[sci.electronics] Fluorescent tubes

dya@unccvax.UUCP (York David Anthony @ WKTD, Wilmington, NC) (04/15/89)

	Here's a question for you net.analog.gurus out there, which
has us normally sane EE types freaked out:

	Consider your normal, low technology 40-wat two tube (four
pins per tube) fluorescent light.  This is the kind which says "Rapid
Start - Preheat" on the envelope.  We have such a beast which appears
to eat tubes; the tubes turn very black at one end.

	Replacing the ballast and checking the line voltage (which is
OK) has no effect; the fixture lunches a tube in about two days of
residential use.

	I was always under the impression that the ballast was some
kind of funky saturable reactor which, when all that nice magnetic
flux wasn't being used to sustain ionization in the tubes, allowed 
the filament current to increase to the level which causes ionization
(and once the tube ionizes, the heavy core saturation, distribution
of the windings, whatever caused filament voltage to fall back so
they last a long time.)

	I was wondering about the one in my plant light as well, which
uses a "starter" (which appears to be something like an Amperite time
delay relay...no, I'm not that old :-) :-) ) but whose filament seems
to burn only at **one** end when starting. (The rapid start lights 
start so quickly, you can't see the filaments heat up...). In the plant
light case, the tube turns black at one end as well, but it is very
finicky about the combination of lamp and starter (usually, several
packs of starters have to be tried every couple of years to find one
which makes the lamp happy...).

	Tubes with burned filaments at one end don't work or glow very
dimly (obviously)....

	Anyway, back to the subject.  The "Rapid Start-Preheat" 40 watt
tubes above my head have been in there about 6 years, operate 8 hr/day,
no problem.  The victim, who claims to have wired the ballast correctly
(and how could you screw this up?) says his "identical" fixture eats tubes
in a few days. What gives???

	And briefly, what is the ultimate limit on the tube life (I assume
lack of filament emission is fatal, but what about the tubes which have
only **two** pins??)  

	Is the structure in the tube just a heater at each end, or a heater
with a cathode tied to one pin (of some sort)....And, is the saturation
of the starter due to self-rectification (cold cathode a la 0Z3 tube) or
AC saturation.

	They didn't teach about these things when I went to college, so
please bear with my ignorance :-) :-) :-)

York David Anthony
DataSpan, INc

larry@kitty.UUCP (Larry Lippman) (04/16/89)

In article <1432@unccvax.UUCP>, dya@unccvax.UUCP (York David Anthony @ WKTD, Wilmington, NC) writes:

> 	Consider your normal, low technology 40-wat two tube (four
> pins per tube) fluorescent light.  This is the kind which says "Rapid
> Start - Preheat" on the envelope.  We have such a beast which appears
> to eat tubes; the tubes turn very black at one end.

	First of all, I must clear up a point.  The "Rapid Start - Preheat"
label means the tube can be used on EITHER a "rapid-start" fixture OR on a
"preheat" fixture.  These are two separate and distinct fixture circuit
designs; i.e., there is no circuit called a "rapid start - preheat" circuit.

	Preheat fixtures have a ballast and one "starter" per lamp, while
rapid-start fixtures have just a ballast.

	In a "simple" preheat circuit (generally used for lamps of 15 watts or
less), all of the fixture circuit elements are connected in series as below:

AC ----/\/\/\/\-----[starter]-----/\/\/\/\-----()()()()----AC
       Filament                   Filament     Ballast	

The ballast has only one winding and acts only as a current limiting inductor
(i.e., by saturation).  Most starters used today are glow switch starters
containing two electrodes of bimetallic material within a glass envelope
containing neon gas.  When AC power is first applied the gas in the starter
ionizes, producting heat which closes the contact.  At this point current
flows through the filaments causing them to light, causing electron flow,
heat, vaporization of mercury and consequent ionization with conduction
through the fluorescent tube.  Also, at the instant that the starter contacts
close, the neon gas ceases to ionize, which will cause the starter contacts
to open in a few seconds due to removal of the heat source.  If the
fluorescent tube successfully fires, the voltage drop now present across the
starter will prevent the neon lamp from again firing, so the starter has done
its job until AC power is removed and re-applied.  The ballast limits current
through the fluroescent lamp to a safe, design level - since a fluorescent
lamp with ionized gas is not a "self-limiting" device.

	Larger lamps (such as 20 or 40 watts) require a higher starting
voltage than 120 VAC - even with the filaments in the fluorescent lamp
lighted.  Consequently, the ballast also functions as a step-up transformer
(of the auto-transformer configuration); the increased voltage can range
between 150 and 250 volts.  This ballast is much more complex (too complex
to easily represent over the Net), and typically contains four windings,
plus two capacitors.  This circuit is called a "preheat lead-lag ballast"
circuit.

	A "rapid-start" circuit contains no starters; its ballast is also
complex and contains a step-up autotransformer.  The important difference
in an rapid-start circuit is that the filaments in the fluorescent lamp
are on at all times, although at a reduced level when the lamps have
successfully fired.

> 	Replacing the ballast and checking the line voltage (which is
> OK) has no effect; the fixture lunches a tube in about two days of
> residential use.
> 	Anyway, back to the subject.  The "Rapid Start-Preheat" 40 watt
> tubes above my head have been in there about 6 years, operate 8 hr/day,
> no problem.  The victim, who claims to have wired the ballast correctly
> (and how could you screw this up?) says his "identical" fixture eats tubes
> in a few days. What gives???

	I bet that the ballast is incorrectly wired.  The wiring of either
a rapid-start or preheat lead-lag ballast is extremely critical, and it
is quite possible for two interchanged wires to still allow the lamps to
light, but result in excessive filament current.  Or the replaced ballast
could be defective (a ballast ain't exactly built like a mil-spec
transformer).  Or it would be the _wrong_ ballast for the fixture.

> 	I was always under the impression that the ballast was some
> kind of funky saturable reactor which, when all that nice magnetic
> flux wasn't being used to sustain ionization in the tubes, allowed 
> the filament current to increase to the level which causes ionization
> (and once the tube ionizes, the heavy core saturation, distribution
> of the windings, whatever caused filament voltage to fall back so
> they last a long time.)

	The above is sort of true, but the ballasts have multiple windings,
also function as a step-up autotransformer, and are more complex than
your statement indicates.

> 	I was wondering about the one in my plant light as well, which
> uses a "starter" (which appears to be something like an Amperite time
> delay relay...no, I'm not that old :-) :-) ) but whose filament seems
> to burn only at **one** end when starting. (The rapid start lights 
> start so quickly, you can't see the filaments heat up...). In the plant
> light case, the tube turns black at one end as well, but it is very
> finicky about the combination of lamp and starter (usually, several
> packs of starters have to be tried every couple of years to find one
> which makes the lamp happy...).

	There are two types of starters: (1) the glow type, as I already
described above; and (2) the thermal type.  The thermal starter is
roughly equivalent to an Amperite time delay tube.

	Some fluorescent lamp circuits heat only ONE of the tube filaments;
this is referred to as an "instant-start" circuit.  Actually, there are
two types of instant-start circuits: (1) the series instant-start circuit;
and (2) the lead-lag instant-start circuit.

	And then there is the "trigger start" circuit, and others...

	I don't have enough information to ascertain what type of lamp
circuit you have, but I still bet you have an incorrectly wired, defective
or incorrectly _specified_ (a common problem!) ballast.

> 	And briefly, what is the ultimate limit on the tube life (I assume
> lack of filament emission is fatal, but what about the tubes which have
> only **two** pins??)  
> 	Is the structure in the tube just a heater at each end, or a heater
> with a cathode tied to one pin (of some sort)....And, is the saturation
> of the starter due to self-rectification (cold cathode a la 0Z3 tube) or
> AC saturation.

	The lamps having only one pin at each end are called - what else? -
cold-cathode lamps.  Their circuit uses no starter, but has yet another
ballast variation containing an autotransformer, two series indusctors,
three capacitors and one resistor.  All inside the ballast, no less.

	Cold-cathode lamps have the longest life of any fluoresecnt lamp,
but they also provide fewer lumens/ft-of-length than hot-cathode lamp
circuits.

> 	They didn't teach about these things when I went to college, so
> please bear with my ignorance :-) :-) :-)

	I never learned about fluorescent lamp circuits while getting my
EE either, but I sure did learn quick-like after college when in the "real"
world (I had to design a UV light source for an analytical instrument).
I did, however, as an EE student learn (what seemed to me) an inordinate
amount about motors, generators, power transmission lines and calculations
involving polyphase circuits (which were boring beyond belief). :-)

<>  Larry Lippman @ Recognition Research Corp. - Uniquex Corp. - Viatran Corp.
<>  UUCP:  {allegra|ames|boulder|decvax|rutgers|watmath}!sunybcs!kitty!larry
<>  VOICE: 716/688-1231, 716/773-1700        {att|hplabs|utzoo}!/
<>  FAX:   716/741-9635, 716/773-2488      "Have you hugged your cat today?"