Hi,
I am running 2X 24V 180 Watt PMDC motors off of a 24V battery . The
application is a powered wheelchair with differential steering. 1:24 gearbox.
max load 120Kg. Using the OSMC
The motors were stalled at 12V the stall current was observed to be 6A. Hence at 24V, the stall current should be 12A or thereabouts, well within the Current limit of the current limits of IRF1405
Here's a youtube video of the wheelchair running on 24V 10A supply
http://www.youtube.com/watch?v=YiFXiUZK-gQ
open loop Control scheme: analog 2 pot joystick-ADC-Atmega128-PWM(15.625KHz)-ALI/BLI. BHI/AHI Hi/Lo
I am using 1xIRF1405 per leg.
The boards failed. The following were the observations;
1.Boards worked fine for long periods (24 hours discontinuous) on a controlled
24v 10A power supply with and without load upto stall
2. With battery power, no load test was not done
With user seated (load condition) on battery power, the wheels appeared
stalled when I pushed joystick backwards and did not turn. When joystick was
released. nothing happened. When joystick was again pushed backwards, catastrophic failure of mosfets was observed
3. The mosfets failed thus;
board 1. Q9: Source-Drain shorted but no visible damage. Q1: catastrophic failure with fire and smoke
board 2. Q16: Source-Drain Shorted but no visible damage. Q8: Catastrophic Failure with fire and smoke
4. Wheelchair was mounted on jack after board failure and tested. The other
legs appeared fine, motors turned fine. The HIP was hot
5. Each component in the power section was removed from board and checked. Except for the mosfets, everything appeared to be OK
Working Hypotheses.
1. Motors were stalled (for some reason) when the joystick was moved backwards.
When the joystick was released, the mosfets were off.
The inductive kickback flowed
through the TVS diode on one leg going towards ground, but on the return journey to the motor, it flowed through the forward biased body diode on the other leg (being a lower resistance pathway), exceeded the power rating of the body diode thus destroying it and showing a drain-source short without any visual damage.
When the joystick was then pushed backward, shoot through destroyed the upper mosfet of the same leg.
only problem is, why didn't the same thing happen to the other leg?
SOLUTION: Use TVS diodes with lower voltage rating (1.5ke33CA) on lower fets and for good measure use them across the upper mosfets as well. Use a forward biased diode in series with drain of the lower mosfets, to avoid any possibility of current flow through the body diode. In addition, use a TVS diode across the motor terminals.
2. The Duty Cycle/Bootstrap voltage/PWM frequency is too high/too low. Thus
causing the lower mosfets to jam open (Drain-Source Short) Thus causing the
shoot through. Thus destroying upper mosfets.
Only problem is why did it work fine on controlled power supply
SOLUTION: Change Duty Cycle/Bootstrap V/PWM frequency
Please tell me if I'm on the right track.
Thanks
Sunil
I am running 2X 24V 180 Watt PMDC motors off of a 24V battery . The
application is a powered wheelchair with differential steering. 1:24 gearbox.
max load 120Kg. Using the OSMC
The motors were stalled at 12V the stall current was observed to be 6A. Hence at 24V, the stall current should be 12A or thereabouts, well within the Current limit of the current limits of IRF1405
Here's a youtube video of the wheelchair running on 24V 10A supply
http://www.youtube.com/watch?v=YiFXiUZK-gQ
open loop Control scheme: analog 2 pot joystick-ADC-Atmega128-PWM(15.625KHz)-ALI/BLI. BHI/AHI Hi/Lo
I am using 1xIRF1405 per leg.
The boards failed. The following were the observations;
1.Boards worked fine for long periods (24 hours discontinuous) on a controlled
24v 10A power supply with and without load upto stall
2. With battery power, no load test was not done
With user seated (load condition) on battery power, the wheels appeared
stalled when I pushed joystick backwards and did not turn. When joystick was
released. nothing happened. When joystick was again pushed backwards, catastrophic failure of mosfets was observed
3. The mosfets failed thus;
board 1. Q9: Source-Drain shorted but no visible damage. Q1: catastrophic failure with fire and smoke
board 2. Q16: Source-Drain Shorted but no visible damage. Q8: Catastrophic Failure with fire and smoke
4. Wheelchair was mounted on jack after board failure and tested. The other
legs appeared fine, motors turned fine. The HIP was hot
5. Each component in the power section was removed from board and checked. Except for the mosfets, everything appeared to be OK
Working Hypotheses.
1. Motors were stalled (for some reason) when the joystick was moved backwards.
When the joystick was released, the mosfets were off.
The inductive kickback flowed
through the TVS diode on one leg going towards ground, but on the return journey to the motor, it flowed through the forward biased body diode on the other leg (being a lower resistance pathway), exceeded the power rating of the body diode thus destroying it and showing a drain-source short without any visual damage.
When the joystick was then pushed backward, shoot through destroyed the upper mosfet of the same leg.
only problem is, why didn't the same thing happen to the other leg?
SOLUTION: Use TVS diodes with lower voltage rating (1.5ke33CA) on lower fets and for good measure use them across the upper mosfets as well. Use a forward biased diode in series with drain of the lower mosfets, to avoid any possibility of current flow through the body diode. In addition, use a TVS diode across the motor terminals.
2. The Duty Cycle/Bootstrap voltage/PWM frequency is too high/too low. Thus
causing the lower mosfets to jam open (Drain-Source Short) Thus causing the
shoot through. Thus destroying upper mosfets.
Only problem is why did it work fine on controlled power supply
SOLUTION: Change Duty Cycle/Bootstrap V/PWM frequency
Please tell me if I'm on the right track.
Thanks
Sunil
