The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
From: email@example.com (Gibbo)
Date: 16 Oct 2002 12:19:43 GMT
Organization: AOL, http://www.aol.co.uk
Subject: Re: How to construct PWM pure sine wave inverter ?
"fred bartoli" firstname.lastname@example.org wrote:
>Bill Sloman a écrit dans le message :
>> "Booboo" wrote in message
>> > >
>> > >Suggestion: google for 'magical sine waves'
>> > >
>> > http://www.tinaja.com/glib/stepsynt.pdf
>> > I think he means a ~real~ sine wave.
>> A pulse-width modulated digital data stream - just ones and zeros - looks
>> exccedingly like a sine wave if low-pass filtered by a good filter.
>> The problem with Don Lancaster's magic sine waves - which are just very
>> binary sequences - is that the high frequency noise is big and and you
>> a multipole filter to get it down to acceptable proportions. Not a problem
>> with a fixed frequency source, but for a variable frequency source, the
>> technique of driving a D/A converter from a phase accumulator via a
>> table is to be preferred.
>> For power applications, where you really need PWM, Don Lancaster's magic
>> sine waves probably have too many switching edges to be acceptable - each
>> transition dumps a lump of energy into the switch and you need to ration
>> number of transitions.
>> Bill Sloman, Nijmegen
>I don't see what's wrong with Don's approch for AC inverters of performences
>like the one here.
>If I've correctly understood the paper after a breaf reading, there are few
>switching edges (from fig 1) 6x4 = 24 pulses over a full sinus cycle. Much
>less than a 20-100kHz classic PWM approach. On the contrary, I would think
>that magic sin waves is much more favourable in this regard.
>Also for the "high freq" noise, the first major harmonics are the 23th and
>25th, so it would be easy to build a fixed low pass filter that have it's
>corner frequency between 1) the highest fundamental freq (80Hz in this case)
>and 2) the 23th harmonic of the lowest freq (30x23 = 690Hz) there's plenty
>of room in this window.
>In fact, what I find less practical in this approach is the size of the low
>pass filter elements, given the low corner frequency and maybe the amplitude
>of the obtained fundamental, but I'm not clear on this point having just
>thought a few seconds to it.
I first saw the references to magic sinewaves many years ago and was
immediately very interested. So interested that I breadboarded a sinewave
inverter using that technique.
I wish I had thought more carefully about it before I wasted so much time on it
There are many, many problems with it as far as a general purpose inverter is
concerned. Problems that are not really apparent until you start to look very
closely at it.
Intuition should tell you (at least I feel it should have told me - but I was
young) there's something not quite right when a PWM signal with 24 switching
transitions apparently has almost no harmonic content yet one can quite clearly
see it on a scope !
The fact is that the harmonic content of a magic sinewave is absolutely
horrendous. It just seems like it isn't when the harmonics are averaged over a
cycle or when a mathematical analysis is done.
The true harmonic structure is much more complicated that it at first appears
(eg very low harmonics that are themselves modulated by a phase shifted
fundamental [which is itself modulated] thus giving a low average figure but
harmonics (of low order) which are in actual fact of enormous amplitude) and
the system only approaches a workable figure when you get onto very long
sequences approaching equivalent switching transitions with a normal PWM
We didn't really get any practical results until using a sequence with over 400
switching transitions. Roughly equivalent to a PWM at 30KHz as far as harmonics
are concerned and 20KHz as far as switching transistions is concerned.
Given that (very fast) amplitude regulation is extremely simple in standard PWM
it transpires that there are absolutely no benefits whatsoever in attempting to
use magic sinewaves for this purpose and in fact one ends up with a system that
is completely unwieldy and over complicated.
The complete lack of commercial units using this techology should also say
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup