UCC37322 - Am I On The Right Track?

The unit I'm working on uses a 20kHz PWM signal (from a PIC - max supply 20mA) to drive a DC load. Input voltage could be either 12V or 24V, and the nature of the load to be driven is unknown - could be inductive or resistive. I'm using an Allegro AC8756 current sensor to monitor current and limit it (via duty cycle) to 30A for now.

The MOSFET I have been considering and playing with is the IRF1404ZGPBF. It's rated to 75A continuous which is more than enough, and has a pretty low on resistance of about 3 mΩ (at 75A). This is probably overkill, but I would like the capability to possibly run these same parts in a 50 or 60A configuration down the road.

I think I want to use the TI UCC37322 to drive this MOSFET since it can use my small PWM output signal and has a 9A max drive output which should be plenty of juice to turn the MOSFET on fast enough for my frequency range - even at 60A. The 37322, MOSFET, and necessary gate resistors will be immediately adjascent on the board and the MOSFET will be on a heat sink.

My concern is the continuous current output of the UCC37322 which is listed as 600mA. I want to be able to run this at a "true" 100% duty cycle which could conceivably see 30A running continuously from D to S. I have been unable to determine whether 600mA is enough to keep this MOSFET fully ON for an extended period of time. I suspect the answer is in the data sheet, but I'm unable to decipher it at this point and could use some direction.

I have been told on this forum by members whom I believe and respect that you typically want your gate current 1/10th of your drain current which would say I need 3A. I do not know if that rule-of-thumb was more for turning on fast or for keeping it on.

So, my questions:
1) Will the UCC37322 be able to keep up at 30A @ 100% duty cycle?
2) Same as #3, but at 60A?
3) Is this a robust combination of components for this type of load, or would you choose something(s) differently?

I appreciate any help or guidance - I'm sure the answers to 1 and 2 are in the documents I've been through, I'm just not seeing them. The answer to #3 is in wisdom that I do not yet possess.

You are confusing BJT's and MOSFETs. BJT's need a base current of 1/10 of the collector current but MOSFETs have insignificant DC gate current. The driver is only needed to rapidly charge and discharge the large gate capacitance of MOSFETs. There will be no problem driving the MOSFET at DC (100% duty cycle).

Awesome, thank you - I had forgotten the rule of thumb mentioned was with respect to a BJT, and that explains why I could not find much information about calculating gate input current for MOSFET's!

crutschow said:

You are confusing BJT's and MOSFETs. BJT's need a base current of 1/10 of the collector current but MOSFETs have insignificant DC gate current. The driver is only needed to rapidly charge and discharge the large gate capacitance of MOSFETs. There will be no problem driving the MOSFET at DC (100% duty cycle).

Click to expand...

Couple of things to watch out for.
Even though your mosfet is rated at 75a, that is steady state (and in lab conditions), in a switching application the mosfet will spend some time in the linear state while it is switching, using the driver reduces this however you need ta take that into account when working out dissipation.
Another one I found when using mosfet drivers was that the waveform from the output of the fet was so square I was getting loads of harmonics and rf interference.

Interesting. I think I will be okay on heat dissipation as the heatsink is large, but thank you for the information.

On your comment about the wave being "so square," does that mean that the dV/dt is causing the harmonics and rf interference? If so, in my application I think I could actually slow it down some: I calculated that I need 2A to get a 50nS rise time. I think I could slow it down - at least on the rise - without any real impact on the functionality of my output.

I'll do some research on harmonics and RF and see what I should be looking at. Thanks again for the info.

MB

dr pepper said:

Couple of things to watch out for.
Even though your mosfet is rated at 75a, that is steady state (and in lab conditions), in a switching application the mosfet will spend some time in the linear state while it is switching, using the driver reduces this however you need ta take that into account when working out dissipation.
Another one I found when using mosfet drivers was that the waveform from the output of the fet was so square I was getting loads of harmonics and rf interference.

Click to expand...

MikahB said:

On your comment about the wave being "so square," does that mean that the dV/dt is causing the harmonics and rf interference? If so, in my application I think I could actually slow it down some: I calculated that I need 2A to get a 50nS rise time. I think I could slow it down - at least on the rise - without any real impact on the functionality of my output.
.............

Click to expand...

Remember that slowing the switching time increases the transistor power dissipation and reduces efficiency.

Proper shielding and filtering should keep the EMI to an acceptable level, if it is a problem. It helps to keep all interconnecting leads as short as possible.

Mika, yes your assumption is correct, that is what I meant.

Excuse my terminology, I'm hands on not so much textbook.

@dr pepper: no excuse needed for your terminology! I'm just trying to make sure I understand what's what and I've found that me not asking often equals me misunderstanding. And just for the record, I am all about hands-on other than the (seemingly endless) hours I spend reading datasheets and white papers trying to understand this stuff. I'm a mechanical engineer, so this is way outside my wheel house!

@crutschow: I will proceed with working towards making my first functional prototypes and hopefully that will help me determine if EMI is a problem or not. My unit has two boards, a low-power (5V) control board and the high power board being discussed here. Both boards are about 2" x 2.5" so keeping leads short is relatively easy. I think I came up with a pretty slick way to mount my 5V regulator and MOSFET to the upper board that makes everything very compact and tidy and lets them connect to a common heat sink. So, I shall plug on ahead and see how it goes.

Thanks again. Now time for me to make a new post trying to understand my circuit protection scheme!