My Adventures with Designing A Full Sine PWM Inverter

[ibedonc]

Hi , I will use this to track my design of a FSW inverter , I am nick naming it Godzilla  this Intverter Will start at 40kw and be expanded to what ever you need

The output already can do 288kw , But will start with only enough DC/DC stepup to do

40kw

I thought about the multiable ways to do this and now I have up with a system that

would be easy to make , because I am using off the self MSW inverters as DC/DC

I take the output of a Cobra 2500w MSW run it into a Full Wave Bridge Rect. doubler

like you would find in a Standard PC smps this gives a +- 169volts  ( 120 x 1.414)x2

( there are 16 of the MSW's paralleled)

I am using much better Caps then you will find in those power supplies and more of them. this supplies power to 4 300amp IGBT's 2 of these will put out 120v ac and other 2 will do the same but 180degree out . so that you get 240v ac just like your (maybe ex ) power company does , each one of these I call a FSW module and they will be 4 of them

of course there will be batteries connected to the MSW's

I have most of the parts here or on the way.

the 60hz Sine will come from a Crystal based circuit

found here

http://www.fairchildsemi.com/an/AN/AN-42025.pdf

it works I breadboarded it tonite , it makes a perfect sine wave right at 60hz , I do not have a freq counnter , but on my scope it is dam close , I have layed out a board based on this and will be making it this weekend

it gets its +5 , -5 from a LM555 running as a neg volts converter .

this will drive a National LM4651 class D chip or I will also be testing a PWM based on

LM555/LM339. The IGBT drivers will be IR2110.

I will post Schmatics and Board layouts as I make sure they are tested

well that is it for tonite

[ibedonc]

one other thing , you could use it for smaller systems , and turn your MSW into a FSW

I started at 40kw because sometimes I need it

[drdongle]

Pictures and Schematics Man! we need Pictures and Schematics!!

Dr.D

[ibedonc]

working on that

[PaulJ]

   My MSW inverter has performed flawlessly for over a year now, but it does have its limitations, specifically running induction motors, some brands of compact fluorescents, and some nicad (cordless drill) chargers. On the other hand, it's very efficient. I don't want to scrap it and spend thousands on a FSW unit, so needless to say I'm VERY interested in anything you come up with. Please keep us posted!

[ibedonc]

I have a working Breadboard Prototype of the PWM working and it is driving mosfets for test and low pass filter out of that is my pure sine

I am working on getting this much into a schematic

[Budgreen]

Just out of curiosity, how will you controll current from the MSW inverters?

what I'm looking at is the overload possibility of them, since you are rectifying these into the +/- 165v to be switched and buffering with caps, when the IGBT's switch it essentially creates a dead short, there could exist the posibility of triggering the overload protection on the inverters.

on a side note I have some small dual H bridge motor driver ic's, has anyone used something like this to drive a larger H bridge?

[ibedonc]

current shunts in the FSW PWM also I will have it run to a breaker panel

>when the IGBT's switch it essentially creates a dead short

not true only if something on the A/C side shorts and there will be breakers between that and the FSW

when the FSW is running with no load the transistors do not even get warm, it will have very low idle current , do not have a amount yet but will test that when I have it running on the high volts test

 I am going with IR21834's  as my IGBT driver

I have a IR2110 working , but do not like it , the 2183 has built in Dead time and

low side invert , only have to feed it from the comparator

as of last nite I have a low voltage version working on my Breadboard

I was having trouble getting the high side to work until I figured out I was missing a wire , put that in and perfect sine out of my inductor/cap filter , connected a 24v transformer secondary to it and out of the 120v side I got 120v

now I will make a board for the high volts test and try and get a veriac for testing

not using the LM4651  , using a LM555 for a RAMP generator , TL082 op amps

for sine and RAMP amps , then one of TL082's used as a comparator

will have a new circuit posted by saturday

[Budgreen]

ok, sounds good!

I also have a simple PWM circuit built on a breadboard in my pile of junk somewhere, no reference sinewave on it tho (was using signal generators) so maybe an inverter design I was testing awhile back was wired wrong or didn't like the fets I was using because they created a dead short to the supply (wasn't a deadtime issue) must have been in my filtering -shrug- eventually I will try gain

best of luck on your project.

[John II]

Hi: I'm new to this forum, I have already spent about 10 hours pouring over the materials. I love it because you guys are more than talk ! Most of you are actually trying out some of the things you talk about ! haha  A great forum indeed !

I live on a small farm, with a ultra low income, have driven nothing but junk vehicles for years, but you should see our PV arrays ! (smile) Our first PV array is a 1 kw array wired for 12v at about 80 peak amps output. I got tired of copper cable about the size of a garden hose. So after pinching dimes forever.. we built a 2 kw array on a heavy steel fixed frame wired for 120v dc output at about 16 amps.

But I ran into a problem that I did not anticipate... that is, 120 v dc input inverters jump into the commercial realm and cost about $6,000 upwards ! Something I simply can not afford at our income level. I can run alot of power tools directly off of the dc though which is nice.

A fellow in Home Power magazine that I have made friends with in upstate NewYork has built a 120vdc input inverter, but it's output is a total ultra harsh square wave ( not just an ordinary square wave). I bought the parts to replicate it, but the more I look at it... the more I fear of overheating and burning out half of the motors on our place !

Sooo Ibedonc, You and others on this forum  that are working on inveters have given me some fresh ideas to work with. So please keep us filled in on all of your progress.

 I'm a electronic hobbiest, but know just enough to get myself into trouble ! haha I have been scanning Search engines for about 2 years trying to find others working with inverter circits so I will have more knowledge about them and have their sucess and pitfalls to build upon. I'd be delighted with even a modest pwm inveter..... anything would beat a fixed square wave unit : )

John II

[ibedonc]

I am Just finishing up on my latest Circuit , and it works , most of the parts can be had a RadioShack , but one of them can only be got at Arrow Elec , that is the ML2035

sinewave gen.  I am not using the LM4651 and I have designed a PWM using TL082 op-amps

also found that a IR21834 mosfet driver is the best

the PWM/driver board only has 3 8pin chips and one 14 pin chip

you could make this for only a few 100 watts or up to what ever you need the only diference is the output mosfets or IGBT's

and it will fit in a very compact size

what gets me is why does the store bought stuff costs so dam much , I have less then  

$10 of chips in this circuit and the only parts that could rise the price is the mosfets and they all ready use those in a MSW inverter, no reason for MSW inverters to  

exist , I guess they think they can charge more because they are better and people do not know that it would not cost them anymore to make a Full Sine  

I am currently laying out the P.C. BOARD , I will upload the latest circuit tonite after I add some text labeling to it

you will need a min. of + -  80volts to get 120ac , you could connect 14 12v batteries

in series to get your + - volts that would be 168 volts

[John II]

I'm new to this forum and just now realized I posted my comments to you in your Diary portion of this forum. Should I post further ideas, comments and questions to you somewhere else other than here ?

I have got to say, I'm quite excited to see what you are working on. I have heard of digital pwm driven audio amplifers... but never saw any explination as to some of the back ground theory let alone actual schematics.. and Like I mentioned eairler, I have done quite a bit of web searching. I guess I just assumed if I could find little to know info on standard pwm inverters, that I wouldnt stand a chance of finding any info on any sinewave inverters.... especially developed for higher voltages.

Could you refer me to any web sites that might explain more of what's going on behind the schematics that you have posted ? I'm a amateur radio operator, and most of my electronic skills ( what little there is ! haha ) is primairly in linear electronics, and I know a whole lot less in digital, especially in power circuits.

When I can get a little time, I'm going to download the mfg's docs to the chips you are using etc.

John II

[ibedonc]

that is what it is here for , post all you want

[ibedonc]

New working Version uploaded

http://www.otherpower.com/images/scimages/348/FSW_Inverter.pdf

[DakotaSIG]

Nice, I've always thought inverters were overpriced too.

Take a look at this for some ideas for your next system: http://www.tinaja.com/magsn01.asp

From the web site:

Digital sinewaves with precisely controlled amplitudes and amazingly

low distortions. Compared to traditional PWM, magic sinewaves can

offer far higher efficiencies and lower distortions. With circuitry that

is elegantly simple and microcontroller friendly.

For use in such applications as induction motor speed controls, electric

autos, solar panels, power factor correction, inverters, home energy

efficiency improvers, 400 Hz avionics, UPS, PFC, and special utilities.

[ibedonc]

My Circuit is pretty simple and you do not have to work with a CPU. it is a lot easier

to make a board from parts that most can be had from Radio Shack or Digikey

[ibedonc]

Well I have Learned that I do not like IGBT's they do not behave like MOSFETS

they do not turn on as cleanly and get to much rounding , the MOSFETS I was testing

with on my Breadboard where nice and square, I will have to do more research on them

or just stay with MOSFETS

[DakotaSIG]

OK, but look at the site?

The guy is selling a preprogrammed 18-pin PIC chip. The PIC, some support chips, user I/O, and your half-bridge drivers would be all that you need. I would guess that he would also be willing to provide limited help with any technical issues too.

[DakotaSIG]

To add: the chips sell for under $20. Source code and one of hour of consulting is only $89! Here's the data sheet and application note: http://www.tinaja.com/glib/mschips.pdf

[John II]

Ibedonc : When You say " + - 80vdc to get 120ac "  Are you talking a split rail votage here ?( a positive reference to ground and a negative reference to the same ground of 80vdc each ? )

Idealy I need an inverter that can operate on an input of between 105vdc and 160vdc. I guess this would be the normal operational levels between a 120v bank of standard deep cycle lead acid batteries ?

I have printed off all of your work that you have published... as soon as I get the time, I'm going to print off the chip specs as well. I have to say... right now I don't understand exactly what's going on within the circuit you have But I hope too soon : )  It sounds so promising, yet alot more simple then a CPU based setup.

Do you belive your circuitry would be adaptable to my situation where I need aprx. 105vdc to 160vdc input and 120vac output ? would your circuit drive a step up transformer so that 120vac RMS would be availalbe at the low end of 105vdc input, and in about 5 to 8 kw in size ?

John II

[ibedonc]

yes +80 / gnd / -80 for 160 volts but + - 100 volts would be better

A DC/DC would be best I will be posting one of those also ,in your case you would put 100 -->> 200 volts in and get +-100 volts out for the FSW. I am going to use 48volts for my DC/DC , hopefully I will be able to soon Document the Schematics, I have been building and when I get a working one built then

I add the correct values

[hilo90mhz]

Really interesting project...

288KW !!! youre even more crazy then me !

Ive been building a 15KW 3 phase pwm sine wave inverter for about a year now.. on and off.

Right now I have a 1kw prototype thats fully functional, Im using a PIC16F628 microcontroller to generate the PWM signal, it outputs on a 19khz carrier frequency modulated to produce a 60hz sine wave, with 128 steps per quadrant of the sine wave.

If youd like to see the code or get a chip just ask.

Im really wondering if feedback will be an issue, were you planning on implementing any kind of real time voltage feedback?

Like when you hook the inverter up to a load with a bad power factor, if there is no feedback the inverter output will become all distorted, most of the professional inverters have real time feedback; they compare the output voltage to the real sine wave value.

I can see you have some kind of voltage feedback on the output, the schematic is a little hard to follow. Is your feedback more of an average output voltage type of thing?  

Chester

http://www.hot-streamer.com/hilo90mhz/

[ibedonc]

Im really wondering if feedback will be an issue, were you planning on implementing any kind of real time voltage feedback?

I will control the output level with a 220 --> 6v transformer which controls the input level of the SINE

yes the schematic is rough , I plan on cleaning it up when I have it working with loads

on yours why so low of carrier freq ?

I am not really interested in microcontrollers , my sine generator outputs a pretty nice sine and you do not have to program it

[hilo90mhz]

I dont have any real time voltage feedback yet... like you im still wondering how much of an issue it will be.

The carrier is low because of the IGBTs.. They have a relatively long turn on/off time and that can make up alot more of the losses than the voltage drop of the device. And thats just the frequency of the PWM module in the micro im using, i can be higher, but then it gets more complex to program.  I think the commercial inverters use about 20 khz..

[ibedonc]

yes , after working many years with MOSFETS , I am not impressed with IGBTs

Mine are rated for 200khz , but even at 50 that do not switch like mosfets

I am testing IRFP360 mosfets now