I have been trying to use a Sylvania (canada) f14t12 tube for a while. I'm running it from an 18/20w choke on 230v mains. The current is 370mA. This is in between the Philips (390-410mA) and the GE (14w at 40v, so 350mA) specifications.

But every time i fire it up, it first works fine for a few minutes, then starts swirling very annoyingly. It will do that for almost an hour, then it gradually gets less. I know that new tubes can flicker and swirl a bit until they've aged, but this thing just keeps on swirling.

Any advice? Is this just a bum tube or is something else wrong?

This lamp have 41V lamp voltage, which is considerably lower than most other T12 and T8 lamps.

According to this 1979 Westinghouse fluorescent lamp guide, it does show an operating current of 385mA and an arc voltage drop of 39V for a North American 14W T12 fluorescent tube:

https://www.scribd.com/document/110155293/Westinghouse-Guide-to-Fluorescent-Lamps-Brochure-1979

I'm unsure if this is an american or a european market tube. They were rarely used in Europe, i don't know if in this case there's much difference.

Either way, i'm pretty close to the right current specs wise.

I will disassemble the fixture and see if i can measure voltage drop over the tube, and perhaps see if i can figure out a way to push a bit more current through it. I don't expect much improvement with just a couple mA extra, but who knows.
The ballast does run annoyingly hot despite being within current spec. Cheap stuff.

I have been trying to use a Sylvania (canada) f14t12 tube for a while. I'm running it from an 18/20w choke on 230v mains. The current is 370mA. This is in between the Philips (390-410mA) and the GE (14w at 40v, so 350mA) specifications.

But every time i fire it up, it first works fine for a few minutes, then starts swirling very annoyingly. It will do that for almost an hour, then it gradually gets less. I know that new tubes can flicker and swirl a bit until they've aged, but this thing just keeps on swirling.

Any advice? Is this just a bum tube or is something else wrong?

Like it was said multiple times before, Joule law of P=U(RMS)*I(RMS) does not work for discharge lamps because waveforms are not sinewave and somewhat not in phase, too. So you can't deduct a current of 0.35A for 14W tube as just 14W divided 40V. In practice, it will be more, for example, ubiquitous 40W tube is typically specced at 103V RMS, 0.43A RMS, but 103*0.43 is 44.3W, while actual power rating is 40W or even 36W for T8 tubes.


Swirling is thought to be related to quite poorly understood interactions of the impurities in the discharge, and typically disappear after some time as impirities are decomposed or captured by electrodes. Time differs, may be some hours, or some tens or hundreds of hours. Only in some exceptionally badly made tubes it is persistent. Fun fact is that swirling mercury discharge is often desirable for special effect sign tubes, but it is practically impossible to be kept for a long time.

Then what kind of measurement and/or calculation do the folks who write down the specs in datasheets use? Philips explicitely states a mains current of 390mA for 1 tube running on a 125v system, or 410mA for 2 in series on a 220v choke. Power factor in the 0,4ish range. I'm taking these numbers from the 1965 world catalog.
Lamp voltage is not specified in that catalog.

I use a trms multimeter, it should account for the 'weird' wave shape that gas discharge tubes create - though not for phase shifts. I'm very aware of the effects of non-sinusoidal waveforms on measurements though dealing with separating real/apparent (translation) power from reactive power is still a bit foreign to me.

You need some proper watt meter like used in electric meters for example, that integrates momentary values of U*I at high enough rate over a line cycle. And it shall be robust enough to survive starter spikes of a few kV. That is not easy.

Practically, for fluorescents you are OK if RMS lamp current is within manufacturer specs, at least if lamp is healthy and it's voltage is within a reason.

Fluorescents are not very fussy about exact current, there is not much lifetime penalty if current is within +-10% or may be even +-20%, as cathode spot on the filament seems just to grow slightly larger at higher current.

If you like to play electronics a bit, from my experience with metering chips at work I'd say you can usually mod a surplus digital utility meter to make it fit to capture tube current, voltage and active power directly. You need to disconnect anything extra like capacitors and links to internal AC/DC supply from voltage input(s). To get an idea how metering chips work and what typical schematics is in use, look up datasheets and and application notes for example for ADE series at Analog Devices site. Just be careful not to touch voltage divider and current shunt circuits to keep the factory calibration.

 

Heck I had Canadian F20T12 on 120 volt  OCV that thing from new Swriled for 9 months of use before it settled.   What I observe the Swriling of the arc.  That the lamp is hunging
Trying  to find it sweet spot on the cathodes.  It will settle
 Down.   

  One thing due such low arc voltage of lamps and High OCV
 U might want to Try adding one more lamp in series.  It should
 Run about right  on 18/20 watt chokes as it tend to overdrives
 The lamps and your ballast would thank u for it too as it would run lot cooler too.

Since the current is 370mA, it is closer to *under*driving it than overdriving it going by Philips specs about mains input current without any PFC capacitor. I can lower the system input voltage to lower current, swirling continues. I gotta rewire my variac to put in *more* voltage to try it out on 400mA - but it's built into a giant Philips Natlab workbench, which itself is not able to be disassembled in its current location.

14w T12 tubes are extremely rare here - finding a second one is nigh impossible (and i just want to use a single one)

There's one thing that i don't think is very influential but who knows - the tube is almost in contact with the grounded fixture along its entire length, because the fixture was made for t8 lamps. I just cut it down to fit the 14w one.

Lateral electric field is in fact very influential on xenon discharges. Without lateral field you just getting a thin straight line bowing slightly upwards because of convection, adding some lateral field component - you have curious loops. But that is for HF discharge > 20KHz. Mercury discharges usually do not display such behavior, especially at 50Hz. But who knows in the case of your tube. Again, main suggestion is just to leave it alone for some time to clean up.

I will run it a bit more. If it's light enough in the room (like when i'm running the 125w SBMV from Aliexpress), the flickering from the swirling is easy to ignore. Just not when i'm actually working at the workbench that it lights.

The power calculation: The real lamp power is Current*Voltage*PowerFactor, the below u ity power factor is given by the shape mismatch between the almost sinewave current vs almost rectangular voltage, so it uses to be around 0.9.

The swirling uses to happen after the lamps are kept longer time off (like weeks or even months) and mainly in cold. After getting powered, it swirlin about an hour or so and then it settles and works normally, once it is being used normally.
To me it looks like the mercury condenses and maybe even settles by gravity, so after the lamp is ignited, the atmosphere inside has uneven mercury distribution. After power up the uneven atmosphere leads to the swirling (the clouds of higher Hg concentration?), until it gets mixed up back evenly.

No and no.

For theoretical calculations, discharge lamp voltage is classically modeled as trapezoidal.

Swirling is sure related to impurities. Organic usually. Incompletely burned binder, dust, vacuum pump oil in poorly maintained machinery. Not exclusive to mercury, also happens in Ne tubes at burn-in for example.

Stirations in cold lamps are a different thing!



 

Yeah when I first fired up some vintage tubes on trigger-start, they started to swirl and freaked me out.
The second time, I just let them run and the swirling went away.