# High Power Square AC Wave Generator?

#### AnalogGecko

Joined Dec 15, 2016
5
Hello all, I've been reading here for quite a long time and finally decided to get an account, because I've run into an issue that I cannot seem to come up with a way around.

Let me preface this by saying that I grew up playing with 480V three-phase in my family's machine shop, so I understand the basics fairly well. I have always enjoyed power electronics, but when it comes to waveforms, I get confused in a hurry.

To cut to the chase, I need to output a reasonably square AC wave at some serious power (~2.5 KW). How in the world would I go about doing that with analog circuitry? It's not that I have anything against micros, I just don't ave a pc at the moment so using one would be exceedingly difficult, and I'd like to learn more analog anyway. It's too easy now to just throw micros at everything.

An H-Bridge is the obvious topology to generate an AC waveform, but how could I drive the H-bridge to generate my wave? I vaguely understand integrated driver circuits, but some clarification would be greatly appreciated. I need a frequency somewhere in the neighborhood of 100-300Hz, so not exactly a super fast switching problem by any means. Precision or period stability is really not an issue here (I'm not really sure how to articulate my point here, what I'm trying to say is that it doesn't matter is one cycle is slightly longer than the next) . Being able to adjust the frequency is a bonus. I understand that I am switching very large currents (200A at peak current) and I will deal with the power dissipation appropriately. Large industrial device packages and plenty of heatsinking. I also understand that to keep switching losses low, I need a high current driver to keep the transistors out of their linear range as much as possible.

To allay any high-voltage fears, this is a very low voltage application. 35v maximum. I seriously appreciate any insight anyone can provide.

#### Papabravo

Joined Feb 24, 2006
16,834
There are some incongruities in your post. With a voltage limit of 35V you would need approximately 71.5 Amperes to reach the power level of 2.5 kW. You don't exactly have a clear description of your requirements nor of the resources at your disposal. Some AC waveforms, like line voltage, are sinewaves that are symmetrical about GROUND. Other waveforms swing between GROUND and some maximum voltage; these can be generated by starting with a DC power supply.Which of these two possibilities did you have in mind?

Would you mind telling us what the resulting waveform will be used for?

#### tcmtech

Joined Nov 4, 2013
2,867
Standard H bridge with the proper high side/low side switching device driver ICs and whichever basic Pulse Width Modulator IC you chose to tell them what to do.

Standard basic analog circuit start to finish.

#### AnalogGecko

Joined Dec 15, 2016
5
There are some incongruities in your post. With a voltage limit of 35V you would need approximately 71.5 Amperes to reach the power level of 2.5 kW. You don't exactly have a clear description of your requirements nor of the resources at your disposal. Some AC waveforms, like line voltage, are sinewaves that are symmetrical about GROUND. Other waveforms swing between GROUND and some maximum voltage; these can be generated by starting with a DC power supply.Which of these two possibilities did you have in mind?

Would you mind telling us what the resulting waveform will be used for?
I apologize for my lack of information, I meant symmetrical about ground. To have the desired effect, the hot output lead will have to swap polarity. This circuit will be connected to a constant current DC power supply, and be used as an output stage for a GTAW welder. The switching devices will have to tolerate at maximum approximately 200A. the 2.5kW figure is only peak power, the 35v figure is only open circuit voltage. full load voltage will be much lower, 12 to 20v, getting much closer to my 200A peak figure.

#### Papabravo

Joined Feb 24, 2006
16,834
I apologize for my lack of information, I meant symmetrical about ground. To have the desired effect, the hot output lead will have to swap polarity. This circuit will be connected to a constant current DC power supply, and be used as an output stage for a GTAW welder. The switching devices will have to tolerate at maximum approximately 200A. the 2.5kW figure is only peak power, the 35v figure is only open circuit voltage. full load voltage will be much lower, 12 to 20v, getting much closer to my 200A peak figure.
OK. I understand now. The number of useable components at these power levels is somewhat limited. Were you thinking of an integrated H-bridge or one constructed from discrete devices? Were you thinking of just a simple oscillator to switch the H-bridge polarity back and forth? Or were you thinking of something more elaborate?

#### AnalogGecko

Joined Dec 15, 2016
5
An H-Bridge from discrete devices for a number of reasons. Power dissipation being the first and foremost. Can you even actually get integrated circuits with that much power handling?

Simpler is better at the moment. I'd like to be able to change the balance in the future, to clarify what I mean by 'balance', I want be able to vary the percentage of the total period that is positive or negative. Any of that is way in the future though. Just a simple symmetrical wave would be best for now, especially if I can vary the frequency.

#### Papabravo

Joined Feb 24, 2006
16,834
An H-Bridge from discrete devices for a number of reasons. Power dissipation being the first and foremost. Can you even actually get integrated circuits with that much power handling?

Simpler is better at the moment. I'd like to be able to change the balance in the future, to clarify what I mean by 'balance', I want be able to vary the percentage of the total period that is positive or negative. Any of that is way in the future though. Just a simple symmetrical wave would be best for now, especially if I can vary the frequency.
Nothing would surprise me when it comes to the geniuses who work in the dark forests of Silicon Valley. I guess I would start with a suitable IGBT. They seem to have some favor among the high power crowd. Can you give me a link to the datasheet for your high power DC supply?

#### AnalogGecko

Joined Dec 15, 2016
5
Nothing would surprise me when it comes to the geniuses who work in the dark forests of Silicon Valley. I guess I would start with a suitable IGBT. They seem to have some favor among the high power crowd. Can you give me a link to the datasheet for your high power DC supply?
I'd be happy to link you to a datasheet, if one existed. My supply is cobbled together from a few old welders.

Any reason for the IGBT? I was planning on MOSFETs.

#### dannyf

Joined Sep 13, 2015
2,197
With that kind of power, your best bet is an off the shelf product, if it even exists.

#### Papabravo

Joined Feb 24, 2006
16,834
I'd be happy to link you to a datasheet, if one existed. My supply is cobbled together from a few old welders.

Any reason for the IGBT? I was planning on MOSFETs.
In designing an H-bridge to work with existing components it is sometimes helpful and often times essential to know the capabilities of devices you are interfacing with. It tends to avoid nasty surprises that crop up when verbal and other non-precise methods of communication omit or cover up important details. There are significant safety questions here and I don't feel comfortable proposing a design without the information I require.

The IGBT is designed to handle high power applications with features for connecting heavy conductors and substantial heat sinking.