Cool effect on F48T12CW: http://www.youtube.com/watch?v=-VwROCaIPrM

Discharging of the large buffer capacitor in an electronic ballast. And the ballast has very poor line regulation by itself (but may get the stabilzed voltage from the PFC stage), so reduce the power when the voltage on the cap goes down, so it take so long to discharge it fully.
Nice effect...

At the radio station I work at the lamps in the production room fade out and glow dimly on the ends, and pulse to the music ( AM radio station) and the preheat 40 watt dual fixture does the same thing....hot RF field...the tower is just outside the building. The control room is shielded and grounded and doesn't do this.

Powell

I don't know if I'd want to spend any time in an RF field that was strong enough to make lamps glow!

Depends on frequency, as the AM transmitter wavelength is hundreds of metres, so most of the pickup is via the wiring, whilst you will not get much just because you are so small that the transferred power is less than that you get from the light from the lamps.

It is true, but the allegedly dangerous power level limit "picked up" by the body is quite low, mainly on lower frequencies (with deep penetration). Higher frequencies are easier to pick up, but do no penetrate so deep, so affect only surface organs. And from these only eyes are sensitive (they are sensitive to temperature changes, but at the same time they do not have the blood flow to cool them down); There are known some skin burns, but these injuries originate from accidents, when wrong switch was used when switching off necessary antenna sections, so maintenance crew climbed directly on the energized section.
So if it is able to light the tubes with few meters of wire, the field is already quite high.
But on the other hand these unprotected areas are not the place, where operators stay majority of their shift, so the "dose level" they get would not be as high due to the short exposure time...

Correct Medved, you can get badly burnt if you get in the beam of a FM or cellular transmitter, as the power density there is very high, and you will be a significant portion of a wavelength and thus will absorb considerable power.

An AM transmitter is dangerous for a different reason, in that most of them do not isolate the antenna from the final anode of the transmitter, and thus the whole antenna floats at 1-2 kV above ground, with an insulated base and insulated links on the guy wires. This is simply because the current is very high, and having a isolating capacitor would be a very expensive and lossy item, and you need to insulate the bottom of the antenna anyway, lightning protection being by spark gaps and inline inductors that form part of the output tuning and harmonic traps. Makes for fun to arrange a supply for the aircraft beacon lamps, you have to use a transformer specially made for the application, with a very large separation between the secondary and the core, and a shielded primary winding, with RF filters on input and output. A FM tower is solidly grounded with multiple cables to a ground plane.

Not that I would like to climb one, some are close to a kilometer high.

I have a F42T6 that does that and the 8fts at my work do that as well.

try 20-50kv on some.
my 2x 4-1000a amp has 7.5k on the plates.

I have a two-lamp F40T12 rapid-start wraparound where I replaced the ballast with an electronic Advance version, and it also does that fade-out thing when turned off. Otherwise it behaves almost the same as a magnetic rapid start.

can be interferences made by amplifier?
If there is an inductive load (transformer, rf suppression inductor, etc), can make voltage spikes on line when load suddently decrease.

Try also to build an simple rf meter (example: http://forum.meteotriveneto.it/viewtopic.php?f=4&t=19974
it can be very useful to detect what kind interference you are seeing on FL tubes.

The F96T8 dual lamp fixture with the solid state ballast does this. And a dual tube 4-1000 amplifier? What is that used for?

Powell W4OPW