DC/AC INVERTER W/ IGBT's

[STABY]

Hi,
I'm trying to design a one phase, 2 kVA inverter using a IGBT "H" bridge.
There are two wave shapes for drive the bridge; one generated with PWM method, providing pulses of different width and same amplitude, of course according of the angle of the sine reference.
The other one can be an stteped wave with 4 steps on each half cycle.
I'm not sure wich is the best method, in special for the filtering purpose, because we need a pure sine wave at the output.
Some suggestion??, thanks.
Jorge

 

[Lewish]

What does your available source voltage look like?
Is it a single DC voltage, or is it a series of batteries?

 

[cbarn24050]

Hi, PWM is the way to go.

 

[peebee]

Re: "because we need a pure sine wave at the output."

You'd better define your output better than "pure," ie., some allowable percentage of THD on your output voltage waveform.

A 4-step wave shape won't even get you close to "pure". PWM will likely have to be deep in the kHz range. Either way, you'll still need some heavy filtering. In general, the faster the frequency of your inverter, the less filtering you'll have to do, no matter how it's configured.

Most commercially available inverters do a pretty lousy job at generating a "pure" sine wave.

 

[peebee]

Here's another thought on generating a "pure" output -- maybe you should consider a motor-generator set. That still won't be pure, but it might be closer and cheaper to pure than trying to synthesize a waveform electronically.

 

[STABY]

Lewish- I'm going to use one or two 12 v. battery in series connection. Please said me the importance of the DC level for choose beetwen PWM or 4 steps for drive my IGBBT's.
as we will use an step-up transformer to obtain 220 VAC, 60 Hz output.
peebe- Thanks, I'll like 3% THD as much.
Don't you think using PWM (kHz range) I'll need less filtering for 60 Hz. sine wave shape, and much for harmonics.
As I understand this is not a commercial site I'll prefer not ask for inverters brand names.
Jorge.

 

[cbarn24050]

Hi, you should not use igbts for low voltage/high current bridges, use mosfets instead.

 

[melone]

Won't there less voltage dropped across an IGBT since the element that is being used to as the switch is a BJT which will have almost a constant VCE? However, when you use a FET, the VDS will vary based on your current.

 

[Lewish]

STABY - to use a 4 step approach, that means you have to have 4 discrete DC voltages available for the input to create the 4 step waveform. Otherwise, how would you get the 4 voltage steps. Sounds like you have only one or step voltages available.

Melone - at input voltages below about 60 volts, use power FETs. The minimum Vce on IGBTs is about 2.0 volts. Some a little more, some a little less. A good power FET will have an on resistance of about 0.01 ohms. At 100 amps, that is only 1.0 volts drop. Yes, the on resistance will rise slightly with less current, but at absolute worst won't increase more than 50%. So, 1 amp at 0.015 ohms, ain't much loss.