13 watt PLS CFL lasted whole 10 seconds!

[JohnR66]

The old bulb in the outdoor fixture was worn out, just blinks, no start. Got
a new bulb and installed. Blinked a couple times and started right up. After
a couple seconds it went out and would not restart. I put it in another
fixture. Heaters would light up orange (no arc forming across either one).
Very faint flicker of the bulb as starter clicked.

I was thinking the bulb lost some vacuum. I disassembled, check the starter
(okay) and connected the bulb leads to 600 volts to see if it would light,
It light up pinkish and flickered.

That was my lighting adventure for the day. Any thoughts on the very early
failure?

John

 

[Andrew Gabriel]

The old bulb in the outdoor fixture was worn out, just blinks, no start. Got
a new bulb and installed. Blinked a couple times and started right up. After
a couple seconds it went out and would not restart. I put it in another
fixture. Heaters would light up orange (no arc forming across either one).
Very faint flicker of the bulb as starter clicked.

I was thinking the bulb lost some vacuum. I disassembled, check the starter
(okay) and connected the bulb leads to 600 volts to see if it would light,
It light up pinkish and flickered.

That was my lighting adventure for the day. Any thoughts on the very early
failure?

Dim pink light from a fluorescent usually means it's lost its
mercury and the argon base gas is discharging. This dim pink
is usually a sustainable arc (at least on series ballast lamps
on 240V supplies). In bright ambient light, you might not see
the pink discharge as there's no fluorescence. I don't know if
this matches the symptoms you now have. If it does, then I
wonder why the lamp worked briefly? Maybe it is contaminated
with something which quickly eats all the mercury ions?

Reading your description again, another symptom is the orange
glow from the filaments. On a 240V switchstart lamp, the
filaments are normally heated enough (with sustained starter
operation due to non-starting lamp) that they cause local
ionisation and hence a white glow at the tube ends without
any voltage across the tube. The glow only drops back to
orange when there's no emission coating left. So maybe the
tube was manufactured with a missing emission coating, or
it wasn't properly fired on by operating the filaments during
the manufacturing process (not sure what effect that would
have).

 

[TKM]

The old bulb in the outdoor fixture was worn out, just blinks, no start.
Got a new bulb and installed. Blinked a couple times and started right up.
After a couple seconds it went out and would not restart. I put it in
another fixture. Heaters would light up orange (no arc forming across
either one). Very faint flicker of the bulb as starter clicked.

I was thinking the bulb lost some vacuum. I disassembled, check the
starter (okay) and connected the bulb leads to 600 volts to see if it
would light, It light up pinkish and flickered.

That was my lighting adventure for the day. Any thoughts on the very early
failure?

John

Sounds like a crack somewhere let in some air. The poor thing had enough
starting gas to die in a glow of color; but dead it is. Glass stresses or
poor seals at the cathodes are likely causes of "air" lamps. Such faults
are hard to test for in the factory since they can happen due to handling or
shock during transport.

Terry McGowan

 

[James Hooker]

OK your problem sounds typical of a lamp where one or both of the cathodes
was not correctly activated during production. These are coated with an
electron emissive material, normally a mixutre of the oxides and tungstates
of barium, strontium, yttrium, calcium etc. I try to describe below the
process leading up to the kind of failure mechanism you have seen.

This emitter coating is first applied to the cathodes in the form of the
metal carbonates eg BaCO3. During production it is necessary to "activate"
the material and convert the metal carbonates to the metal oxides. This is
done simply by passing an electric current through them to heat them to
incandescence. CO2 gas is evolved in the reaction below:
BaCO3 >> BaO + CO2

There is a further reaction between the CO2 and the tungsten coil to produce
tungstic oxide:
CO2 + W >> WO3 + CO

This in turn reacts with the metal oxides and partially converts them into
the metal tungstates:
3BaO + WO3 >> Ba3WO6

The combined mixture of metal oxides and tungstates provides an excellent
emission mixture for high efficiency operation of the lamps, with low energy
losses at the electrodes. During the high temperature treatment a tiny
amount of metallic barium is also produced, and a monolayer of this diffuses
over the electrode surface and dramatically improves the situation. The
evolved CO and CO2 is finally pumped away, and the lamp backfilled with its
Argon/Krypton mixture.

Occasionally, lamps manage to go through this production step with one of
the electrodes not being fully activated. This is generally due to a poor
electrical connection to the filament wires. Or sometimes if there is an
excess of emitter on the filament, then there may not be enough time to
fully activate all of it during the short time available in the production
process.

When the customer then lights the lamp, the electrodes heat up and this
activation process begins inside the sealed lamp. CO2 gas is evolved, and
its presence causes the lamp voltage to rise. Usually within a few seconds
the CO2 pressure is high enough that the lamp voltage rises above what the
mains supply can deliver, and it will flicker and go out. Of course if you
then put your lamp on a higher voltage supply you can re-start it and
sustain the discharge in the CO2 atmosphere. I think this is what you were
seeing. Of course on the normal mains supply the cathodes will glow orange
as usual, but the lamp won't be able to start up.

I have seen this problem during visits to many Chinese CFL factories. Since
they make the lamps with a lot of manual labour, its quite a frequent
occurrence that the operator will forget to connect the current supply wires
to the lamp filaments while they're on the vacuum pumping bench. Or often
they have a bad connection, most factories simply twist the wires together
to make a connection and it's not always very reliable. Some of the more
advanced factories do a 100% inspection before shipping finished lamps, when
they sort out the defects. But most simply ship 100% of what comes off
their machines, good or bad.

In the better Western factories there is normally a burn-in period followed
by a spectral check on the lamp, which monitors gas purity by looking at the
relative intensities of certain spectral lines in the discharge. This
instantly shows up the concentration of both CO2 and hydrogen in the lamp.
Automatic feedback systems will then reject the lamp. Of course even this
system is not 100% reliable but it does reduce the chances of a customer
receiving a lamp like this.

James.

 

[Clive Mitchell]

These had very high ballast compartment temperatures, much higher than
the Chinese products. I suspect that this may be related to the
premature failure, but I don't know for sure.

I'm finding that the cheap Chinese lamps outlast the Philips Geni lamps
in style. The Geni lamps suffer early blackening of the electrode area
and often seem to suffer from mercury absorbtion meaning they are pink
at startup and take a while to reach their full output.

I guess it might be down to the sheer mass production allowing a much
better analysis of faults and allowing a lot of experimentation to iron
them out.

The Chinese manufacturers seem to have no qualms about incorporating
ventilation into their designs. The air usually flows in at the gaps
round the tube base and out through concentric vents at the back of the
cap.

 

[JohnR66]

The old bulb in the outdoor fixture was worn out, just blinks, no start.
Got a new bulb and installed. Blinked a couple times and started right up.
After a couple seconds it went out and would not restart. I put it in
another fixture. Heaters would light up orange (no arc forming across
either one). Very faint flicker of the bulb as starter clicked.

I was thinking the bulb lost some vacuum. I disassembled, check the
starter (okay) and connected the bulb leads to 600 volts to see if it
would light, It light up pinkish and flickered.

That was my lighting adventure for the day. Any thoughts on the very early
failure?

John

I guess these bulbs are bad. The next one lasted 10 days. They are 6500K
color temp and are labeled Bayco on the aluminum collar.
John

 

[TKM]

I guess these bulbs are bad. The next one lasted 10 days. They are 6500K
color temp and are labeled Bayco on the aluminum collar.
John

Sounds like they need to adjust their sealing fires. Too bad. All too
typical, though. On automated lamp manufacturing lines, if you make one bad
one, you often make many -- very fast. It takes on-line and fast-acting
test equipment to keep the quality high.

I watched an incandescent line making standard A-line lamps once which
operated at 3,000 per minute. The lamps were a blur going by; but every
process step had quality-control sensors and a big paddle that immediately
wacked lamps that didn't pass into the recycle bin. Not much broken glass
the day I was there; but I noticed that the recycle bins were full.

Terry McGowan