Hi-Low Side Driver Techniques (Without Driver IC)

Thread Starter

monk_house

Joined Nov 9, 2019
3
Good Evening.

I have constructed several H-Bridge (high-low) power switches before to control large DC motors in my robotics projects. The key downside to such simple bridges (featuring a high and low side MOSFET and a high side inverting NPN as a driver) is the speed at which I can drive them. I use relatively low speed PWM switching, which makes my motors noisy and ringy to the human ear. I am attempting to redesign my driver setup so that I can drive the FETS much faster (above 20khz) and get quieter motor operation.

Parts:

High Side P-Channel MOSFET: AOD403
Low Side N-Channel MOSFET: IRLR8726

( I know the AOD403 FET is obsolete, but it is super cheap and has a nice current rating. I will replace it eventually )

I use a C1815 as my high-side driver NPN.

I have attempted to create a NEW design capable of faster operation...

Proposed Design:

Proposed Bridge.jpg
By carefully selecting the base resistors of the NPNs, I can limit the CE current to within their ratings.
The current will still be MUCH higher than the 40mA my Microcontroller can supply, charging the gate capacitance faster and thus allowing for higher PWM frequencies. I added the 12v zener diode to prevent the High Side gate voltage from exceeding ratings. I plan on operating motors between 12 and 25 volts. Note that both Power MOSFETS are logic level.

Is there anything wrong with the above circuit? I have never seen NPNs used to pull up a gate that has been pulled down, but with fast switching I would need small pull up/down resistance, and an inverting setup would require dissipating power over this small resistor when idle. I would like to avoid gate driver ICs if possible...
 

crutschow

Joined Mar 14, 2008
23,542
Is there anything wrong with the above circuit?
A few things.
The P-MOSFET is upside down as well as the source and drain designations being reversed.
The P-MOSFET must have the gate at the source potential to turn off and 10V below the source to turn on.
The N-BJT must thus have its base go to the source voltage to do that.

Better to use the NPN to drive a PNP (PNP emitter connected to the P-MOSFET source.

Note that the AOD403 transistor has a high gate charge of 50nC nominal which will require a gate-source resistance of no more than about 250Ω to switch within a microsecond.
Some type of totem-pole or push-pull driver might be preferable to avoid the power dissipated by such a low resistance when the MOSFET is on.

What is the maximum operating current?
What is the maximum time the motors will be on (duty-cycle).
 
Last edited:

Alec_t

Joined Sep 17, 2013
10,470
I appreciate you don't want to use an IC as a driver, but unless you've got a bunch of BJTs you're trying to find a use for, why not use an IR2011 or similar as the FET driver? It's designed for the job and enables N-channel FETs to be used for both high-side and low-side switching.
 

Thread Starter

monk_house

Joined Nov 9, 2019
3
why not use an IR2011 or similar as the FET driver? It's designed for the job and enables N-channel FETs to be used for both high-side and low-side switching.
I have used IR2101 and 2011 before, and I had a good experience but they are relatively large and difficult to rout on a double-sided board smaller than a graham cracker (it’s a quad motor controller). The AOD403 has such a low on resistance that using a high-side n-channel merits apparently little benefit...
 
Top