Author Options:

Help upgrading CFL circuit? Want to drive cfl's up to 25w. Answered


I have built this circuit, it can drive cfl's up to 15w, how do i modify it so that it can drive cfl's up to 25w? 
I am using D313 for his project as the D882 didn't gave me the results i needed. Using 12v battery.

Here is the original instructable link https://www.instructables.com/id/A-simple-low-voltage-inverter-for-fluorescent-lam/?ALLSTEPS

(This question is optional)
Do u think the transformer i made for this circuit can be used for the famous Jeanna's light (Joule's Thief). If yes then how it would connect on the circuit? I will be making two or three more transformers like this. Do u think it needs modifications for its working like number of turns for the primary, secondary and feedback, different wire gauge?


Just to let you know that it is not my circuit, i just followed through the Instructable of the person who made this. Its link is above.

Anyway you said i should use capacitor in series with cfl, let me know what is the value of capacitor i should use?

The cap is only needed if you design a quite powerful transformer that is able to feed a let's say 20W flouro but want to use a small 8W lamp on it.

Basically just to llimit the available current so the lamp won't get too hot.

If you use quite thin wire on the secondary you have a good chance the current won't go too high anyway - simply test your setup with a big lamp and if it works but a small lamp on the same setup get's hot you add tha cap.

Start with something in the 470nF range and if the lamp refuses to work take a higher value until it works. If the lamp still heats up use a lower value like 220nF for example.

In a perfectly adjusted setup the input power used should only be slightly higher than the rating for the lamp.

So if you use a 11W lamp but your circuit draws close to 3A when running it uses to much.

Smae story if it is the other way around, if a 11W lamp only draws 500mA it won't reach full brigtness or might even only glow at the ends of the tube.

Thanks for your reply, i appreciate.

I did and experiment yesterday and added more turns on the secondary about 100 more making total 350. It ran at first without giving any more power or brightness in the cfl then transformer started giving strange noises and sparks, turned it off and since then its not running anymore. When i connect the power all that is happening is the transistor is getting extremely hot within 5 seconds.

I suspect you shortend the transformer or blew the transistor.

You might want to consider a ZVS circuit with two little mosfets instead.

If your primary was not properly insulated and the voltage too high for it there would have been arcs between the windings, maybe the sound you heard.

Yeah, you suspected it right. I will take this ZVS circuit into consideration.

I am also interested in this Jeanna's Light (Joule Thief) Not sure if i can use this transformer for the same purpose, i mean the windings are the same, all i have to do is connect the windings accordingly , i am gonna try using 2N3055. The idea is that this one seems more energy efficient.

Can you tell me what should be the primary and secondary gauge margins? I mean if i were to use 30 awg wire for the primary what gauge of wire i should use for the secondary?

Can't really give you recommendations on the wire :(

I take what I have available at hand but for HV only the thinnest wire possible.


Well i need your advise again...

Like you already know the issue was my transformer was shorted, So yesterday i winded it again and just as i was about to make it permanent on pcb, i thought i should check it one last time and then this happened. Just as i turned it on my transformer started giving weird noises and CFL went down from bright to low and then completely off at the same time my transistor was damn hot. I have no idea why it happened, i mean i have already tested it before for 30 minutes and it was running fine and everything was cool, i have big heatsink installed on the transistor. Still when it was working during that 30 minutes, heatsink was safe to touch, but now it might even melt solder.

I tried different transistor, capacitor, resistor but the transistor is still getting hot.

Just to let to know, when i was packing it all up, i saw one of the high voltage wire was touching the heatsink of the transistor, do u think this has anything to do with the issue?

Also i used D313 transistor, also tried with TIP41C

If you feed back HV to the transistor it really won't like it.
The weird noises, if the transformer is not getting hot is either vibrations of the wires due to low frequency or arcing between the layers of the HV wiring due to poor insulation.
If you magnet wire is not coated properly or the coating is damaged (salvadged wire from trnasformers or similar) you componsate by sealing the transformer with resin or wax.

A little vacuum pump and suitable container is advised as ohterwise you will trap air and in the air pockets still get arcs.

Well, i rewinded the transformer, the third attempt, and it is working again. I guess in the end transformer was shorted due to HV secondary was touching transistor heatsink. This time i made sure that i there is proper insulation between every layer not just secondary. I am gonna put everything back together in pcb today. Breadboard testing was successful and cfl was able to light up about 60% bright. Gonna need a bigger heatsink though. Also it can drive cfl's up to 20W.

My next project would be making Jeanna's Light. My last two attempt of making it was a complete failure and mess. I tried rewinding toroid again and again but it didn't even blink. Then i found out the toroid i was using was a iron power made, and cfl's requires ferrite made toroid. Well i don't have toroid of such material but i am gonna use transformer. I also have ferrite core from a flyback transformer, do u think i can use that for this project?

In most power supplies you can find suitable round cores as the input filters.

Flyback works too, depends on the size you have in mind.

For your CLF project you might want to consider a ZVS circuit to reduce the wasted power on the transistor.

Although the overheating could be due to a missing filter too, have you tried a RC filter parallel to the primiry of the coil?

Well, i started winding the flyback core and wondering about one thing, it must be silly of me to even ask but does it make difference whether i wind primary and feedback beneath secondary or above it? I will surly consider building zvs circuit, i just wanna finish what i have on hands right now.

What is RC filter? as i said i am not that much experienced in electronics so i don't know what that is.

You should always wind the primary first, or in case of a flyback core on the other side of the core.

This prevents over arcing from the HV side to the low voltage side.

Basically if you do them stacked you have the primary about 20% longer than the secondary and insulated with a few layers of insulation.

Using two coils on a flyback is much easier as you can use a plastic bobbin and and electric drill to wind the secondary, when done you simply slip it onto the core and close it.

You don't need the RC filter for the ZVS circuit.

It is basically a capacitor (C) and a resistor (R) in series, this RC filter is parallel to the primary coil.

The transistor switches the coil on and off but without protection the stored energy will fire back onto the transistor when it switches off.

The capacitor in the filter stores that energy and the current is limited by the resistor.

As a capacitor, that is empty, act like a short in DC it will be only limited by the resistor, once it is getting fuller the internal resistance rises.

So it is only the initial "bump" that comes back from the coil that is filtered out, the rest can be handled with ease by the transistor.

Hello once again. First of all thanks for all the help, i was able to complete that last project pretty decently. I had an old led casing so i used it and assembled everything in that. Attached an image of it, its not so clear, i think my camera lenses are not good these days.

What i wanted to ask is about the second image of the schematic diagram.

Is this the example of a ZVS circuit you were talking about? The two A's are connected to one another same goes for the 12-15v, i don't understand the GND, the gnd wire upper one should connect somewhere but its not, or i am missing something here.

Picture 53.jpgF8MC68JGTN5PZEX.LARGE.jpg

The circuit you found is not bad at all.

I don't really know why where is more on the secondary coil than just the CFL, especially considering the TIP is only good for 80V.

The transistor fires by the impuls from the feedback winding, so not a realy ZVS like used with mosfets but does the trick.

L1 is the choke coil and IMHO it could do with less than 50 turns.

You can try the circuit as it is and if not working to your satisfaction try the CLF directly on the secondary without the transistor and capacitors.

If I understand it correctly the parts on the secondory are supposed to form a resonant circuit which results in a much higher output voltage.

Alright thanks i will try it, but what about the -GND wire just above the secondary? Where does it connects to? or should i just ignore it?

Also what should be the winding order for this? I will wind it on flyback core. So 400 turns secondary on one side and rest of the coils on the other side according to the schematic diagram?


L4 on one side of the flyback, L2 and L3 on the other, the choke coil, L1 on a ferrite toroid core.

All GND connection go to the common ground, so all go onto the negative connection of your power supply.

L3 and 3 should have the same winding direction, for the rest it does not matter.

Please tell me what windings should have same direction? L3 and ?. Thanks i appreciate

They should have the same direction but as I said: If the circuit is not working at all and all parts are fine simple reverse the polarity on either L2 or L3.

You mean L3 and L2 should have same winding directions? I want to confirm it as you seems to have typed 3 two times.

Yes, unless the circuit specification call for it the other way around.
The reason is that you want to fire the transistors with the voltage generated from the feedback coil.
On the other hand it also means that if your circuit won't work at all and all transistors are fine you should reverses the polarity on the feedback coil to check if that solves the problem.

The CLF and transformer form a resonant circuit.

Well, at least they do in commercial versions of a driver.

In a design like that the losses are minimal as a ZVS driver is used.

Your design reminds me of the cheap circuits found in the flour lamps for car use.

So all the work has to done by the transistor so to prevent overheating you need quite high frequencies and a matching transformer.

The secondary must provie enough voltage to fire the CLF and enough current to produce enough light output.

A capacitor in series with the CLF can be used as a current limiter, this way you can design one transformer for all lamps and for smaller ones you add a cap.

Frequencies up to about 250Hz would allow you to use a metal core for the transformer but your transistor might be doing overtime.

I would try to get in the 1-5kHz region and use a good size ferrite core with a bobbin.

You'll need to change out the transformer.