I picked up this portable 4AA powered F4T5 light.  The lamp current is for some reason DC and it leaves one end unlit, specifically the negative side/cathode.  


The current waveform clearly shows it is running in unipolar direction.  It is not the lamp as the "dark end" follows the socket even if I reverse the lamp.


I suspected the inverter was putting out DC, but when I checked the lamp voltage, the voltage swings to both polarity in almost identical magnitude, so i really can't figure out why the lamp current is unidirectional.  I already ruled out lamp being more conductive in one direction, because if I flip it around, it still stays dark on the same side socket.  
Voltage waveform:


If it was putting out a voltage with significant offset, I can see why this is happening, but since voltage is AC with almost symmetric magnitudes, I don't really understand! Anyone have a clue??

The oscilloscope was centered and set to DC coupling, so any offset would have showed up.  The center line didn't show in pic, so I drew it in. 



Power analyzer shows the lamp power is about one watt, though I don't really count on it being that accurate since the frequency is around 60 to 80KHz and its outside the range of instrument's accuracy (50KHz)

Because it's being fed DC power, it will be run on DC power

Read the thread again... the AA battery feeds inverter, which steps up and produces high frequency output.  The voltage across the lamp, which is shown, is clearly AC swinging approximately +/- 50v in both directions with the lamp running.

I'm going to have to play around with it using a resistor that mimicks the properties of a 4W T5 lamp.

It is because of "FlyBack" type of inverter, what generate HV pulses (during starting) in only one direction. So only in this direction it exceed the breakdown voltage and ignite the glow discharge. This then cause only one electrode to heat up (the cathode; by ion bombardment), so then it turn into arc (hot cathode mode) only in one direction. And as in the opposite direction still flow no current, there is nothing to heat up the other electrode, so this (rectification) mode become stable.
  You might do an experiment: Add heater source (separate 3V, e.g. a battery) to the "anode" filament for start. It would allow the current to run in both directions and most likely the rectifying effect disappear at all. But be aware, then this mean increase in input power, what might overload both inverter and batteries, so you might try it only for short time.

Bingo!

So the lamp was completely blocking current in one direction causing the DMM to effectively report current as DC.  Once it was replaced with a resistor, voltage turned into substantially asymmetric waveform. 

I think I need a better power analyzer.  The bandwidth on this one ain't cutting it. 

I think I need a better power analyzer.  The bandwidth on this one ain't cutting it. 

I think the best power analyzer for these things is the scope with a current probe (simplest is a current transformer with shunt resistor on the secondary, but this does not work from DC; Probes operating from DC are available, but are by far not cheap). Other option is using a scope with insulated channel inputs and use one to sense the voltage across a shunt resistor.

And if you want more automated analysis, you need a digital scope with the capability to store waveforms into [*.csv]-like file format and load it into Excell (or similar) SW. And there you might easily program analysis math you need...

Bingo!

So the lamp was completely blocking current in one direction causing the DMM to effectively report current as DC.  Once it was replaced with a resistor, voltage turned into substantially asymmetric waveform. 

I think I need a better power analyzer.  The bandwidth on this one ain't cutting it. 

Luminaire: I think your lamp rectifies the AC current pass through it to DC. This phenomenon may be happen with all fluorescent and low pressure sodium lamps or sometimes even HID lamps operated on ballasts from the mains supply, when only one electrode is depleted emitter and the other electrodes are good. Once rectification of the fluorescent lamp begins, it is flickering at the mains frequency and severally overdrives its ballast. Mercury migration may also results, This is the most destructive EOL mechanism of the lamp for its ballasts.