Fluorescent starters

[Oldcodger]

Could be classed as homebrew, but as I'm only interested in the internals of starters, I'll post it here.
Wired up my lightbox with one ballast, two tubes and two starters as per diagram on the makers page, but no light. Maker's diagram shows circuit for two tubes ,with this ballast .
TBH- I've never bothered looking into how these lights function, other than the ballast is there to limit current when tubes strike.
Looking on line, starters are said to be closed switches which open after a set time, so I'd expect one to be S/C with a meter , but on mine ,all I see is ~5nF . ( something like 30Mohm at 50HZ)I do see a large ( >>Mohm slowly going OC) resistance at first.
Anyone shed some light ?

[dragonser]

HI,
an interesting question.
I may be wrong on some / most of this ...
as far as I know the early non electronic starter has a neon bulb with a heat sensitive strip in it. when the voltage is high enough the neon strikes and heats up the strip which then closes ... then the neon is extingushed and the strip opens giving a surge in the current which starts the lamp.
I think the ballast is there both to drop the voltage for the tube, and act as an inductance to produce a voltage spike to start the tube conducting.
I don't know how the electronic starters work !
there are different types of starters available , some are for a single tube, others for tubes in series....
I think there should be continuity between the pins on the same ends on the tube, as they are the heater elements.
any corrections / comments welcome.
regards Peter B

[TonyDuell]

Most fluorescent starters are the so-called 'glow' type. They contain a small discharge lamp (often filled with argon, although some are neon) with bimetalic strips for the electrodes. There is an supressor capacitor in parallel with it which is probably the 5nF you are reading.

At power-on there is not enough voltage across the fluorescent tube for that to strike, but the lamp in the starter does. Its electrodes get hot, and bend (being bimetallic strips) and touch, effectively making the starter a short circuit. Current now flows through the ballast (traditionally an inductor) and the filaments at the ends of the fluorescent tube. Those get hot of course. But the fact that the starter is now a short circuit means that there is no glow discharge in it and it cools down. The electrodes separate, interrupting the circuit. You now get a large back emf from the ballast which strikes the discharge in the fluorescent tube between the still-hot elecrodes.

Once it has all got going, the voltage across the fluorescent tube is too low for the glow discharge in the starter to strike.

My experience is that if you have 2 tubes in series on one balast you really need special starters. Are you using those? What circuit are you using?

[high_vacuum_house]

This operation that is described above is correct.

The 5nF capacitance you see is a small film capacitor wired in parallel with the neon glow switch to minimise arcing in the switch.

Check the rating on the starter itself. There are different variants for different wattage tubes.

I have made this mistake before and after fitting a new tube and starter could not work out why it still did not work. I had fitted a starter intended for high wattage (i.e. long length) tubes whilst the fitting I fitted it to was for short tubes.

Christopher Capener

[Oldcodger]

Starters = 4-65 W, bog standard . Tubes= 8w UV tubes. Ballast, one picked from RS as suitable for two tubes in series.
Tony- circuit is Mains- heater in tube 1to heater in other end ,via starter to tube 2, then via starter to tube 2 other end and return to N.

[matthewhouse]

A series starter is normally used in such an arrangement: S2 starter.

[Oldcodger]

matthewhouse - circuut diagram for series starter , would be nice.

[Boater Sam]

Page 109 of

https://www.osram.com/.../single-and...english-part-1

URL won't work but if you search you will find, google is your friend.

[matthewhouse]

See attachment.

[Mooly]

Way back in the early 80's (or even very late 70's) Elektor published a solid state starter circuit that could be built into a standard starter case. I built one and it was in use for a couple of decades. Just a triac and an R/C network on the gate using a couple of high value resistors and a 0.1uf cap as I recall. Pretty instant starting with it.

To check for starter volts, short one out then try, one lamp may light.

[Radio Wrangler]

Electronic starters are off-the-shelf items, though a bit dearer than the glow ones. I think they're all intended for the plain circuit, not the series one, so you might as well have two smaller ballasts if series mode would have taken two starters.

David

[bobskie]

This may be a bit of a hijack but I think it's relevant to the question. What would cause a starter to simply glow purple and the tube to not start? I have tried 2 starters and both exhibited the same symptoms. It is a nice glow that comes off them though

[Boater Sam]

If the tube ends are not preheating as well, the contacts in the starter are stuck open.
Try dropping it on the floor, sometimes it wakens them up!

[julie_m]

The starter switch is open-circuit to begin with. When first switched on, the voltage is enough to establish a dicharge. Then the bi-metallic contacts close, stopping the discharge. The bi-metallic contacts open. The sudden interruption of current through the ballast coil creates a kickback voltage effectively in series with the mains. If the voltage is enough to strike the fluorescent tube, then the voltage across the starter will be too small to start a discharge there. Otherwise, the cycle just repeats until the fluorescent tube starts.

[Oldcodger]

Originally Posted by TonyDuell

My experience is that if you have 2 tubes in series on one balast you really need special starters. Are you using those? What circuit are you using?

Thanks ,Tony ,and all who contributed. Penny didn't drop, till I looked at the circuit ( as Mathew house posted). Self bum kicking time. Of course 220-240 starters won't work - ballast at startup has no effect ,so mains is divided between two tubes ( aka 120 each), so I needed a series ( or 110-130 v starter).
Fitted and it works. Always helps- when you meet a problem ask for help, and somewhere out of the blue there's an answer .
bobski- have a look at https://www.osram.com/media/resource...---starter.pdf

Now for the timer .

Admin - problem solved , post can be closed.

Ah so, problem solved. That's why I love this forum, help is available from more knowledgeable people than me.

[Oldcodger]

And Merlin, as I said in thanks- sometimes ( as I often found out on the bench) , when stuck- it's often worth bouncing ideas off folks. Too often we run up and down the same path, missing a side path, till someone points it out. Tony's suggestion of "special "starters did that, and made me look at the problem from another POV. (I needed lower voltage starters).
TBH- Fluorescent lamps are not something I've dabbled with. So any help was welcome.

[julie_m]

Gah, I just re-read my earlier posting and realised I missed a whole paragraph .....

With two tubes in series, one starter usually opens before the other. It still creates an inductive kick across the ends of its tube, since the other starter is still closed. When the other starter's contacts have cooled down, they open and there is a second inductive kickback, which ignites the other tube. Or, if the two starters open within nanoseconds of each other, the voltage appears across both tubes in series and may be enough to start them both conducting -- if not, they will close again and restart the cycle.

[bobskie]

Originally Posted by Oldcodger

bobski- have a look at https://www.osram.com/media/resource...---starter.pdf

Hi thanks for the link, nothing about them glowing but I reckon it's just a dead tube. Mine is a homemade lamp too but I didn't make it, it's in my hall cupboard. The wiring is rather open and dangerous to be touched but it looks safe enough from a technical point of view! I'll just get a whole new light to be safe though I think I have one lying around somewhere actually. Good luck with the project.