Need Help Designing a Full Bridge IGBT Inverter Driver Board

Thread Starter

Fried Circuit

Joined Mar 22, 2017
4
I have been working on this crazy idea I have since I was around 12 years old. (Now I'm 60) My problem is I do not know how to calculate all the variables of the components to work together to make it happen. I keep burning up parts.

My goal at this level of the design is to create a circuit to drive a h-bridge as an inverter.

My plan so far is to use:

TL494 KA7500 Driver Board powered by a 12v battery to generate a signal to the optocouplers.
HCPL-3120 optocouplers
I am getting my isolated power for the IGBT's from TRACO POWER 1MH1215D
Attached you will also see the H-Bridge configuration I am trying for.
I am using IRG4PC50S standard speed IGBT's that I will send 120 volts through.
I am using a 104 capacitor
I am using a 7 ohm gate resistor

I am having trouble answering these questions myself;

TL494 KA7500 Driver Board
1. I am using the signal from PWM1 and PWM2 in an anti-parallel configuration as you see it in attached IGBT board diagram. Is this the correct way to do this?
2. It is my understanding that I can connect the Feed + and the Feed - feet through a potentometer to change the voltage of PWM1 and PWM2. Or I can just attach Feed - to the ground for max voltage. This does not seem to work for me.
3. I am using the same 12v source to power the TL494 KA7500 Driver Board and the (4) TRACO POWER 1MH1215Ds. Is that a problem?

HCPL-3120 optocouplers
1. I seem to be frying these too. Am I sending too much current to the LED? What resister would I need? Would I need one resistor on each of PWM1 and PWM2 because of their anti-parallel config?
2. Right now, it will fire the optocouplers, but it will only give me 5 to 6 volts at the IGBT gate when it switches on the isolated power module. (There is 15v available) This does not seem to be enough voltage to fire the IGBTs. Could that be because the LED is not getting enough voltage? Or is it because my Isolated power units are not giving enough power and are exhibiting a power drop for that chosen IGBT?
3. When I use a AC voltage measurement across the PWM 1 and PWM 2 when it is hooked up to the 4 optocouplers,, it shows a .6 volts. When I use a DC voltage measurement across the PWM 1 and PWM 2, it shows as 2.5 volts. I am out of my league of understanding to know if this is right.

Thank You for your help to make this come alive!
 

Attachments

ronv

Joined Nov 12, 2008
3,770
I have been working on this crazy idea I have since I was around 12 years old. (Now I'm 60) My problem is I do not know how to calculate all the variables of the components to work together to make it happen. I keep burning up parts.

My goal at this level of the design is to create a circuit to drive a h-bridge as an inverter.

My plan so far is to use:

TL494 KA7500 Driver Board powered by a 12v battery to generate a signal to the optocouplers.
HCPL-3120 optocouplers
I am getting my isolated power for the IGBT's from TRACO POWER 1MH1215D
Attached you will also see the H-Bridge configuration I am trying for.
I am using IRG4PC50S standard speed IGBT's that I will send 120 volts through.
I am using a 104 capacitor
I am using a 7 ohm gate resistor

I am having trouble answering these questions myself;

TL494 KA7500 Driver Board
1. I am using the signal from PWM1 and PWM2 in an anti-parallel configuration as you see it in attached IGBT board diagram. Is this the correct way to do this?
2. It is my understanding that I can connect the Feed + and the Feed - feet through a potentometer to change the voltage of PWM1 and PWM2. Or I can just attach Feed - to the ground for max voltage. This does not seem to work for me.
3. I am using the same 12v source to power the TL494 KA7500 Driver Board and the (4) TRACO POWER 1MH1215Ds. Is that a problem?

HCPL-3120 optocouplers
1. I seem to be frying these too. Am I sending too much current to the LED? What resister would I need? Would I need one resistor on each of PWM1 and PWM2 because of their anti-parallel config?
2. Right now, it will fire the optocouplers, but it will only give me 5 to 6 volts at the IGBT gate when it switches on the isolated power module. (There is 15v available) This does not seem to be enough voltage to fire the IGBTs. Could that be because the LED is not getting enough voltage? Or is it because my Isolated power units are not giving enough power and are exhibiting a power drop for that chosen IGBT?
3. When I use a AC voltage measurement across the PWM 1 and PWM 2 when it is hooked up to the 4 optocouplers,, it shows a .6 volts. When I use a DC voltage measurement across the PWM 1 and PWM 2, it shows as 2.5 volts. I am out of my league of understanding to know if this is right.

Thank You for your help to make this come alive!
Just some ideas....
It looks like the optos want about 10 milliamps. If tour driver is running at 12 volts the resistor would be about 1K ohm.
When you switch the IGBT there needs to be some deadtime so the top and bottom IGBT on the same side of the H won't be on at the same time. It takes about a Usec, for the IGBT to turn off.
You probably need clamp diodes as well.
This may not be fun without a scope.:(
 

Thread Starter

Fried Circuit

Joined Mar 22, 2017
4
I just bought a PC scope. It is a Hantek 6022BE.

The signal output is not even close to 12v. As mentioned above, it is at the best .6 volts. will have to put it on the scope to see what I get.

Where would I put the clamp diodes? That is the diode across the collector and the Emitter?

How would I add some dead time with this PWM board?
 
Last edited:

crutschow

Joined Mar 14, 2008
24,705
How would I add some dead time with this PWM board?
Below is the LTspice simulation of a non-overlap circuit that gives dead-time between the two outputs of ≈0.85*R3C1 (or R4C2), so you can adjust the values of R3C1 and R4C2 (both the same values) to give the desired dead-time.
The values shown give a dead-time of 1.68μs.

It has three CD4050 buffers in parallel at each output to improve the capacitance drive capability.

Power and ground pins aren't shown for the IC's but must be connected.
Any unused inputs must be connected to ground.

upload_2017-3-22_22-49-32.png
 
Last edited:

Thread Starter

Fried Circuit

Joined Mar 22, 2017
4
Wow! Thank you for all the help. I just put it on my new scope and it seems that dead time is already in the chip.

Also, how I had it hooked to my optocouplers was an anti-parallel configuration. That only gave me at the most 5v at each signal. When I hooked the scope probes to to both channels and then grounded the probes to the input ground, it came alive with 10 volts at PWM1 and PWM2
pwm1 and pwm2 with ground connected.png

Correct me if I am wrong, but would that not still need only 1000 ma resistor?
http://www.rapidtables.com/calc/electric/ohms-law-calculator.htm

And I would only need (1) 1000 ma resistor for both channels put on the common neg ground??
 

Thread Starter

Fried Circuit

Joined Mar 22, 2017
4
CRUTSCHOW, I appreciated your answer back. I gleaned from your comment that I probably need to make sure the chip outputs are grounded properly.

Thanks
 
Top