So I found this CFL. Took it home and it didn't light. So I dismantled it and checked the innards. Both capacitors were dead. One vented and the other shorted. And the safety fuse on the board had also blown. Aside from the capacitors blowing what else on the ballast could be broken? I looked up the part number on the transistors and it looks like they're discontinued. Would there be modern equivalents I could use if those are also fried?

Here is the data sheet for the transistors:

https://www.datasheets360.com/part/detail/mje13009/3992491617836637936/

The capacitors are 200v 68uF. Would it be safe to use a single 400v 120uF capacitor in place of these two?

The transistors might or might not be dead... Desolder them and test with a DMM on diode setting (showing the forward voltage drop, not resistance) for the Base-Emitter diode functionality in them

Either way, the 1300x are fairly standard transistors used in many cheap switching power supplies (also in ATX power supplies), they are fairly mass produced. Some specific versions could be discontinued (like a non-ROHS version, or a specific case, or from a specific manufacturer) but the general compatible family is not going anywhere

Here are a few, they differ virtually only in cases (that matters if they are pushed to their thermal limits, which may be the case in the CFL, so choose the ones with non isolated tabs if the originals were also with non isolated tabs)
https://www.digikey.com/products/en/discrete-semiconductor-products/transistors-bipolar-bjt-single/276?k=13009



The capacitors can be just in series, or they can act for 2 other purposes :

1. Being a voltage doubler circuit, especially in an 120V CFL or if the tube is of some extremely high voltage that would need a doubler from 240V (i think the latter is unlikely). If it is an 120V CFL and uses a doubler, you could convert it to a non doubler circuit and supply with 240V instead

2. Providing the center voltage point to a half bridge in the HF converter. You would see that one end of the CFL tube connects to the center point between the capacitors or so. This also can be hacked to an extent - Since the HF circuit does not need all this capacity (its needed for smoothing the 60Hz input), you could make a half bridge divider with 2 film capacitors of much lower values (requiring only that they are rated for the tube current, that would be on the order of 1A), and the electrolytic would be one for 400V across the 320V DC bus only

Some half bridge implementations don't use the center point between the electrolytic caps but use a pair of film capacitors in the first place, or (cheaper implementation often found in CFLs) using just one film cap to one of the rails of the DC bus and thats it...

I swapped in some new capacitors and a fuse and powered up the circuit. As soon as I did the capacitor heated up extremely fast and so did one of the transistors.

An electrolitic cap would heat up fast like that either from reverse polarity, or from overvoltage

Reverse polarity could happen if it is soldered in in reverse, or if some diodes in the rectifier bridge (or single diode for that capacitor in case of a doubler circuit) are shorted and the capacitor is getting AC

Overvoltage in a divider could happen if one side of the load is shorted (due to shorted film capacitor or transistor, ...) and the full voltage falls on the other side. Say if you apply 320V (rectified 230VAC) to a divider of 2 200V capacitors, and due to a shorted component the entire 320V goes to one of them

Overvoltage in a doubler is not possible

You have there some other dead components - possibly the transistors or film capacitors

Besides, they are possibly a result of EOL of the tube - Have you checked the tube ? (say with a T5HO 80W ballast)

An electrolitic cap would heat up fast like that either from reverse polarity, or from overvoltage

Reverse polarity could happen if it is soldered in in reverse, or if some diodes in the rectifier bridge (or single diode for that capacitor in case of a doubler circuit) are shorted and the capacitor is getting AC

Overvoltage in a divider could happen if one side of the load is shorted (due to shorted film capacitor or transistor, ...) and the full voltage falls on the other side. Say if you apply 320V (rectified 230VAC) to a divider of 2 200V capacitors, and due to a shorted component the entire 320V goes to one of them

Overvoltage in a doubler is not possible

You have there some other dead components - possibly the transistors or film capacitors

Besides, they are possibly a result of EOL of the tube - Have you checked the tube ? (say with a T5HO 80W ballast)

I've checked the tube with a continuity tester and both cathodes are intact. I've also checked the vacuume and it seems to glow when high voltage is applied. The ends of the tube also appear white. I can also visually confirm through the clear ends of the tube that the cathodes are intact... I think the ballast failed prematurely.


I did notice that one of the mylar capacitors is shrunken as if it shorted then began overheating.

So basically I should go the route of replacing all of the semiconductor components and caps?

Do you think the choke could have been damaged by the short out or is it more robust than that? The company thag designed the board did implement a fuse luckily.

Capacitors can fail randomly so this could be the initial cause of all the rest of the damage

Test the components for shorts (capacitors) or for correct operation (transistors) with a DMM and replace accordingly. In an NPN transistor, testing across Base (+) to Emitter (-) must show an intact diode with drop on the order of 0.5V, when the transistor is desoldered from the board

The choke and the tube are in series - Any overcurrent through the choke would go through both of them, in which case i think a cathode in the tube would blow way before the choke heats up. The same would probably happen in case of isolation breakdown in the choke. I think the choke is OK

I checked both of the transistors and they're dead shorted. I checked the big choke and it only reads 0.5 ohms. Is the "2.0k" on top what it is supposed to read?

There is <1 meter of Copper wire wound in that choke, its resistance is near 0. The 2 Ohm seems about right for resistance of the choke + the resistance of the multimeter probe wires

Those chokes are not "off the shelf" part, they are made to order by a choke manufacturer for the CFL manufacturer (or made in house). What's written on the choke often have no meaning other than a part number. In this case some of the numbers could mean 1.0uH, 2.0uH, EE-15 core, ... or just be meaningless part numbers or version/revision numbers

I checked both of the transistors and they're dead shorted. I checked the big choke and it only reads 0.5 ohms. Is the "2.0k" on top what it is supposed to read?

The 0.5Ohm is something I would expect. The "2k" means the inductance (2k uH = 2000 uH).
The transistors should be possible to buy, different makers use different prefixes (the MJE came originally from Motorola/ON Semiconductor). So e.g.  E13009, D13009, FJP13009, KSE13009, MJE13009, ST13009 are all essentially the same transistor.
When the transistors are out, check the resistors as well. They tend to blow up too once the transistors fail.

Thanks for the help so far guys!

I have another question. So there is a DIAC in the circuit. Specifically a DB35T. How likely is this component to fail? I have no way to test it unfortunately.

Also I heard that diodes are usually suppose to test at .6v but all of these diodes are testing at .5 and one is even down to .4 are these still good?

The diac can be in fact tested. Connect it in a circuit like whats below (component values are approximate) to 120VAC and connect a DMM as the volt meter. If the diac is OK, you'll see a DC voltage that rises and abruptly drops to a low value, repeating every few sec

In place of the 10M resistor you could use a Neon indicator screwdriver connected to the Phase (same as when used to deect the Phase), and connect the rest of the circuit between the indicator's touch terminal and Earth

But I've never seen the diac failed yet.
But there is an easy way to test them once the transistors are out of the circuit:
Connect a loudspeaker instead of the base and emitter of the bottom (negative; the one to which base is the diac connected) and power the circuit up (assume youve fixed the rectifier). A whiney (or sharp buzzing, or virtually any audible) noise means the kick start circuit components (resistor, capacitor, diac) are all OK.

I would be more worried about the output resonant circuit: Check, whether the coil winding isn't visibly "fried" and then the health of the resonant capacitor.

Thanks for the help so far guys! It is very much appreciated! I will be ordering parts soon hopefully.