum changes depending on the duty cycle).

Perhaps it's switching too fast for the MOSFET driver?

 

Faster, as in higher frequency? Or faster falling/rising edges?

Higher frequency requires faster falling/rising edges because higher frequency means more transitions, which means more time spent transitioning, which means more losses (heating) from spending time in the lossy transition region. Without changing the frequency (and thus the number of transitions), the only thing you can do is to make each transition faster. Faster falling/rising edges means more noise.

If the driver cannot keep up, then the rising/falling edges can't be produced as quickly as you need them to be. Which means less interference, not more. However, there is such a thing as turning the MOSFET on too fast causing unecessary nois.- that's why gate resistors are used to control the rise/fall times for a driver that is oversized for the MOSFET. So even if you reduced the PWM frequency, you wouldn't reduce the primary source of noise since the noise is not coming from the PWM frequency as much as it is coming from the speed of the rising/falling edges.

__________________
Tanaka Sensei (avatar) says: Please spell it "ridiculous" correctly! Not "rediculous". ^^

 

I'm thinking higher frequency. I ran the Tiny12 at 1Mhz, the Tiny13 runs at full speed (~4Mhz). Although I am not sure if this directly affects the PWM clock.

 

Your circuit is badly shielded. Sharp square waves create harmonics that can ouput AM frequency well above the base you're driving it at. Obviously your new fets are MUCH better at sharp on and off times from your given drive signal.

__________________

 

It does. PWM settings run directly off of the clock and PWM frequency will vary proportionally as you change the clock frequency if you do not change the configuration.

__________________
Tanaka Sensei (avatar) says: Please spell it "ridiculous" correctly! Not "rediculous". ^^