Connecting converter to the inverter

[sam781]

Yes, problem solved after connecting a resistor as load. Thanks...

 

[sam781]

I've made all there bridge and filter added to them. I checked them separately and found around 18v which is alright to drive MOSFET.

Now, some questions have come to mind regarding the h bridge converter circuit given in post #7.

1. -ve terminal of lowest coil's bridge and the -ve terminal of 326v will always be connected together. Will it be any issue?
2. Should not be any issue form my idea. But is is necessary to check the polarity of the coil and connect to -ve terminal of 236v? As bridge rectifier has been used, -ve of bridge is always -ve so coil polarity should not be any matter. This is my understanding.
3. When high side MOSFET will be on. -ve terminal of upper coils' bridge will be at 326v. Will it be any issue? Should not be any issue in may sense as as like point#3.

- - - Updated - - -

Correction -
I've made all three bridges and filter circuit also added.

 

[BradtheRad]

It should be okay for you to make the connections you describe. It can work because the transformer secondaries themselves are not connected to each other.

 

[sam781]

It should be okay for you to make the connections you describe. It can work because the transformer secondaries themselves are not connected to each other.

Okay, I'll do it now...

- - - Updated - - -

Doesn't work.
Should I use high voltage (500-600v) diode? I've used low voltage (less than 100v) diode.

 

[BradtheRad]

I see. None of your phototransistors have a definite path to ground. No current can flow. As a result the mosfets do not turn on consistently.

It will require additional devices, to apply proper bias to the mosfets.

Also it improves reliability if you turn off the mosfets by pulling their gates to ground with a 10k resistor.

The high side mosfets will also be a problem. It will make it easier if you choose P-devices.

 

[sam781]

All the MOSFETs' Gate and sources are connected to each 18v volt's +ve and -ve terminal. MOSFETs are turning on. I've checked with a separate 12v battery.Battery connected to high side MOSFETs' drain and low side MOSFETs source by disconnecting filtered output of 326v. It gives output. When I connect 326v, no output is obtained at the output terminal; after few seconds a sound came (not sure from which parts it came, probably from the converter transformer). I immediately disconnected the power supply.
I thought I may be due to low rated voltage of rectifier diode.

 

[BradtheRad]

The only way to be sure about the diodes, is to disconnect each diode and check whether it still passes current in one direction and blocks it in the other direction.

after few seconds a sound came

That's one indicator something is not right.

If you were lucky, you disconnected power before a component got fried.

You are working with high voltage. It is easy to fry a component before you even know which one was overheating.

It will be best if you were to reduce the voltages which you are working with. Can you apply 6V as the power source? Then you have time to touch components to find out if any are overheating.

Additionally, trying slowing down the frequency. You need to verify that your H-bridge timing signals are correct. You probably know that shoot-through is a common hazard. Your mosfets must not allow shoot-through from one supply rail to the other. They will burn up. Your load will not get voltage.

 

[kappa_am]

disconnect FET bridge. use your designed driver to drive single high-side and low-side FET ( do not connect as a single leg) just 2 separate load. if there is no problem, you have a mistake in power section. look at your FET and opto datasheets and estimate delay time. you have to consider a dead-time according to difference of rise time and fall time.
report the result to discuss other possible errors.

Good luck

 

[sam781]

I've checked with 12v battery as inverter input; it worked but the output waveform was not as expected. The waveform was 50Hz square (almost) wave. As per my understanding the output should be as like as below figure (as I did not use filter) -


I've checked the waveform at the 6 optocouplers output separately and found the expected waveform.
Then, connect the opto output to the MOSFET's Gate and source and see almost smiler waveform.
Then, I gave power to the bridge from a separate 12v battery. Got square wave output.
Then, I connect 326v dc from converter output. Same issue... Some sound stared from the converter transformer. :sad:

 

[BradtheRad]

You are jumping directly from a low voltage test, to a high voltage test.

I still suggest that you try a less drastic jump in the applied voltage. Why not increase it by increments? That way you might get an opportunity to see the problem develop. Or get a better idea from noises, etc.

Another thing to try:
Your project contains several sections. It may be a good idea to test sections individually, when you need to isolate a problem.

 

[kappa_am]

is it possible to show me the DC-link current?

 

[sam781]

is it possible to show me the DC-link current?

How can I do that?

 

[kappa_am]

Do you have current transformer of isolated current measuring ICs?
if not you can use a small resistor in series between dc link and your bridge then measure it voltage.

 

[sam781]

After some modification and tests, I've got the below results -

1. To test my inverter at a low voltage, I made another secondary winding with 16/17 turns which gives 45v dc after rectification and filtration (LC). Then I got the following output -

It seems okay.

2. Then I connect a LC filter with the L and C available with me (L=1.8mH, C=0.68uF). The output is as below-

Seems okay too.

3. Finally, I connect the desired voltage 326vdc to inverter input. To check the waveform, I connect a normal iron core transformer (230v to 12-0-12) available at the moment to the inverter output to reduce the voltage to connect my pocket oscilloscope (risky to directly prove more than 60v). I found the following output without filter - (no load condition) -


Some sound comes probably from the filter inductor (12v to 326v DC section), some noise seem to come from ferrite core transformer of converter section. But no components get hot. Seems okay. But, the waveform is not satisfactory. I'm guessing the distorted waveform is due to - 1) the use of iron core transformer, 2) not connected any load, 3) no filter. These three reasons. how can I check the actual waveform with my pocket oscilloscope? I've 10x prove but that is not usable.

 

[BradtheRad]

Scope image #1 looks as expected. The carrier seems to be a lower frequency than you plan eventually, right?

Image #2 appears good, but my simulation has a more jagged waveform.

Image #3 went through a transformer and got filtered somewhat. It's still square-ish with switching jaggies. I think that's expected.

Here is my simulation using 2mH and 1uF, in a low-pass butterworth filter. With a load of 10 ohms the waveforms appeared reasonable.



If your load is lighter, then I believe the inductor and capacitor need to be higher values.

 

[sam781]

@BradtheRad: Thanks for your valuable feedback.I tried to make circuit in Circuit Simulator v1.6i to check the waveform as you gave in the above post but didn't find the Triangular wave generator.


It would be great if you could help in this regard.

 

[BradtheRad]

@BradtheRad: Thanks for your valuable feedback.I tried to make circuit in Circuit Simulator v1.6i to check the waveform as you gave in the above post but didn't find the Triangular wave generator

Open the edit window for an AC source.

Click the waveform selector to see the choices.

 

[sam781]

Still unable to make the same thing that you had made.:sad:

Source settings (from top)are -

1. Max voltage=5, sawtooth, f=1kHz, offset=5v
2. Max voltage=6, A/C, f=50Hz, offset=0
3. Max voltage=5,sawtooth, f=1kHz, offset=5v
4. Max voltage=6, A/C,f=50Hz, offset=0

 

[BradtheRad]

One of the sawtooth offsets should be -5V. (I believe you should change #3.)

My sinewaves are 10V amplitude. No DC offset.

----------------------------------------------

Notice my op amp outputs each go to 0V during half of the cycle. The outputs are not high impedance during that time. Your own H-bridge must do the same thing. You must allow the inductor to continue to conduct during idle gaps between pulses. The mosfets have a body diode , but I'm not sure if that can perform this role effectively.

 

[sam781]

Still unable to generate the expected waveform. Could you please share your file?