----- Original Message -----
From: "hermans" <on4kj@skynet.be>
To: "'K8RI on TowerTalk'" <K8RI-on-TowerTalk@tm.net>; <WarrenWolff@aol.com>;
<TOWERTALK@contesting.com>
Sent: Saturday, November 18, 2006 6:09 PM
Subject: [TowerTalk] RE : Query RFI
> Those who own the Math. Will tell you have a look at the FOURIER Series...
> If I remember well, a square wave generates very high pieks odd
The rise and fall times dertermine the frequency. The wave form is composed
of many harmonics. The amplitude of the harmonics drops off rapidly with
increasing frequency for a given wave form.
> harmonics.The higher the harmonic the higher the piek.
The shorter/faster the rise time the higher the frequency and number of
harmonics that make up the wave form and thus the total energy in those
harmonics. However as you go up in the number of harmonics the energy drops
off with each sucessive harmonic. "I think" it drops off by the square of
the number. IE each sucessive one has a quarter the energy of the previous
one...but that sounds a bit high.
This is drifiting a bit, but when it comes to light dimmers, triacs and SCRs
(Silicon controlled rectifiers not Saturable core reactors which work much
better but weight a bit more<G>) They vary the brightness by varying the
point at where they turn on the switch during the AC cycle. It's a
relatively simple circuit, but I've forgotten the name of the triggering
device. (Diac maybe?) Be it a triac or SCR there are only two ways of
turning them off. They automatically turn off when the voltage wave form
crosses zero (or rather close to it), or hitting it with a reverse voltage
spike. I know of none that use the second method but if you think today's
dimmers create hash you should see what they do when forced off. <:-)) The
turn on generates quite a spike with a fairly rapid rise time. The turn off
which happens just before the zero crossing, due to internal voltage drops,
also creates a spike, but of a much smaller amplitude.
Actually these things are a treat to make and play with. They can be a
lesson in interactive RFI and control. The basic units are not much more
than a pot, diac, and triac. The spikes generated will travel down the ac
lines for some distance and when they find another device ... <G> they just
might turn it on, or provided the timing is right, turn it off which
generates even larger spikes. With three or four of them all active you may
find turning one up, also turns another one up, but when you turn that one
down it may affect others. So, turning some up turns others down while
others go up. It's much like the old Laurl and Hardy movies where they are
being chased through a series of rooms off a hall going in one door and
coming out of another room entirely.
The Fourrier Transform is really an infinite series of sins and cosins. In
reality it is to some limit n where the power in the harmonic has dropped to
an insignificant value. I'd guess more than half on TT know more about this
than I do.
> At least fourty years ago..........hi hi
I have a math minor and it's only been 16 years but I couldn't do a
derivative or integral if my life depended on it. BTW my first courses in
grad school (CS) were digital image processing and the design and analysis
of algorithms. The second was easy. We only went to 5 level simultaneous
equations. By the third week in digital image processing we were doing
fourrier transforms (not FFTs) on images. From there we moved on to matrix
algebra...and the prof said "not much math". Just goes to show how
different a PHD and a lowly grad assistant view the term, "not much math".
<:-))
Roger Halstead (K8RI and ARRL 40 year Life Member)
N833R - World's oldest Debonair CD-2
www.rogerhalstead.com
>
> Jos on4kj
>
>
>
> As Dave has already said, Dimmers do make noise. Generally they are
> nothing
> more than a triac (bi-directional SCR) and create a terrible wave form.
>
>
>
> _______________________________________________
>
>
>
> _______________________________________________
> TowerTalk mailing list
> TowerTalk@contesting.com
> http://lists.contesting.com/mailman/listinfo/towertalk
>
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|