Issue with building a full-bridge Inverter

[uzair_r]

Hi,
I am building a Full-Bridge Inverter using four IGBT and four drivers for each IGBT. Driver is HCPL-3120, all four drivers are supplied with 15VDC isolated supply. I am switching using M bed micro controller LPC1768 generating two PWM signal working alternately, first signal turns IGBT 1 and 2 on and second turns IGBT 3 and 4. Inverter is supplied with roughly 36V while each driver with 15V. I have uploaded picture of schematic where four IGBT can be seen and one driver although in actual four drivers are used. After testing circuit i am unable to get a sine wave however I have upload snaps of my output. So what I am doing wrong here and what type of output i should be expecting from this circuit .
Any help here is much appreciated.

 

[Baluncore]

Welcome to PF.

I would have expected some form of LC filter on the output before any load to interpolate the sine wave and to remove the high frequency RF switching harmonics from the load.

What load are you using during testing?
Are those two traces the signals on the two load terminals?
Why do we see less than 2Vpp on the traces?
What is your ground reference connected to?
Oscilloscope coupling is AC. Why not DC couple with oscilloscope reference the (–)36V rail ?

 

[uzair_r]

Thanks,
Got it about LC filter, I tested this circuit without load as I am having some issues with calculations. Picture DSC_0142 was taken at output with reference to ground while other picture across two output terminal. Problem is I don't know why I am having output just of 2V as I should be having much higher output. I have also followed voltage in circuit and seems its the mid-point between four IGBT i am loosing voltage. If I got your question right on ground reference lower rail in my drawing is ground meaning IGBT 4 and 2 emitter legs are connected to ground. During testing I was not getting anything on DC couple on scope across output or with reference to -36 rail.

 

[Baluncore]

You are not getting an output switching between 0V and 36V, so I suspect you have a problem with your switch module design. I would suggest that you build and drive just one HCPL3120 and IGBT with a resistive load from the positive supply rail. A 100 ohm load would draw 360mA and need to be rated at better than 13W. You might consider using four 12V light globes in series as the load. Once that is switching cleanly you can complete the bridge.

Check your 36V supply is still OK. If you have turned on both IGBT pairs at the same time you may have caused some damage to the 36V supply fuse.

 

[Windadct]

The HCPL3120 is not a bootstrap driver to make this you need 3 isolated 15V DC supplies for IGBTs 1,3 & 4+2. So this will not work. You would only be turning on IGBT 4 and 2. Then with no load -- the output is practically just the bottom drivers output V - since the scope is AC coupled, I suspect it is really 0 to +4.5V .

I agree with Baluncore -- make a DC Buck chopper and then understand it. With an isolated DC supply you could do this in ether High or Bottom IGBTs in a leg - and use the Freewheeling diode of one of the other IGBT (use a jumper in this IGBT between the Gate and Emitter- to make sure it is OFF) -- the Isolated DC supply will turn on a TOP IGBT - but not a top and a bottom.

 

[uzair_r]

This morning I changed HCPL3120 supply to 3 isolated 15V DC as you said, at a point i got a signal and I have upload it. Your feedback on that will be very helpful. Like you said I will put a load with filter and test one leg, however I am planning to integrate inverter with a DC-DC booster where its output of 36V will go to inverter.

 

[Windadct]

It would be good to look at the Gate Signals - you should see the PWM signal clearly. It may be that you killed a component somewhere. --

 

[Baluncore]

This morning I changed HCPL3120 supply to 3 isolated 15V DC as you said, at a point i got a signal and I have upload it.

When you attach a trace, we need to know which signal you are viewing and the load conditions.
We also need to be able to read the vertical voltage scale, or you need to report the Vpp of the observed waveform.