F
Foley U. Matthews
Guest
On Stardate Tue, 23 May 2000 10:05:25 +0800, "David Emrich"
<demrich@ihgtech.com.au> applied digits to the keyboard and routed the
information from some kind of brain (presumably), thusly:
are no e's in my true email | Lotteries, Gambling... The Executive Producer
Visit the Ellen Foley Info | is to blame!... and perhaps, Love (is/can be)
http://www.go.to/ellen-foley | "Fully expecting to be Hurt!"
<demrich@ihgtech.com.au> applied digits to the keyboard and routed the
information from some kind of brain (presumably), thusly:
Foley U. Matthews. there | I believe in : Paying NO voluntary taxes, i.e.D.A.Kopf <dak@dakx.com> wrote in message news:3929BE49.7F69DFAC@dakx.com...
Fred McGalliard wrote:
"Rowan Crowe (NoSpam)" wrote:
Hi all,
I am interested in experimenting with flourescent lights, powered from
a
nominal 12VDC source. I have a few questions...
1. What voltages and type of waveform are required to light a typical
flourescent tube?
The tubes light by low vacuum arc. You can rap one with a few KV to
start the discharge, dropping to a few hundred volts if you heat the
filament first, or use an RF discharge and a coupling foil near one end.
Once you initiate a discharge, you need to drag the charges across the
tube with enough voltage to see that every mean free path length they
gain enough energy to create a photon at the mercury line, and enough
more to occasionally generate an ion pair to replace the ones on the way
out. In a standard fluorescent tube, about a meter long, this drops to
about 60V, as I recall. It is directly proportional to distance between
the end plates. If you use a 60 cycle AC, the tube will try to quench
(the arc tries to discharge all the free ions and stop being an arc)
fast enough that you can get a nasty flicker. If you were driving this,
I would opt for a square wave with close to 100% duty cycle, so the
current in the tube is zero only very briefly during the transitions.
The problem is two fold. Generate enough voltage to excite an arc, and
then to feed it a steady current at a voltage determined by the tube
length gas pressure, and temperature. This is all the big ballast
transformers do. You do not need to heat the filament to start the bulb.
And in fact, with RF power, you can make it glow nicely without even
contacting the filaments at all.
Or use current-limited DC (e.g. the rectified output from a saturated-core
transformer) of a couple thousand volts per series lamp. No flicker, no
need
for starting filaments (recycle burned-out AC tubes). Over the sweet spot,
the
efficiency is independent of the voltage as the brightness mostly varies
with
the absorbed power. At least in theory, if it works let me know!
Of course, if you're doing a DC supply, you can always use a "cold end"
current mirror circuit (just make sure the transistors have enough) to limit
the current, and a cockcroft walton multiplier with high value capacitors
(like 47uF or so) in the first couple of "rungs" and low value caps (like a
couple dozen nF or so) (all high *voltage* of course). This way, before the
tube starts, the whole multiplier is active, and generates a high voltage on
the order of a few KV, but once the tube fires and starts drawing current
(limited by the current mirror), the low value caps don't hold enough charge
between cycles to keep up output voltage, and the multiplier effectively
becomes a two-stage multiplier (only the high value caps have any effect,
and the voltage dropped across the remaining diodes take a few volts off).
The same principle was used in a successful laser power supply in an
Electronics Today magazine, probably nearly 25 years ago now. Essentially
the same principle (high KV start, 1 KV run), just have to fiddle the
numbers a bit.
David.
are no e's in my true email | Lotteries, Gambling... The Executive Producer
Visit the Ellen Foley Info | is to blame!... and perhaps, Love (is/can be)
http://www.go.to/ellen-foley | "Fully expecting to be Hurt!"