Newsgroups: sci.electronics
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!think.com!ames!xn.ll.mit.edu!xn!tonyb
From: tonyb@titania.juliet.ll.mit.edu ( Tony Berke)
Subject: Re: light bulbs buzz when dimmed
In-Reply-To: ins_atge@jhunix.HCF.JHU.EDU's message of 4 Apr 91 01:22:37 GMT
Message-ID: <TONYB.91Apr4110656@titania.juliet.ll.mit.edu>
Sender: usenet@xn.ll.mit.edu
Organization: M.I.T. Lincoln Lab - Group 43
References: <1991Apr2.162037@maximo.enet.dec.com>
	<1991Apr3.223655.18202@cbnewse.att.com> <7893@jhunix.HCF.JHU.EDU>
Date: 4 Apr 91 11:06:56

In article <7893@jhunix.HCF.JHU.EDU> ins_atge@jhunix.HCF.JHU.EDU (Thomas G Edwards) writes:

   In article <1991Apr3.223655.18202@cbnewse.att.com> random@cbnewse.att.com (David L. Pope) writes:
   >> an annoying buzz when dimed.  Is there a better circuit for dimming
   >> the bulbs that will prevent that !@#$%^&*&^%$# buzz?

   >and min for the bulbs you are using). The actual cause of the buzz
   >is the length of the fillament in the bulbs, so maybe you could
   >find large diameter bulbs designed for dimming (that is, with short
   >fillaments).

   O.K...but exactly why do the bulbs buzz when dimmed, but not when at
   full power?  My guess on this always was there was some kind of
   clamping of the waveform going on, making it look more and more like
   a square-wave with associated audible harmonics?  In that case perhaps
   a variable-transformer would make life better.

When you use a normal SCR dimmer, 50% power is produced by shutting
the juice off for 50% of each waveform -- right at the peak voltage!!.
This makes a really sharp edge, which indeed has lots of high
frequency harmonics (it's not a square wave, though).

I've always assumed that filaments buzz because they are coiled and
have some of the qualities of an electromagnet -- even if the filament
isn't coiled, when you vary the current through a conductor you get a
magnetic field.  As the current changes, so does the field -- and when
you use an SCR dimmer, you vary the current quite abruptly, causing
the field around the filament to go bannanas.  This results in the
filament wagging all over the place under the influence of the field.

Please note: this explanation might be total bullshit, I'm only
guessing.

The following is not a guess, however: A company (recently defunct) by
the name of LightSync made some dimmers that were essentially
switch-mode power amplifiers -- they used giant MOSFETs as output
devices and put out DC.  Their product was called HUMPHREY (a play on
"hum-free").  Switching amplifiers are not simple to design, however,
and I recall that they had quite a time keeping the RFI down (40kHz
200V square waves are NASTY!).

I bet that zero-crossing dimming would be quieter than the normal scheme,
but I think you would wind up with visible flicker if you try to use this
scheme for dimming much more than 50%.

Have fun!     Tony Berke 
