Can a PL quad pin CFL be driven via a capacitive dropper ballast? And if so, 1) how do I size the metal film capacitor and the series resistor for the tube? 2) What do I select/size for a semiconductor starter? 3) How much will this effect life and lumen output?

Reason I ask besides having the material around (13 watt quads and a few CFLs with failed ballasts) is that I've seen this done with night lights years back. A capacitor and resistor in series with the tube would act like a ballast, and latter versions had a triac (I'm guessing) as the starter. Here is one example but with a glow starter:

https://www.youtube.com/watch?v=NtzAVlucpqQ   


Doable?

I'd guess it cant because the capacitor does not provide restarting impulse after zero crossing. In the nightlight lamp they use starting probes (similar to Mercury lamp) that help in the reignition, but there arent any in a PL..

Life is not great with capacitive systems, especially with purely capacitive ones



Capacitor is calculated very approximately as :

Vcap [ V ] = sqrt ( Vline - Vlamp ). For lamps with low arc voltage you can skip this calculation and assume Vcap = Vline

Xcap [ Ohm ] = Vcap [ V ] / Ilamp [ A ]

Cap [ F ] = 1 / ( 2pi f [ Hz ] Xcap [ Ohm ] )



So for example for 180mA (PL 7W) on 230V it would be :

Vcap = 230V

Xcap = 230 / 0.18 = 1278 Ohm

Cap = 1 / ( 2pi * 50 * 1278 ) = 2.5e-6 F =
2.5uF



Line voltage is not clean, it contains harmonics. The harmonic content may be small but due to its higher frequency it may increase the current significantly over whats calculated, so in practice it may be good to use smaller capacitor to prevent overdriving

The bottom line is, capacitive droppers are not suitable for ballasting discharge lamps

Ash, huge help, thanks! I may make this partially resitive by adding a resistor in series with the cap.


In regards to the starting probes- I could be wrong- but the 5 inch night lights I took to bits seemed to be much like regular linear T5s with two pins at each end and a filament cathode. It had a transistor looking device as the starter. The one inch versions had the probes.

Also including this for future use:

http://keisan.casio.com/exec/system/1258032632

In theory I could add a 5 ohm resistor in series with little impedance change, or a slightly larger cap with a higher impedance resistor. As I am reading this limits inrush and makes the lamp more stable.   

With resistros the "Vcap = Vline" approximation does not work anymore, so you really have to calculate Vdrop. The resistor Vdrop is in Phase with the Arc, and in 90deg out pf Phase with the capacitor

Mind power dissipation ! P = I^2 R. In your case, 5 Ohm resistor would dissipate 1/6W, so 1/4W resistor would be hot and 1/2W would be very warm. With 100 Ohm (which is just enough to have any significant effect), the power would already be over 3W..

But I could still model both in series with the above calculator to obtain equivalent impedances? Good point on the dissipation, will have to keep an eye on that.   

Question, I am a tad confused where you derived 180ma. At 7 watts @230 volts 30.5 ma. Of course since the circuit is reactive (guessing) not all the current flowing will be converted to watts, so I would guess you would need 60ma of through current? 180ma seems like a really low power factor. Just confused atm. My apologies.

You could model them in series and get impedance. But you really are interested in the current, which is V( cap + resistor ) / Impedance

And V( cap + resistor ) is the thing you dont know. As long as there is only cap (or the resistor is insignificant like 5 Ohm) and the lamp's arc voltage is very low, you can very roughly assume Vcap = Vline. But with a more significant lamp arc voltage or resistor value, you cant do this shortcut

180mA is approximately the rated current of a 7W PL-S lamp. Provided as example. The arc V for such lamp is somewhere around 40..50 V

The power factor of a capacitor plugged directly into AC power (without lamp or resistor) is 0, it draws current but does not draw any power. When something that draws power is in series, but most of the voltage drops on the capacitor, the power factor will indeed be very low

The 7W PL is meant to work with a choke, with power factor that can be as low as 0.25

That makes perfect sense now in regards to voltage drop across the capacitor and power factor! Last 2 questions- and I'm sure its obvious but I am not a fluorescent expert... where do I find the rated current for any particular bulb? Can I just take the current off an equivalent magnetic choke and assume that as being correct?

https://5fc98fa113f6897cea53-06dfa63be377ed632ae798753ae0fb3f.ssl.cf2.rackcdn.com/product_images/files/000/094/769/legacy_product_detail_large/data?1431466985


And I notice that when 13 watt US lamps are driven its always with a 120 volt source, however as long as I size the capacitor to maintain 30-40 operating volts, I should be fine with a 240 volt source?      

Dont limit yourself on questions, we are pleased to answer them

You could take the rating off a Magnetic ballast or sometimes from the manufacturer's specs. Few manufacturers have a concentrated datasheet with various lamp data



The general rule (which can be broken sometimes) is, that to run a lamp with a choke you need a voltage source approximately double the voltage of the lamp, or higher. So the lamps with ~40..60V arc (up to 9W PL-S, 18W PL-L, 18W T8, 20W T12) can work with a choke at atleast 120V, while the lamps with ~95..110V arc (11W PL-S, most PL-C/PL-L/2D, 36..58W T8, 40..80W T12) can work with a choke at atleast 230V. Any lamp can work with a choke at higher voltages (choek value will be different)

The US 13W will need a suitable choke to work on 240V. I think 2 of the 13W 120V chokes in series may be close enough. you can likely also add a 2nd 13W PL in series i.e. run 2 PL 13W + 2 chokes all in series on 240V

Europe have a 13W PL-C that have higher arc voltage and lower current than the 13W PL-S

Ahhh, I see it now! I know from HID lamps that is certainly the golden rule, ie a 35-150W 55 volt HPS arc tubes run on a 120 volt choke, 150-400W 100 volt HPS lamps run on a 208 or 240 volt supply, 1000W 250 volt on a 480 volt choke- same with Mercury vapor and some pulse start metal Halides.


And in terms of limiting myself- I shall pass One more thing: What can I use as an electronic starter? A PTC? Triac? I've seen a transistor like device used to start pre-heat tubes, but unsure what its called or its forward voltage.

Most likely a GTO. First it is switched on for full time and acts as a short for preheating. Then it is switched on from the beginning of the half cycle and off at a time when current is flowing, this makes the inductive pulse to start the lamp

The forward voltage of them is insignificant, the reverse gotta be somewhere in the >1500V atleast



With capacitor you can't get a starting pulse. You can only preheat, and if you fire the thyristor at different times you'll vary the RMS preheating current, nothing more

Maybe some lamps will be ready to light up on line voltage after preheating

Maybe you can do with a GTO and some small value choke just for the pulse (you account for it in the capacitor impedance, but the capacitor is still the main ballast). I guess thats how it is done in the US cheap 4ft shoplights where the ballasts are in the end panels of the luminaire

With a capacitive dropper you may take a diode with resistor (something with insignificant losses at the lamp arc voltage once the lamp starts) in series, and connect that parallel to the lamp. When the lamp is not lit yet, it will charge the capacitor to the peak mains voltage, which will make double of the mains inthe following half cycle. TnT voltage will be then available to ignite the lamp...
And parallel to that you may connect a series inductance and triac, with firing circuit designed so it will trigger near double of the mains voltage (a resistive divider with a diac). This trigger should incorporate some small delay (small capacitor parallel to the bottom resistor), so it will give the lamp some time to ignite if it wanted to. Plus thecapacitor will be the energy sourcefor thetriac gate...
This arrangement will make a half cycle just after the triac fires, boosting the voltage across the capacitor. Plus thecurrent pulse will provide cathode heating...
If I remember well, something like that was in the ballast of the first OSRAM circolux CFLs (230V, FC6T9 or FC8T9 tube; with the FC6 it consumed  about 20W, with FC8 about 26W, so about 4W losses in tbe smoothing resistor in series with the main ballasting capacitor).

Anyone have a schematic of this lamp? Never seen of heard of such a ballast, but if tried on a consumer scale might be worth looking at.