Connecting 3 Phase inverter and buck converter

[Girish88]

Hello ,

I am trying to simulate a 3 Phase inverter connected to a buck-boost converter at the output. When I run the simulation of circuits individually, they seem to work well but when I connect them, the result is very bad. The DC link voltage drops to half the nominal and the current through the circuit is high regrdless of increasing the output resistance. I am making a big blunder but am not able to figure it .. Could someone please help as it is important.

[robertpf]

Hi Girish88,

Could you please describe your expected behavior, simulation settings etc.?

After running the simulation for a minute on NI Multisim 13 the XMM2 shows 30V, and XMM1 shows 20V. Did you expect other values here?

Do you also get the error that the transient time point calculation did not converge, canceling the simulation?

Best regards,

[Girish88]

Hello Robert,

I am basically trying to simulate a inverter circuit followed by buck converter. 

1. In the inverter circuit, I have not attached any filter on the 3 phase side because it was damping my dc voltage. Without filter circuit, when I run the inverter part alone, for 20 V i get 42 V out which is theoritically expected result.

2. The buck converter also gives propotionate result (Depending on duty cycle) when the input of 42 Volts is given seperately in another file.

3. But when I connect both circuits together and simulate it, as stated by you the voltage in inverter out becomes around 26 and the output of converter becomes 24V for 100% duty cycle.

EXpected result would be 42 V out of Inverter and dutycycle related voltage i.e. 42V for 100% Duty from Buck converter.

I am not very well versed with handling the settings of Multisim and I am using the same settings as the example inverter circuit had.

Could you please give your comments for it will be very helpful.

Thank you !

[robertpf]

Hi,

I think I might be missing some parts of your project, when opening your program all of the nodes s1-s6 are set as on-page connectors, meaning the gate signal of Q1-Q6 will be undefined. How are you controlling the node signals s1-s6?

Have a look at this design guide if you are working with LabVIEW.

Could you please send me your entire project files?

Best regards,

[Girish88]

Hello Robert,

I am not controlling the gate signal in the above application as I am letting act as a full bridge rectifier (Making use of the Bidirectional mosfet to provide rectification) . But when I make the circuit work as a inverter, I use the 3 Phase gate signals. 

1. I will attach a circuit showing the normal inverter acting as a rectifier.

2. Circuit that shows inverter with 3. gate signals

[robertpf]

Hi Girish88,

I have now done some testing and hopefully this will bring you to a solution:

I agree with you that the INVERTERBIdirectional runs fine. In comparison with the INVERTERBIdirectional+buck program I see the following major differences:

• A huge capacitor C1 of 1F. A possible source to the simulation error due to non-converging transient time calculations?
• The VGnd in INVERTERBIdirectional+buck is connected to ground in the Q2-Q4 branch. This is most likely the reason why the output voltage is seen as "divided by two", and the short circuit will give you high currents.

If you lower the capacitor value and remove the connection to ground in the INVERTERBIdirectional+buck program, does it behave in accordance with your expectations?

Kind regards,

[Girish88]

Hello Robert,

 When I remove the ground and change the capacitor value( 1mF), the simulation runs for a some time and shows convergence error. But when I do a transient analysis, it shows the required output. May I know why is it so as I am not able to run the simulation in normal mode.

And I also read that the circuit cannot be run floating as Multisim requires some ground reference to make the nodal analysis. Because my ultimate goal is to take the voltage out of buck converter and make it charge a battery.

[robertpf]

Hi Girish88,

You can always do a measurement relative to ground without grounding a whole branch. The lowest voltage in the lower branch connecting Q2, Q4 and Q6 is below zero, and putting a gound node there will therefore give you half of the expected voltage (in this case).

The convergence error is a SPICE related problem, I recommend starting with this guide on solving SPICE convergence problems

Kind regards,

Robert