Pure Sine Wave Inverter design

[keith03]

I would like to design a Pure Sine wave inverter for DC to AC rectification. I want to drive motors with this applicaiton. (At most 120VAC .5HP) I don't want to specify too much, because the main point here is design basics. With the right information, I should be able to design accordingly. I have found many inverter deisgns, but none suitable for powering motors. This is part of a senior project for my BSEE. I am receiving little support in this area. With the design basics in hand, I will run as far as I can before a reply. Thanks ahead of time.

Any good reads on modern design would be appreciated too!

 

[berkeman]

I would like to design a Pure Sine wave inverter for DC to AC rectification. I want to drive motors with this applicaiton. (At most 120VAC .5HP) I don't want to specify too much, because the main point here is design basics. With the right information, I should be able to design accordingly. I have found many inverter deisgns, but none suitable for powering motors. This is part of a senior project for my BSEE. I am receiving little support in this area. With the design basics in hand, I will run as far as I can before a reply. Thanks ahead of time.

Any good reads on modern design would be appreciated too!

Welcome to the PF. Are the related threads at the bottom of the page any help for starters? You can also do a search here in the EE forum (use Advanced Search) for the word Inverter in the title of threads.

 

[Phrak]

Interestingly I have a very similar design problem in the same power range. But, if all you are driving are motors, you may be able to get away with a square wave output. What kind of motors?

 

[Bob S]

A square wave voltage has a lot of 3rd harmonic in it, and might increase the motor core losses (eddy currents, etc.). Read about inverter duty electric motors here:
http://www.usmotors.com/products/inverter_duty_motors.htm
Bob S

 

[keith03]

Thanks for the quick reply. I did not find help from theother posts due to many misinformed requests. I am looking from an engineering standpoint, and would love some ideas on making the "taditional" square wave inverters (ie. two fets to primary) to a more sinusoidal form. I want to make this an all purpose inverter. It is a design challenge let's call it. This is a huge bite considering I have never built an inverter, but all I have is time and a helluva lot of books. lol. Any working designs?

I thought that a Square wave design would cause humming in an AC motor. I would like to make a few of these, each slightly different depending on the type of load and frequency of use.

 

[keith03]

I unfortunately must power an already instaled hot tub motor, furnace motor, and dryer motor. My project is a windmill design from generator to inverter. The best part is that I am incorporating an already existing project into my BSEE project. I would like to know some design basics that will better my understanding of the inverter concept overall. I have a few books that thouroughly describe creating a square wave output, but it was written in 1978. Inverters have become much more advanced, but they are hard to find design info on.

 

[Phrak]

OK. Here's a few hints. You can approximate a sinusoidal output by driving the output voltage between some maximum and minimum output voltage and include bias to ground between each min and max. This gives you (if I remember correctly) about 30% harmonic distortion with the properly chosen time duration at each voltage level.

You should know that a 60Hz AC motor looks like a CAPACITIVE load--not inductive, to 60 Hz voltage components. This is a big deal when considering back emf. At higher frequencies, it's still inductive.

The second approach is to simulate a sinewave to very good appoximation by pulse width modulation. You inductively couple the load to the supply. Say we have developed +30 and -30 volt, low impedence supply rails, capable of both sourcing and sinking current. Switching between them at, say 30% and 70% duty cycle will develop an average voltage of 0.3*(-30) + 0.7*(+30) = 12V. Vary the duty cycle over time and you can develop an averaged sinewave voltage. Send it though the output inductor and it is smoothed.

Does this make sense?

 

[Phi Tran]

I would like to design a Pure Sine wave inverter for DC to AC rectification. I want to drive motors with this applicaiton. (At most 120VAC .5HP) I don't want to specify too much, because the main point here is design basics. With the right information, I should be able to design accordingly. I have found many inverter deisgns, but none suitable for powering motors. This is part of a senior project for my BSEE. I am receiving little support in this area. With the design basics in hand, I will run as far as I can before a reply. Thanks ahead of time.
Any good reads on modern design would be appreciated too!

Pure sin wave is 100% analog; A PWM can steps to simulate; I am wondering how pure you have in mind. You may not be able to have a "pure" (analog) from a stable DC. However there is method to smooth out the steps.
Good luck.

 

[keith03]

Good info from all. I don't know how pure a Sin wave I need. I just don't want to damage anything with a bad inverter. As far as I know, higher end inverters power motors just fine. Why would an AC motor appear as a capacitive load to an inverter? I never would of expected that. Is it because of the lack of ability for current to properly "alternate?"

I was wondering something on the way home today. Since I have AC power available, would the circuit design benefit in any way using the "good" AC signal as a model to help shape the inverted DC?

 

[keith03]

Also, Phrak, I think that I understand you, with the exception of tying each min and max to ground. I have seen designs that make the alternating DC voltage by sending the squarewave voltage across a primary of a transformer. Are you referring to refrencing the secondary to ground using a center tap? Could you send a schematic? It might be easier to follow a discussion with. Thanks in advance.

 

[mugaliens]

OK. Here's a few hints. You can approximate a sinusoidal output by driving the output voltage between some maximum and minimum output voltage and include bias to ground between each min and max. This gives you (if I remember correctly) about 30% harmonic distortion with the properly chosen time duration at each voltage level.

This works from a transitor-only approach, but with the losses you site.

You should know that a 60Hz AC motor looks like a CAPACITIVE load--not inductive, to 60 Hz voltage components. This is a big deal when considering back emf. At higher frequencies, it's still inductive.

If I'm not mistaken, this is due to the inductive motor's tendency to suck power while coming up to speed, and purely one side of the inductive effect on one side of the phase. At speed, that also holds true.

The second approach is to simulate a sinewave to very good appoximation by pulse width modulation.

Early (1987) power-smoothing devices employed an electrical engine-run generator. Huge. Heavy. Very expensive.

Send it though the output inductor and it is smoothed.

Modern switching power supplies can smooth the sinusoidal current to as fine as you want it, if you're willing to pay the price, and many do away with transformers altogether, relying instead on waveform moderators based on timing circuits.

Does this make sense?[/QUOTE]

Pure sin wave is 100% analog; A PWM can steps to simulate; I am wondering how pure you have in mind. You may not be able to have a "pure" (analog) from a stable DC. However there is method to smooth out the steps.
Good luck.

Good info from all. I don't know how pure a Sin wave I need. I just don't want to damage anything with a bad inverter. As far as I know, higher end inverters power motors just fine. Why would an AC motor appear as a capacitive load to an inverter? I never would of expected that. Is it because of the lack of ability for current to properly "alternate?"

I was wondering something on the way home today. Since I have AC power available, would the circuit design benefit in any way using the "good" AC signal as a model to help shape the inverted DC?

Also, Phrak, I think that I understand you, with the exception of tying each min and max to ground. I have seen designs that make the alternating DC voltage by sending the squarewave voltage across a primary of a transformer. Are you referring to refrencing the secondary to ground using a center tap? Could you send a schematic? It might be easier to follow a discussion with. Thanks in advance.

 

[hisham.i]

Usually the 3rd harmonic is eleminated by a 3 phase transformer connected in Y (star)connection.
but what about """"You should know that a 60Hz AC motor looks like a CAPACITIVE load--not inductive, to 60 Hz voltage components. This is a big deal when considering back emf. At higher frequencies, it's still inductive.""""

 

[Phi Tran]

Good info from all. I don't know how pure a Sin wave I need. I just don't want to damage anything with a bad inverter. As far as I know, higher end inverters power motors just fine. Why would an AC motor appear as a capacitive load to an inverter? I never would of expected that. Is it because of the lack of ability for current to properly "alternate?"

I was wondering something on the way home today. Since I have AC power available, would the circuit design benefit in any way using the "good" AC signal as a model to help shape the inverted DC?

By that I mean I know now what you want: "pure sine wave" grid tie inverter? - Because you already have the AC. Not as you said earlier to run your equipment,

For the sake of science; Yes you can! We have tried; It is a "waiting big big breakthrough", because it could be a single, few cents chip to work as such. Please be noted that the real problem of the renewable energy industry (including solar too . . .) is NOT to harness electricity but to integrate into the grid

Hope I do not go too far off topic.

 

[keith03]

I do not wish to grid tie. It will be a stand alone instrument as far as taking DC and converting it to AC for the load. I was mentioning the available AC from the grid as a spitball to help create this pure sine wave. Instead of cutting a circle out by hand, couldn't we just trace one?

What Back EMF protection should be in place? Significant diode protection? But wouldn't this then not allow for alternating current?

good info guys!

Not to down any information that has been presented, but I have no intention on buying anything commercial. I understand that big money is involved with high quality inverters. I need help finding out why and how.

Thanks to all again!

 

[Phrak]

Also, Phrak, I think that I understand you, with the exception of tying each min and max to ground. I have seen designs that make the alternating DC voltage by sending the squarewave voltage across a primary of a transformer. Are you referring to refrencing the secondary to ground using a center tap? Could you send a schematic? It might be easier to follow a discussion with. Thanks in advance.

Independent of the manner in which you implement it, the idea is obtain a node that will cycle through low impedances at levels 160VDC, -160VDC and 0V, 60 times a second.

This will be lower in harmonics than a square wave. It’s the half-way in-between solution between square wave and sine wave.

I’m not sure it’s even worth implementing over a sine wave output—it may present problems that aren’t quality/cost effective. The only way to tell is to paper-design both, and compare.


I don't know of a good way to attach schematics to this forum. I could use MS word as a vehicle but *.doc attachments are suspect for viral contamination, unlike others like .pdf n'stuff.

 

[keith03]

Phrak,

Thanks for the advice. I am unfortunately unsure how to approach this design. My first thought is to start with the two fet design, and play around with some filtering to see what result I can come up with. I have found that here in school, I often know the information and skills needed to solve a problem, but fail to call them without a push. Guess being able to do this is what makes a good designer.

As far as the schematic, could you put it in a word doc, and then print to PDF? Thanks again.

 

[Phrak]

Phrak,
As far as the schematic, could you put it in a word doc, and then print to PDF? Thanks again.

How do I do a print to PDF? I have Word 2003, by the way.