Reverse engineering an electronic fluorescent starter.

[@mauriceappleby6667]

Hi Clive. I was the original designer of the Pulsestarter in conjunction with Texas Instruments who were responsible for the original main thyristor power device. The data sheet you reference was the original design which was manufactured over a number of years by a company called Arlen in very large quantities. The original concept was based on the fact that preheating the tube cathodes (proposed by a very clever guy called Mike Lester) would increase the expected life of a flourescent tube. Resultant testing proved the fact so much that some extended life testing we had running was eventually turned off after 1 million successful strikes on, I think from memory, a batch of 58W T5 tubes. The idea was that after the preheat cycle, a number of pulses would be applied to the tube as the small thyristor would turn on but not have enough holding current to stay conducting at first but would eventually latch after a number of cycles.This meant that you were always attempting striking with hot cathodes and, unlike a standard glow starter, you avoided overheating and possible fire issues with a dead tube trying to keep on striking. BTW, love your videos which I only found recently but now watch often.Keep up the good work.....and the gin, which is one of my favourite tipples :-).

[@bigclivedotcom]

Good job. I remember the Arlen brand when they first appeared. Did they feature on Tomorrow's World?

[@melplishka5978]

Nice. Ty for your input Maurice.

[@pemac8856]

Hi Maurice. Am I right in assuming that the TN22 keeps getting retriggered, but driven off by the gate being pulled below the cathode, or is DiodeGoneWide's answer correct? And also, what is the reason for the resistor in series with the timing capacitor? Is it just to provide a bit of ripple to give the transistor, or the P0124 a bit of a nudge on? Another possibility I thought of is that the ripple from it peaks as the current peaks and makes the darlington drive the TN22 off then to get a good discharge from the inductor. Is that true? But that would not explain why they had it when they had a C106D, which is not a GTO, so maybe it helped trigger the P0124, and now unfires the TN22 at maximum current. I'm curious, please write me a reply. Cheers!

[@W4BIN]

I always thought that the General Electric research labs was responsible for the development of the SCR. Ron W4BIN

[@Conservator.]

@@pemac8856Hi, DiodeGoneWild has done another video on fluorescent lamp starters: kzbin.info/www/bejne/gmHCn313jNmsZ68

[@DiodeGoneWild]

Once the 47uF capacitor is charged, the transistor turns on and stays on. It steals most of the current from the gate of the thyristor. A low current flows gate-cathode through the thyristor, it is not enough to turn on the thyristor, but it is enough to keep the transistor on. And as the transistor is on, it still eats most of the current, so that there's never enough current for the gate to turn on, but some little current to keep the transistor on.

[@DiodeGoneWild]

and of course the gate-cathode junction of the thyristor acts as a diode in forward direction.

[@stargazer7644]

The SCR gate is shorted to ground when the transistor is on (that's why the transistor is there). How can it get power from the SCR to keep the transistor on if the gate is shorted? How can the cathode have any voltage on it?

[@DiodeGoneWild]

The transistor is not fully saturated. Its voltage drop is enough to allow tiny current flowing through the gate-cathode junction and then through the base-emitter junction. A negative feedback keeps this current stable.

[@justpaulo]

Which I believe it raises the question on why the SCR even turns ON the 1st time ? In other words, even on the 1st time, as the 50V 1uF cap voltage increases, the gate-cathode current increases, raising the base-emitter voltage of the "NPN" , which could eventually start stealing enough current to prevent the 1uF cap voltage to further increase, stopping the SCR of ever turning ON. My guess is that the 47uF comes to the rescue here, when you 1st turn ON the light because it is fully discharged. The times constants are fine tuned such that the SCR turns ON way before the "NPN" is even close to go ON. But once the SCR triggers now there is plenty of current to complete charging the 47uF at the fastest pace possible. On another topic, nowhere I have seen that a SCR, once ON, can be turned OFF by playing with the gate terminal. Usually they just talk about the hold current. Having said that I am convinced it is the "NPN" that turns OFF the SCR, though I am not sure how.

[@joop2295]

Yes, DiodeGoneWild is right. There only has to flow a few micro-amps through the gate into the darlington's base to keep it on, which is not enough the turn on the thyristor.

[@revbobuk]

The old starters were brilliant - use a couple in series with small coloured lamps - orange and red - and you could create a very realistic bonfire stage effect.Lovely random bi-colour flicker.

[@H3wastooshort]

oh thats very cool

[@operator8014]

Oh dear, if Clive doesn't immediately understand it, it must be deep magic indeed. This makes me feel much better, knowing that I'm not the only one who gets bewildered by seemingly simple circuits.

[@Spanaussie]

You Scotts speak funny... A milliamp is pronounced as written.. Milli amp... Not millamp... 😉

[@operator8014]

+Michael De Angel - Buddy, what in the world are you babbling about?

[@slawor4]

Brad Gefroh Maybe he responded to the wrong comment

[@operator8014]

+That Guy - Haha, I think he meant to reply to the video in general, but replied to ME by mistake.

[@slawor4]

Brad Gefroh That's also a possibility

[@themaritimegirl]

The Pulsestarter doesn't simply open when it's finished preheating - it opens and closes very rapidly over a half-second period, or until one of the openings causes the lamp to start. Also, the diode D1 (at least in my unit) is a peculiar one, with a ~3V forward voltage. This might help with some of your confusion with the circuit.

[@cup_and_cone]

The company I worked for years ago illuminated their stores with skylights and fluorescent lights. They got swindled into light sensing controllers that would shut the lights off when the sun was bright enough and would turn the lights on when sun went behind clouds. Well on a cloudy day, the lights can be turning on and off every couple minutes. Needless to say they were going through bulbs/starter/ballasts like crazy...and to top it off, the energy bill wasn't really any cheaper. If you live in the US, you can still see these at Home Depots, some grocery stores, maybe even some Walmarts...and you'll see half the tubes out.

[@brianartillery]

Thank you, Clive - I've always wondered how those starters worked, and you also settled it for me as to what makes that really annoying, and sometimes quite loud 'Klak-klak' noise. My local Co-Op had a particularly irritating one the other week. Cheers.

[@Bodragon]

I've watched many, many KZbin videos all claiming to explain how the fluorescent starter works. This is the only one that explains it fully, clearly and correctly. I especially like the bit about not connecting a starter directly to the mains. Yes, I have done this too! And yes! A very large pop it did make too! Thanks Clive. This video made my day.

[@RaunienTheFirst]

I'm still not fully understanding why. Is it because they're only expecting the full mains voltage for a fraction of a second? But then what if the tube fails to strike? What stops the circuitry from constantly attempting to strike it, and why, given that an unlit tube doesn't drop the voltage, does the starter not simply explode?

[@fillg]

I haven't seen one of those for 20 years or more. I didn't realize they are still used elsewhere. I'm frequently amazed at how different things are here vs over there. I used to replace lots of ballasts, both 120 V and 277 V depending on which building I'm at, but now whenever a ballast goes out it's cheaper to just rip it all out and rewire the fixture with line voltage going directly to the bulbs, then pop in a couple LED tubes. Hopefully by the time those need replaced I'll be retired.

[@russellhltn1396]

I hope you add a sticker to warn against installing a regular florescent, or someone could have a nasty surprise.

[@fillg]

Oh yeah! Every time I buy the bulbs they give me a stack of warning stickers and I definitely put them in each fixture I convert.

[@uK8cvPAq]

I think they're mostly legacy installations these days or fixtures that refuse to die in kitchens and garages.

[@MichaelBeeny]

In my early days I just loved florescent lights but hatted the turn on flashing. One day I was looking through the RS catalogue (as one did) and noticed the electronic starters. I replaced ever one I had and just loved them. It only ever struck once after about 2/3 seconds, noiseless and zero flicker. I still have a few in my bits box today. No idea why I keep them as I am 100% LED these days.

[@torqued666]

So THAT's what causes the ends of the tubes to turn dark! Thanks, BC!

[@jlucasound]

I never knew. It is like there is "welding" going on in there, when there shouldn't be.

[@paulevans7876]

The old starters were abysmal! I once lived 50 metres from a 'corner shop' [Spar]. One morning I go on the air on 20m with my ham station and there are ENORMOUS clicks about every 0.7s across (as it turns out) every band. I scratched my head for a few hours (at work) and finally concluded it must be a light. Sony ICF7600D in hand I find a huge signal when outside the shop. Look in and presto! One of the tubes is flickering on and off. I went in and pointed out what was happening and luckily they changed it (the tube) right away [it helped that it even showed on TVs and I pointed out I would tell the whole neighbourhood that their TVs were being 'attacked' by the local shop!!!!]. Luckily, after a couple of occurances of this they fitted electronic starters and the problem more or less never re-occurred.

[@guyfawkes9951]

I learn something from every video you do, BC. Thanks for increasing my knowledge.

[@DoRC]

I think the regular paper actually looks better in the camera. no reflections .

[@twocvbloke]

I have an RS branded electronic starter, had it for years, and it only stopped being used when I replaced the fluorescent tube in the kitchen with an LED tube, never seemed to fail unlike the traditional starters... :) And as for traditional starters, I recall seeing them used in someone's halloween project to create a "faulty wiring" effect, which you recreated in this video, quite impressive and a simple circuit too... :)

[@scottfirman]

There are floresent light fixtures rated for colder temps. Here in Northern Mi,we have to make sure all the fixtures are rated for cold temps. I remember having the older ones in our barn and when you turned them on in minus 20 degrees,they would only flicker off and on. It was way too cold for the preheater to get the light to florescent. We always had to make sure we had regular clamp type lights in the work areas. Problem being it was so damn cold your fingers would freeze to the tools when working on tractors,snowmobiles or our trucks. My dad put block heaters in all of our vehicles. If you forgot to plug in the vehicle,you had to put a clamp light under the hood for an hour,usually down by the starter. Before the block heater it was a 100 watt light bulb in a work light with a wool blanket over the engine. I remember frost would build up in the carburetor, then it would suck it in to the intake. You had to use ether to get the damn thing started. Especially the ford tractor.

[@mitch3064]

I was finally able to see what the name of that switch assembly was, and then you said the name of it. I have been wondering for some time what it was but never heard you put a name to it in the videos I have watched. Thank you Clive.

[@hadibq]

Great topic, these fluo lamps. They were a big mystery to me when I was a little kid and no one around to explain to me how they work. Unlike the so intuitive incandescent light bulbs😊

[@ciprianwinerElectronicManiac]

Back in the day I was using starters for making Christmas lights blink. You had to try quite a few starters until you got the desired effect and as the starter ages the frequency also change. It was very exciting. :D

[@mrfrog8502]

11:15 Technically speaking there's a different terminology used depending on application. Choke - to block RF and Ballast - to limit AC current, also typically ballast would be much bigger and have higher inductance.

[@Pithead]

I used to like charging those little caps up with 500 volts (with an insulation tester) and throwing them to people while saying 'here catch'. This was back in the 80's when you were still allowed to do bad things to people at work.

[@Firecul]

Rabbi Zyklon Brausebadstein when I first heard about that I was shocked. Sorry I had to, I used to heard those stories. Unfortunately most people are already cautious when I throw things at them.

[@stevejones8665]

Rabbi Zyklon Brausebadstein or get them to hold on to your wind up 1000V Megger LORL Great fun....

[@Miata822]

Yep. Happened to me my first day at electronics school.

[@RS-Amsterdam]

Or asking a fellow student to hold a granate and walk away with the pin. And then after 20 times you had private lessons. Yes we had a tough school in Aghanistan

[@CAIDENROX]

In my electronics class, everyone just melted blobs of solder and catapulted them from the irons. We had no way of charging any capacitors because our teacher knew we would try that lol.

[@timh8273]

This video just answered every question i had when i was a kid

[@phorzer32]

I love this electronic Starters. They are so reliable and the tubes lasting for ever

[@Roy_Tellason]

I had heard quite some time ago that flourescent tubes had an expected lifetime that was based on the number of times you started it up, and it was never quite clear to me why this should be, until you explained it in this video. I'm guessing that this is the reason that a lot of commercial properties like large office buildings are left illuminated all night long in places like New York City, as opposed to turning the lights off. Their cost for the electricity would have been less than having to replace the bulbs more often, I guess.

[@johnpossum556]

From what I gather the resistor below the transistor's base forms the bottom of a voltage divider which must account for why it works. I am more shocked to see an actual circuit with discrete components in that starter case than anything. All the ones I have torn open never had anything that complex in them. I think I even have an old one down the basement that has a removable starter that is resettable. I got that FL unit from a city worker where they used to keep the snowplow trucks -- apparently it is extra good at staying lit in our frigid sub zero temps. Good vid, Clive.

[@theirisheditor]

The electronic fluorescent starter could work well as an April fool's - Wire it in series to a table light (assuming incandescent) and set up a candid camera. When the victim switches on the light, it goes out, leading them to replacing the "blown" bulb and of course the same with each spare bulb they try. I have wondered how they claim to increase tube life and indeed the longer warm-up period and single strike explains a lot. Another clear benefit is that the cutting off also saves electricity on failed tubes. With a standard starter, the ballast is effectively shorted with endless start attempts, consuming more power than with a working fluorescent tube.

[@simonparkinson1053]

Seán Byrne They are very effective at increasing tube life because of the mandatory preheat. Where circular fluorescent tubes are used in ceiling lamps, they do not have guard rings on the cathodes, so the tube ends blacken very quickly. But with the electronic starter they hardly blacken at all. On many tubes started by these, the phosphors in the tube wear out before the cathodes!

[@dayleedwards3521]

Hello Clive, I enjoy the content of your videos. There is a large AC ripple content appearing on the bridge output , this being unsmoothed ., the positive going pulses appear at the transistor base via the 47 mfd and 56k, pumping up to Vbe, thus keeping the transistor on after it initially triggers. It cannot turn on prior to the initial trigger because the 47mfd/56k time constant is longer than the 15k/1mfd. Once the tube fires, there is no voltage available through the zener, the SCR remains off. If the tube doesnt fire, the ripple voltage continuously removes gate voltage via the transistor shorting the 1 mfd until the power is cycled. Its a bit of a trick circuit to be sure......

[@davadoff]

Ah thank you! I think I get it now. I watched the video and read the comments explaining it further and I didn’t honestly understand it until I read your comment. I felt like I was the only person that didn’t get it still after reading the further explanations. I can’t believe I am the first person to give you a thumbs up 👍

[@davadoff]

If the 24k resistor was lower value then that would steal the current from flowing into the base and cause the transistor to turn off, thus allow multiple attempts at starting the tube?

[@SuperBrainAK]

So THATS what those are! i saw a bunch scattered in the bed of a truck from this pawn shop and i couldnt figure out what they were. Thanks for this video (literally within a week of seeing them for the first time lol)

[@lumpyfishgravy]

Believe it or not my _other_ boss had a hand in this one. I say a hand, the design was done but needed tweaking. We got involved because they were being produced locally at the time by BG Electrical. Now my boss is very smart, and he struggled with it. It's a good demonstration of a design that's almost too clever for its own good. Boss is on holiday right now and the job precedes electronic records so I don't know how much we have but I'll see what I can find.

[@SqueeDee]

That was super informative Clive. I wanted to know how starters work since I was a kid but never got around to reading up on them.

[@bigclivedotcom]

Have you seen this video? It covers the technology in fluorescent lights. kzbin.info/www/bejne/p33EdWRnga6HmLM

[@matthewday7565]

Finally understood the reason for the "high holding current"... When the drive to the thyristor is removed, it WON'T wait for the zero crossing, but will cease conducting while there is still current to provide a kick.

[@Seegalgalguntijak]

I've seen these starters in my childhood in the 80s when one broke. They all looked like the white plastic housings, but they did blink the lamps several times. Also after they were replaced, I haven't seen fluorescent tubes go on without prior blinking until the end of the 90s or even early 2000s.

[@simont3686]

6:58 Hehe. Memories. I when I was 12 I think, I hooked a little bit bigger neon glow lamp up to the mains without resistor. Man. I got scared. Little glass shards all ober the workbench and floor. I should have known what I was doing. Thank you for sharing your knowledge and good night from Germany.

[@williefleete]

I think the reason it only tries once is the 47 uF cap on the transistor via the resistor, the 82 and 24k providing a discharge path when power is turned off. Once that cap charges the current flows through the transistor instead and not providing a gate cathode drop high enough to trigger the SCR and perhaps leakage from Collector base when the transistor is on is enough to keep the cap charged.

[@ciprianwinerElectronicManiac]

It's obviosly that the 47uF capacitor that charges a few seconds after the thyristor is triggered then the darlington transistor will shunt the thyristor's gate, since the transistor is a darlington it uses very little current so the capacitor should discharge slowly, I suspect that after the capacitor is depleted it should retrigger the thyristor again. I hope i got it right. Cheers :)

[@Agent24Electronics]

That's how I interpret it as well. I can't see how else it would work.

[@stevejones8665]

ciprianwiner yes that's right the Darlington could be held off from retriggering the thyristor by a tiny current even putting your finger on the open Base would hold it off so i guess it would take weeks for the capacitor to lose every last nA. ....

[@stargazer7644]

I don't see how that would work. The 47µF cap through a 56K resistor has an RC time constant of only 2.6 seconds - and it was only charged to about 0.7V to begin with. For that matter, this MPSA14 darlington has a Vbe of 2V, how does it ever conduct in the first place with the base shunted by the diode to 0.7V? Very odd indeed.

[@ciprianwinerElectronicManiac]

That had me puzzled too but it was the only explaination i got, maybe the diode is not a regular diode? or Clive had it draw incorrectly, maybe Clive could check the forward drop across that diode? Who knows, maybe it's even an actual zener diode and he didn't check it. Cheers :)

[@matthewday7565]

I'm baffled how the transistor gets turned on, since a darlington has 2x Vbe, and the drive to it is fed by a single diode drop - unless that is not an ordinary diode! The "R4" is defined as the reset, I'm guessing it's meant to shunt the leakage to lower the voltage to the point where the switch off SCR / transistor unlatches. I guess the value is too high, with a lower resistor there, it may retrigger.

[@bigclivedotcom]

+Matthew Day That did go through my mind too. I may desolder one end to see if it's a different diode.

[@matthewday7565]

Just read about another system, the half a second "fast starter", though it sounds somewhat abusive of the system. users.tpg.com.au/pschamb/light.html Described as using "DC heating" and producing a characteristic buzz and magnetic disturbance, I instantly figured out the trick - while there would be a 50% disadvantage in only using half wave control, more power would reach the heaters as the ballast would present only resistance. The other half cycle could be used to continuously detect tube firing. It would be a rather brutal system though, and I can't see it being good for the life of tube or ballast

[@mrtl5676]

What an amazing video sharing! ..... I wish my science physics teacher able to explain in the same manner (during my school days)

[@bigclivedotcom]

One of the joys of KZbin is that there are a lot of really good educational channels which are not only enjoyable to watch, but are free too. It's what the Internet is for.

[@amaze11]

This reminds me the number of June 1982 of elektor magazine. They (and me!) used a TIC106D and the circuit still works today.

[@saifuddinlakdawala8988]

The MPSA14 is a germanium transistor with junction voltage of 0.15 to 0.2v. Once the tube starts the voltage across the tube is square (AC peaks clamped at about 100v) & the zener does not conduct, therefore no further conduction of main thyristor. But if by any chance tube does not strike then re-striking does happen. In Old tubes the clamped voltage is beyond 140v. Tubes with still higher clamp voltage has danger of bursting.

[@demolishedman50]

wow Mr Clive, my fount of knowledge on the dark art of electronics, flummoxed by the recipe for this spell, tubus onestrikus. Broken Britain, I tell you.

[@bigclivedotcom]

I've been mulling it over in my mind and I'd guess that even when a lamp has failed to strike the thyristor will fire briefly to keep the capacitor topped up.

[@TheMowgus]

Would the voltage at the zener diode be lower because the fluorescent tube is now lit and therefore, no voltage would charge the capacitor to turn the thyristor on again (and thus, it would not try again)?

[@jaaasgoed]

But what if the tube does not light the first time?

[@tomfy44]

+jaaasgoed I wonder whether this is just the point. Maybe a second starter cycle is intended, just not as frequently as with the mechanical version?

[@Graham_Langley]

You've got the way a thyristor is triggered the wrong way round - the Darlington shunts the gate voltage here, stopping the the thyristor triggering. Given long enough the cap on the base will discharge and the starter will have another go at striking the tube - there's a clue in R4 labelled 'Reset control' on the datasheet cct. And now I've looked at the full datasheet that is what happens. However there's there's a problem with the ccts as shown if I'm thinking straight. If the diode (D1) in series with the thyristor is a normal one its forward voltage won't be enough to turn on a Darlington and I'd not like to bet it would reliably turn on a thyristor either. Edit: When hand-drawing a Darlington I just put two arrows on the emitter.

[@nickbird7742]

Well this brings back memories of the 70 ish I was about 14 and was amazed to discover that putting a starter and a light bulb in series, the bulb would flash. So I started building disco lights for friends. However my parents soon put a stop on my electrical explotes, not on safety grounds but because my testing was messing up their tv watching downstairs with rf interference.

[@semifavorableuncircle6952]

Looks like a better design than the ones i bought some years ago. First one i got worked fine, so i bought another ten or so. All of which failed within half a year, burning out some almost-new tubes and one even catching fire from a cold solder joint arcing. These had a relay inside and used a single diode to effectively saturate the ballast core with DC current to allow for a larger preheating current (which burns out the tubes and causes a really loud buzzing noise while starting), as for voltage detection they used a varistor (which is not intended for anything other than surge absorption and did burn out). I didnt bother to reverse engineer further, only replacing all the fixtures with electronic ballasts that finally work properly.

[@TheBalconyWorkspace]

Damn dude, I just bought 6 of them today for about 50 cents each, now I'm carefully considering again for the application, thanks a lot for your experiment , really appreciated.

[@EldaLuna]

i have an old late 80's salad fridge from a resturant i work from and it uses those things all 3 tubes. nice to know more about them as i always wondered one of them is original glass tube one in a tin shell but other ones were replaced by me when unit was still functioning years later before retired it and taken it home. i do like the clicking though it just sounds neat specially when tube flickers alot sure it ruins life of it but i dont know just sound and seeing it do that is fun

[@mpalaskokkalis1476]

Τέλεια τα σταρτερ για λάμπες φθορίου μπράβο παιδιά.

[@BritishBoy1971]

Clive really knows his stuff! It's fascinating!! :-)

[@richardkaz2336]

Once the Darlington pair transistor fires the collapsing choke field produces a Back EMF discharge that is enough for the fluoro to conduct resulting in the voltage drop across the fluoro is low enough that the Zener does not conduct and cause a re-trigger the heater element circuit. Once the flouro arc is struck there are enough charged particles in the tube for the reverse cycle to reignite the arc.

[@misterhat5823]

I'm guessing the the leakage current through the thyristor allows a small amount to current to flow after it's attempted to start the tube. Since the transistor is a darlington an extremely small amount of base current would allow the collector to stay below the threshold voltage of the thyristor and never letting the trigger current flow into the gate.

[@djonesboni1]

Here is how I think this circuit works. The gate-cathode junction of the thyristor conducts just like a diode and the thyristor will not fire if the gate current is low. When the transistor base voltage drops by discharging the capacitor, the transistor turns off and the gate voltage rises, recharging the capacitor. The transistor keeps the gate voltage below the threshold. The gate current (capacitor charging + transitor base + diode) must be kept below the threshold to keep the thyristor off (for lower consumption). In the case of the gate current of the thyristor is above the threshold it will conduct from anode to cathode and will recharge the capacitor the same way the gate-cathode junction would.

[@tartrazine]

I love how Clive has a mastery over both fluorecent ubes and neon. Like a sort of Bearded Plasma Wizard!

[@hadi96100]

Back in the 70's (in Iraq), I made a poor man's Sound-To-Light unit using starters & different coloured bulbs... It randomly flashed those on & off and had no connection to the music or sound, but it was fun, and I lived to tell the tale...

[@TheExcessiveDose]

The high holding current is to allow for a decent voltage pulse to be generated by the choke. If it had a holding current of, say, 1mA it would open in the moment wehen the choke would carry only 1mA and thus had no stored energy left to strike the tube. (Opposite reason of why you switch off huge tranformers right at the zero crossing of the current.)

[@DrRogla42]

I think the zener diode prevents the circuit from retriggering. As you said, the tube voltage when it is ON isn't 220V. Or did I get that wrong? So, the zener voltage is high enough to prevent gate current on the tiristor when the lamp is already ON. This one should retrigger I think.

[@bac302]

he is talking if the tube failed to turn on why does it not try to turn it on again

[@DrRogla42]

OK, I got it wrong :) It makes sense that I get things wrong.

[@bac302]

i agree he did not explaine it clearly

[@crapcbm]

a starter was the "poor man's disco lights" back in the 80s hehe, we did this a lot

[@cosminogloocosy1154]

...or Christmas tree lights :)

[@crapcbm]

right, and make more strings with a starter in each string, will give a nice christmas light disco hehe

[@DavidCookeZ80]

I don't think the diode that is in series with the main thyristor is a zener. The band on it is to the negative rail. If it were being used as a zener it would be the other way round with the band to the thyristor cathode. It could simply be that this diode's forward voltage is high enough when the heaters are on that it will exceed the voltage required to bias the darlington on (after the time delay) and reverse bias the main thyristor, turning it off even thought it still has a voltage across it's anode-cathode. If the tube doesn't strike at this point, I can't see how the starter is going to latch off. The 47uF timing capacitor is going to discharge via the diode (without the heater current it's forward voltage will be lower than that needed to bias the darlington on) and the "reset" resistor. I suspect it just keeps switching on and off, giving the tube further chances to strike. It would be interesting to watch what happens on a scope with a broken tube - after having taken precautions as a scope's probes are usually referenced to earth (and therefore the neutral in .uk).

[@dancoulson6579]

Hi, Bigclive. Thanks for making this video. I remember making a request for it ages ago.. Shortly after your 'sudden science - driving fluorescents' video was released a year or two ago. Have always wanted to know how these work. Time to grab a beer, sit back, and be enlightened. Keep up the brilliant videos. You're my favorite subscription :D

[@crimsonhalo13]

The light-bulb-on-flourescent-starter is an old favourite of mine for Halloween displays. Cheap and dirty flickering. :)

[@ExStaticBass]

It's a simple capacitive pull down switch that's being used as a timer in this instance. Once the capacitor get's enough charge it's no longer pulling the base/gate pin to it's active voltage or vice versa. Think of the dual LED flasher circuit and how it uses transistors and capacitors to switch back and fourth. This works on essentially half that circuit. Since it has nothing to reset it back the other way, it just stays off.

[@kiefac]

I don't know much about circuitry but I would assume the Darlington transistor is supposed to turn on and off, and then the track with the 1uF capacitor is just where the current goes.

[@PsiQ]

So, the tiny amount of current flowing through the "off" thyristor (which annoys you if you have a sensitive led which tries to start on it after loading its caps) keeps the transistor active enough to prevent the thyristor from starting... interesting use of something i always viewed as a bad/negative aspect of Thyristors. ( i always hated that you cant really switch off a thyristor dimmer without another mechanical switch to really cut off the leakage current)

[@xyzzy177]

My guess is that you got a mis-read on the darlington transistor, and it actually is an SCR. Try checking the part number on it again. Otherwise, excellent video and great explanation. Thanks!

[@sheadjohn]

i remember cold start and rapid start starters. also for the army we had some florescent lights that you could turn on a small tungsten light to warm it up. Physically warm up the florescent tube.

[@bigclivedotcom]

The early dimmable fluorescent lights had a tungsten lamp to provide continuous current to the cathode heaters and make them dim more stably over a wider range.

[@TechTed1]

A very good explanation I was kid when i learned how to turn on fluorescent tubes without straters only with a push button or just 2 wires 😂

[@ChristofErmer]

Well explained. I look your video in germany and recommend these my studends

[@jrmcferren]

I don't think the second circuit actually latches. When the 47µF capacitor discharges below the threshold the 1µF capacitor charges and latches the SCR on. The SCR turns on just long enough for the 47µF capacitor to reach (or actually exceed) threshold and it turns on the transistor and shunts out the 1µF capacitor and cuts of the SCR and the cycle repeats when needed (the 47µF goes below threshold). This cycle also happens so quickly that the fluorescent tube filaments (or your ballast lamp) do not have time to build up sufficient heat to start glowing, let alone starting thermionic emission and the inductive pulse created would also not be enough to attempt a tube start.

[@seanregan730]

I don't think I've seen one done before but what would make a great reverse engineering item and interesting to see the explanation on how they work would be a HF electronic ballast for fluorescent tubes if you might consider it Clive cheers Sean

[@pemac8856]

The SCR is a type called a gate turn off, or GTO. That high hold current rating is only with the gate grounded. Driving it negative will drive it off. Because the timing capacitor stays topped up, the darlington would quickly drive the gate down to about the same potential as the cathode, so it could be retriggering, but turning off again quickly. You could check that with a cro. The reason for making it a GTO would be to make it turn off while there is still at least 175mA flowing through it to produce the inductive spike to start difficult tubes. I think the 56k in series with the timing capacitor might be to produce a bit of ripple to turn on the darlington, or fire the cutoff SCR while there is current flowing through the TN22 to produce the spike. In mine it was a 6.8K. I don't know what it was for, as the original thyristor was a C106D, not a GTO. Just recheck the orange capacitor value to make sure it is not 10uF like mine. Let me know whether it is or not, and the number on the zener please, as well as the thermal fuse number, as mine is a standard metal one rated at 2A 125 deg C. The capacitor in a glow switch starter is to weld the contacts together briefly. I tried one without and the contacts just buzzed and it 1/2 lit.

[@analoghardwaretops3976]

In turn off mode , GTO's are equally power hungry devices that require at least (15-25)% of rated load current ( of reverse polarity) to turn off in my view , back then these were manufactured only for high voltage & high current industrial power control applications ..and such small / low power rated devices were economically unfeasible for general purpose everyday use. Whilst the GTO's turn on circuit was similar & simple as for SCR's , turn off circuitry was more complex & power consuming.

[@lescrossan27]

Once the fluorescent has fired there is going to be no voltage at the transistor collector that's why that 170v zener is there? Therefore there is no voltage at the gate of the SCR. The transistor is there to take the gate away if tube fails to strike within a certain time, it should reset itself? (Edit) the original circuit using the scr to control the gate would need a power cycle to reset.

[@techjdu]

Here's a great one for ya. My aunt brought by a pool pump for me to fix. She said it got wet and wouldn't turn back on.I just told her it's too complicated to fix. People don't get safety mechanisms.

[@BillyNoMates1974]

bigclive, i suspect the transistor shorts out the 1uf capacitor thus drags the operating voltage down for the power supply and then is held low. hence the thyristor cannot be triggered again regardless of the power supply coming into the starter

[@amrkoptan4041]

See how neat you make your videos, amazing !! thanks for the video

[@HIDLad001]

When the tube is operated in cold cathode mode on an electronic ballast and the emissive coating wears out, it is still able to strike (because it is a high voltage of about 600V`), but it will get hotter and hotter until the lampholder melts or the tube acutally loses vacuum.

[@Xanthopteryx]

I have never ever had an electronic lighter failing to light the tubes. They work like, Amazingly good! And they last for decays! My oldest one are like 20 years old, or more!

[@tonesoft1956]

Used to make disco lights using a starter on each one off theses RED GREEN BLUE spotlights and there also great for flashing Christmas lights did you know if you alter the value of the capacitor across the flasher you get different flash rates

[@JerryEricsson]

While I don't have any idea why the darlington transistor pair is set up that way, the setup reminds me very much of your Ghost Detector circuit built from Rabbit Ears! BTW I built one of those but instead of making it with rabbit ears, I used a torrid wrapped with a BUNCH of copper wire, it is very sensitive indeed.

[@whitcwa]

An interesting thing about fluorescent lamps is that some will not start without a grounded reflector within a half inch (12mm ). The capacitance helps the lamp start. If no reflector is present, you can grab the bulb to make it start.

[@bigclivedotcom]

You used to be able to get tubes with a conductive strip down one side to help starting if tubes were used in "architectural" designs with no ground plane in their vicinity.

[@mysock351C]

Always wondered the same thing about those starters back in the days before electronic balasts.

[@jrmcferren]

In the US we have old desk lamps that use Manual starters. The Japanese versions of them are even simpler. I will describe the Japanese makes here, but the American models simply have a latch in the off button added that is unlatched when the start button is pressed. When the desk lamp is plugged into the power outlet the line voltage is present across the lamp and ballast, but the lamp is not lit since a starting cycle has not occurred (note American units the off switch is usually still latched open). When the on button is pressed the start contacts are closed (dual tube units have two sets of contacts here). The on button is held in until the ends of the tube glow and then is released. If the tube doesn't start (either due to dodgy on button or not being held down long enough this cycle can be repeated. When the light is on the tube continues to conduct until the power is interrupted. When the power is interrupted (ether externally, by someone removing a tube, or by the off button for a few seconds) the lamps cool down enough that they will no longer conduct. The circuit in the lamps is rather simple. The off button is a normally closed momentary (latched in the American versions) pushbutton and is in series with the entire circuit. The on button contains a set of contacts for each tube and these contacts are placed where the starter is in the circuit. What's the big deal between latched and unlatched off buttons. The unlatched off button only breaks the potential from the circuit when held down. The tube can still be induced into a dim glowing state or started by external means (static discharge, discharge from a piezo ignitor, nearby lightning strike which is more common on the 14 inch tubes, CRT degaussing within a few inches, etc). The latched version keeps the normally closed switch open until the on button is pressed. Now these designs are not common outside of 100 to 127 volt countries as the tubes must be in series to prevent instant starting with the OCV of the mains in 220 to 240 volt countries. In these countries a standard automatic switchstart circuit is used.

[@mysock351C]

And for what its worth some of the cheaper ballasts have or had that same circuit built into them from what Ive seen. Obviously no longer removable, so if it takes a dump the whole ballast was trash. And yeah lol remember those back in my school on some of the desks since it was a historical site from the mid 1800's. Had to push that SOB button for a while to get light sometimes.

[@jamesvandamme7786]

The starter doesn't operate after the tube fires because the voltage on the bridge rectifier droops down and the zener prevents any voltage from going into the gate of the big SCR. If the tube doesn't fire or the light is switched off then on, full mains voltage is applied again.

[@hoggif]

It seems that when voltage is high, current through zener charges capasitor which eventually fires thyristor. Once fired, current runs through thyristor and "diode". I guess the "diode" is actually zener to get voltage higher. (darlington need about 1.2V at base to conduct and due to resistors the "diode" voltage must be more than this.) Once transistor cap is charged, the transistor latches the base of the thyristor down and the thyristor turns off near zero crossing (next low current point below the holding current of the thyristor). Now, if the lamp has fired, high voltage zener prevents the circuit charging and therefore thyristor will not fire again. However, if the lamp has not fired, the cycle should start all over from the beginning. If it does not seem to make another try, perhaps it just heats it up long enough to get things going at the first time?

[@Graham_Langley]

I agree with you, and nice to see I'm not alone in picking up on D1's voltage drop not being enough.

[@68MalKontent]

Yes, but it worked also with a standard bulb connected in place of the fluorescent tube, meaning that there's no voltage drop once it opens the circuit. That's the problem. The diode is most probably a dual/triple diode in one package.

[@saad85]

68Stan, The standard bulb was not in place of a fluorescent tube, it was in series with the starter.

[@68MalKontent]

And the filaments of a fluorescent tube are in series with the starter... The difference is that once the tube strikes, the effective voltage on the starter drops down to 70-100V, whereas in case of a bulb it goes back to the full mains voltage.

[@waltergeeraert3667]

As I see it: The fluorescent lamp should get lit by the coil spike after the thyristor gets shut off. After that the voltage over the circuit is low enough to prevent a restart, the zener doesn't provide current. If the lamp shouldn't start the voltage rises high enough to conduct the zener again and the cycle starts over once more. The R2, R3 and R4 should be dimensioned such that the darlington doesn't conduct anymore at a restart. Now I have a question: the darlington input needs 2x 0.65 Volts, is the series D in the main circuit a version with a higher voltage drop?

[@billywhizz6965]

Not to get boring I will just give a couple more thoughts that sprung to mind and call it a day. One is that the thyristor needs to be off while triggering the tube into life. First, it is possible because of the time it takes for the ballast or inductor field to collapse and create the starting voltage it is possible that the thyristor may trigger at the same time as the field collapses preventing the high voltage being created to fire the tube. Second and may or may not be so obvious is that when the high voltage is created the thyristor could fire and present a virtual short across the tube stopping it from firing.

[@vladomaimun]

The resistors and the capacitor behind the transistor have rather high values. You disconnected the circuit shortly after the lamp turned off. Maybe the capacitor was still holding charge. Or maybe the thermal fuse went off?

[@mcgyver272000]

If I am understanding correctly, this is basically an automatic version of the manual red "start" button on some desk fluorescent lights? I do remember my father's old shop having the lights flicker when you turn on the switch. Good old magnetic ballasts. I remember seeing several cards of starters in a local home center about a month ago.

[@Lenny-kt2th]

Basically, yes, but I would never have thought of the term "automatic" for those glow light starters that are still quite commonly used over here. It wasn't until copper became so expensive that magentic ballasts started fall out of favour and were replaced with electronic ones in new fixtures. But there are still loads of magnetic ballasts around.

[@ymaylove]

Awesome video! 40 years ago I broke the starter glass to get the bi-metal switch, that was extremely foolish , Thank science. Radiation Safety Aspects of Fluorescent Lamp Starters incorporating radiation source Radioactive substances used in the starters are 85 Kr, 147Pm, 3 H and 232Th.

[@fuzzy1dk]

if the voltage across the tube drops once it is on the thyristor should stay off because of the zener, so the transistor should just clamp the gate when the current has reached some point so the lamp can strike. Though it is hard to see how the voltage drop across a single diode can turn on a Darlington which has 2xVbe

[@blg53]

Could it be Clive that you had wrong specs assuming that the IC was a Darlington pair? A transistor does not "latch" like this, a special latching circuitry is required to enable it to do so, only a thyristor (or some mechanical relays) can do this sort of thing by itself. It does not make sense for makers of the starter to deviate from the recommended manufactures' diagram in that one detail only.

[@franciskent3551]

Clive any possibility of showing the circuitry of flourescent emergency lights which incorporates an inverter and batteries along with the ballast. 🍻 cheers.

[@MrStevetmq]

Using an old type starter on a incandescent bulb with red gel is a grate fire lighting effect. esp shining though a slow running fan.

[@celsostarec6735]

AFAIK starters are suposed to boost startup voltage by sudenly opening the circuit causing the changing current across the choke (di/dt) impose a higher voltage across the fluorescent tube than what is available from the utility. This is necessary for the firt ionization of the gas, that when heated up with current flow will now conduct at regular voltage. After that initial ionizing cycle, the starter is not needed anymore, as the lamp will conduct by itself. That is why it's called a "starter". No wonder an electronic starter would trigger above normal conducting lamp tension (the zenner part), and will have both a means to start conducting and stop conducting across it's pins. They are probably calibrated to conduct through the starter and the electrode heaters just enought for initial startup and then submit the lamp to the overtension across the electrodes. As all of this is stressful to the materials of the lamp the shorter the stress imposed the higher the durability. Even though solid state lighting is taking over this technology for good, industry is already sabotaging durability with planned obsolescence through bad drivers with lousy inappropriate capacitors to make virtually unlimited duration LED lamps fail. Where I live (Rio de Janeiro, Brazil) we've made pilot instalation of some 500 T8 Led tubes. They've been working of for 4 to 5 years. Half a dozen have had failed 10 led segments out of near 200 Leds. No whole lamp failed. We've scaled it up recently to some thousands while in a retrofit. Contractor could select any of the equivalent lamps he wished, and told us they would go with the best garanties. 5 to 10% lamps failing in the first month of use... Our national standars organization ABNT had made norms in line with American and International ones, and the country will impose compulsory testing and certification of the devices including the driver and it's capacitor (well known object of willful sabotage in electronic balasts and led drivers). Industry is alway lobbying against. It will be compulsory for everyone by the end of 2017 unless industry manages again to postpone it. That's the way it goes.

[@johnhill5745]

I used to use the old type starters to flash the old chistmas tree lights back in 60s/70s.

[@billywhizz6965]

My guess is that the main thyristor must fire every so often, depending on the state of charge on the capacitor on the base of the Darlington transistor. If enough charge is held then it does not fire every half cycle but it must fire every so often. Just enough to keep that capacitor charged enough to turn on the Darlington transistor.

[@ABaumstumpf]

Are you sure it only fires once? Might just be that the capacitor holding the darlington high last for a minute or two?

[@SuperBlackReality]

Interesting, would you also take to bits those compact fluorescent bulbs ?

[@geofflotton5292]

Nice bit of product placement with the meter :). We used to have starters with cut out built in. When a tube failed to light, the starter would trip out. They were called watch dogs. Why won't twin series 20w tubes work with a universal starter?

[@bigclivedotcom]

The twin tubes in series require a starter with different voltage thresholds because of the much lower open circuit voltage across the tubes. I was never particularly fond of the single ballast for two series tubes approach.

[@jlucasound]

I heard you are on your 5th closed caption Stenographer. They keep saying "I can't keep up with this guy! He keeps Waffling!" God, I love Waffles. I love Clive videos. What was the point of this? ;-) Maple Syrup and Butter. Mmmm.

[@AttilaAsztalos]

Through all of my childhood, a classical starter like that was the thing that kept the string of Christmas lights blinking on the tree... :)

[@geoffgeoff143]

That was probably a bi metal flasher built in to a case the same as a starter.

[@misterhat5823]

The high holding current means that it turns off at a higher current, instead of waiting until the current is almost zero. Since the energy stored in an inductor is I*I*L the higher current ensures more energy and a better chance of the tube striking. That said, why not use an avalanch rated MOSFET? It could be made to turn off at any chosen current. Better yet, a PIC and a MOSFET. That way it could choose the needed striking energy and lessen the blackening of the ends and still strike every time.

[@swedishspymuseum]

The Darlington have a gain of 20k accoring to ON. Maybe its running on some leakage from the either the thyristor or the cap

[@sparkyprojects]

Only thing i can think of is a slight leakage that keeps the cap alive Many years ago i was asked to put a warning light on a freezer room door (room was in a cold room), so if the door wasn't closed properly they would know from the corridor, i installed a sealed industrial limit switch in the door, and fed it out to a red bulkhead fitting, to make it more obvious i wanted to use a flasher unit, but then i thought about using a starter, just the normal mechanical (glow type), but it worked well for many years, and because of the random flashing it was much more annoying and obvious.