M
Mook Johnson
I have a 3 phase PWM motor driver this is generating too much heat.
Motor phase current peak is 1.5A, Bus voltage is 500Vdc PWM freq is 10Khz.
This is for a sinusoidal drive so all Mosfets are switching all the time.
Variable speed induction motor.
What is the realistic efficiency I can achieve when delivering say 600 watts
through this drive to the motor.
I'm shooting for 98% efficient. That leaves 12 watts of heat dissipated by
the six mosfets. The current prototype is at 90% and gets really HOT (60
watts of heat) and requires large heatsinks and a fan.
The mosfets are http://www.st.com/stonline/books/pdf/docs/9572.pdf
So, Pdc ~= 1.5A^2 * .34 ohms Rds is .765 watts per FET X 6 = 4.59Watts
(worst case). The rest must be AC switching loss. If the Psw = the Pdc them
I'd reach my goal of dissipating less than 12 watts.
The Fets are currently driven through large pot core gate driver
transformers by 1.5A mosfet drivers. They are large dure to the relatively
low swithcing frequency. I'm thinking that the transformer leakage
inductance is not allowing fast enough switching speed and the switching
losses are dominating FET losses. (haven't had a change to put a scope to
the thing to see for sure)
I'm looking at using some high side drivers with floating supplies so I can
turn the gates on and off very fast.
Using this approach, or a better one, how efficienct can I reasonably get
this driver.
I realize the EMI will be greater if when the switching edges are faster but
I think we can handle this with proper enclosures and shielded wires.
what are your thoughts?
Motor phase current peak is 1.5A, Bus voltage is 500Vdc PWM freq is 10Khz.
This is for a sinusoidal drive so all Mosfets are switching all the time.
Variable speed induction motor.
What is the realistic efficiency I can achieve when delivering say 600 watts
through this drive to the motor.
I'm shooting for 98% efficient. That leaves 12 watts of heat dissipated by
the six mosfets. The current prototype is at 90% and gets really HOT (60
watts of heat) and requires large heatsinks and a fan.
The mosfets are http://www.st.com/stonline/books/pdf/docs/9572.pdf
So, Pdc ~= 1.5A^2 * .34 ohms Rds is .765 watts per FET X 6 = 4.59Watts
(worst case). The rest must be AC switching loss. If the Psw = the Pdc them
I'd reach my goal of dissipating less than 12 watts.
The Fets are currently driven through large pot core gate driver
transformers by 1.5A mosfet drivers. They are large dure to the relatively
low swithcing frequency. I'm thinking that the transformer leakage
inductance is not allowing fast enough switching speed and the switching
losses are dominating FET losses. (haven't had a change to put a scope to
the thing to see for sure)
I'm looking at using some high side drivers with floating supplies so I can
turn the gates on and off very fast.
Using this approach, or a better one, how efficienct can I reasonably get
this driver.
I realize the EMI will be greater if when the switching edges are faster but
I think we can handle this with proper enclosures and shielded wires.
what are your thoughts?
