BLDC driver, sanity check

Hi, I’m trying to design a 3-phase driver for brushless dc motors. I have no formal training in electronics, and I would like to have some experienced eyes to check my schematic for faults leading to malfunctions or a destroyed Arduino Due.

The motor I would like to experiment with is a car alternator with the diodes removed. I already have the motor running on a sensorless ESC, but i would like to experiment with sensored control, PID control, varying rotor supply etc. That’s why I’m trying to design my own controller instead of going for a commercial product. I’m planning to run the “motor” at up to 48V and 100A.

The controller is designed as a shield for Arduino Due. The 3 half bridges will be on a separate module along with a current sensor.

The three IR2183’s is for driving the half bridges. The IR4427 is for driving a mosfet for rotor-supply, and an extra mosfet for miscellaneous. The LM311 is for triggering an interrupt in case of overcurrent.

My concerns at the moment is:
Can i share the output from a single current sensor between the LM311 and an analog input?
Have I done enough for protection against a potential noisy environment, due to the high power being switched?
I have a hard time understanding the bootstrap circuit. Do i need to pulse the LO side drivers prior to motor start, or do i need extra circuitry, or am I ok?

Thank You, Peter

Datasheets:
IR2183: http://www.irf.com/product-info/datasheets/data/ir2183.pdf
IR4427: http://www.irf.com/product-info/datasheets/data/ir4426.pdf
LM311: http://www.ti.com/lit/ds/symlink/lm211.pdf
Current sensor: http://www.allegromicro.com/~/media/Files/Datasheets/ACS758-Datasheet.ashx
Hall Sensors: Error - Melexis

Oops, sorry about the large picture.

Yes the current signal can go into a comparator and an analog input simultaneously,
no problem.

If the high power stuff is opto-isolated you'll sleep easier, since a high power fault
cannot hit the microcontroller. I assume current sense is hall-effect isolated sensors?

Low-side shunt current sensing is pretty safe too.

The bootstrap circuit requires the low-side driver to be on every few ms or the
high side driver will lose power. Thus if you PWM don't go all the way to 100%, so
there is always low drive. Park the bridge with low-side FETs on, high-side off, and
its ready to start.

The bootstrap circuit charges the 12V cap for the high side driver whenever the source
of the high-side FET is shorted to ground (because the low-side driver is on, and high-side
off) Once that arm of the bridge goes high the cap starts to discharge - you only have a few ms typically before you have to restore it.

Make sure the driver chip decoupling caps are about 20x the bootstrap caps and the bootstrap
caps are about 10x the MOSFET effective gate capacitance or more (or put another way
convert nC of total gate charge into nF of bootstrap cap - then the bootstrap cap only
drops 1V or less at the point of switch-on (ie from 11.5V down to 10.5V). The bootstrap
diode needs to be fast and have a high enough voltage rating for the motor supply.