Hello,

Iam designing a four quadrant DC drive. The problem i am facing here is, speed control of DC motor. Check the schematic.

Iam using the PWM technique to control the speed. The mosfet used is STW7NA100. I think the problem is with the snubber. I have selected the R & C values arbitrarily. A reverse connected diode (MUR860) is also used across drain and source.

Result: Three mosfets dead.

How to calculate the snubber values?

Motor parameters:
KW - 0.75
RPM - 1500
Volts - 220V DC
Amps - 5A

Excitation: 220V DC & 0.4A.


Switching frequency: 500Hz. Duty cycle [0 - 90%]

Experiment conducted:

1. Initially 220V DC is applied to field
2. Slowly armature voltage is increased upto 80V and maintained constant at 80V.
3. A voltmeter connected across drain and source reeds 80V (Vds = snubber voltage)
4. When the duty cycle is changed, Vds crossed 100V. and the armature voltage dropped to 40V. Here the mosfets are being dead.

see this videohttp://youtu.be/PIb6ZQdza1w

 

I found a pdf in this forum.

http://www.cde.com/tech/design.pdf

In page no.7, 1st paragraph

In my project, what should be the Eo, Io, Coss?


Next in page no. 10:
What is that Lp? Is it the motor armature inductance?

 

How is the power supply of the LM339 referenced to the rest of the circuit? In other words are you driving the MOSFET correctly?

 

It's hard to tell exactly what is going on in your video, but it looks like you have several devices mounted on your heat sink. I don't know what the other devices are or maybe I'm seeing it wrong. To drive the big motor that you are showing, you need a hefty heat sink for your mosfet and you probably also need a cooling fan as well. With the current that your motor is drawing, you will be dissipating a lot of heat in your mosfet.

 

How is the power supply of the LM339 referenced to the rest of the circuit? In other words are you driving the MOSFET correctly?

xr2206 and lm339 are mounted on a pcb with its own dedicated 12v power supply. The ground of this electronic part is connected to the source pin of mosfet. And i have tested the circuit for a 12v dc supp'y and a 12v dc motor. It gave a very good result.

 

It's hard to tell exactly what is going on in your video, but it looks like you have several devices mounted on your heat sink. I don't know what the other devices are or maybe I'm seeing it wrong. To drive the big motor that you are showing, you need a hefty heat sink for your mosfet and you probably also need a cooling fan as well. With the current that your motor is drawing, you will be dissipating a lot of heat in your mosfet.

first i used a small heat sink. Then i thought its not sufficient for the application. I have taken a little big heat sink from an old inverter. What you see on the heatsink are the inverter mosfets irfz44 which iam not using here. I have removed one mosfet from the heatsink and mounted my big mosfet. And the motor is rated 5 amps but it is running under no-load. So it cant take more than 1amp.

 

How is the motor terminal voltage (80V) measured? Oscilloscope or voltmeter?

80V with a voltmeter would mean it's the rms/peak value if we assume the signal to be a perfect square wave.

I would not measure a voltage level of a partially unknown waveform with a voltmeter, use always the oscilloscope.

1. start with a lower DC voltage and increase the duty cycle. Observe the waveform. Are there peaks building up that could be harmful to the MOSFET?

It's a 1000V MOSFET.

If you see the motor as an inductance consider the following:
- The moment the MOSFET turns off the inductance will create a negative voltage at the source pin, therefore increasing Vds.

That means you need a freewheeling diode at the motor terminals, even though that will need to be a quite powerful one.
Or you use a full bridge.

Maybe the motor specialists here can help you out.

 

How is the motor terminal voltage (80V) measured? Oscilloscope or voltmeter?

80V with a voltmeter would mean it's the rms/peak value if we assume the signal to be a perfect square wave.

I would not measure a voltage level of a partially unknown waveform with a voltmeter, use always the oscilloscope.

1. start with a lower DC voltage and increase the duty cycle. Observe the waveform. Are there peaks building up that could be harmful to the MOSFET?

OK. now i have connected a powerscope across drain and source. 90V spike observed at tailing edge. I did not connect any snubber in this case.

If you see the motor as an inductance consider the following:
- The moment the MOSFET turns off the inductance will create a negative voltage at the source pin, therefore increasing Vds.

That means you need a freewheeling diode at the motor terminals, even though that will need to be a quite powerful one.
Or you use a full bridge.

Ok. MUR890 is connected across armature terminals. Still the mosfet is burning. I have used a cooling fan also. I observed the heat sink is getting too much heat. I'll try with a larger heat sink