PCB top

electronic ballast repair

WARNING: Electronic ballasts are dangerous, not only because they are usually connected directly to the mains but also they generate very high voltages: the one described below has 600V output, much higher than the 110/220V mains voltage.

So by no means try this at home without proper training and certification! Please read Safety Guidelines for High Voltage and/or Line Powered Equipment

I had a fluorescent desktop lamp go dark for no apparent reason. I opened it and found an electronic ballast with the following markings:

Electronic Ballast for Fluorescent Lamp
Model: EBCF-127-120V-LPF
Input: 120/60Hz 0.38A
Lamp: PL 27W
Output Voltage: 600V

After a few minutes of eye inspection (and nose:) ), I couldn’t find anything with burning marks or any foul smell. The small fuse had continuity along with the connecting wires, which made me think the culprit was the bulb, as the glass was fairly dark near the filaments, although in hindsight the latter came from its age and constant use. The bulb has 27W and 6500Kelvin color temperature. Without much thinking I got a new bulb but when I received it a few days later it didn’t switch on.

So going back to the ballast I went through the circuit, testing the likely culprits in the high voltage section: open/short capacitors, diodes or transistors and finding none, the mystery deepened. I cycled the power and then I went around poking the voltages to find that the mains rectifying circuit was working fine but when prodding with the multimeter probes next to the diac suddenly the bulb flashes! Humm, maybe the diac broke since it’s used to start the circuit, somewhat replacing the old mechanical/thermal starter that pairs reactive ballasts.

Testing a diac is more tricky with just a multimeter so I checked the circuit around it, finding to my surprise an open 510kΩ resistor (R1 in the circuit) with no burning or overheating marks. This resistor is responsible for the charging of capacitor C2 which, when reaching a high enough voltage to overcome the diac breakdown voltage, turns on transistor Q2 to provide the initial kick in the oscillator circuit. Replacing this resistor brought the circuit back to life and now the lamp was working again with no issues.

I took the time to extract the schematic as seen below , which can be downloaded in eagle, png and pdf formats, along with relative placement of the parts on the original pcb.

Some interesting references can be found at

13 thoughts on “electronic ballast repair

  1. K Ramakrishnachar

    I have repaired several ballasts having this kind of trouble.What is the remedy ? is it better to use a higher wattage resistor !

    1. Nuno Post author

      I wondered the same, and you might be spot on, but what power rating should be used? The original resistor R1 seems to have a power rating of 1/8W or maybe 1/4W. From Using a simple calculation ( P = R * I² ) gives a maximum current of about 0.5mA. Since the discharge diode (D3 on the schematic) is constantly being triggered by transistor Q2 we could assume that average current could be attained.

      But then looking at the circuit voltages, we would probably get a maximum of about 120*√2 DC as V+, which would give a maximum dissipation power of about 56mW ( P= U² / R ), well below the resistor rating, if my calculations are correct.

      Maybe a full circuit simulation would solve the mystery :)

      1. Hartmut

        Since a voltage doubling circuit is used here, U+ = 2*sqrt(2)*120V=339V => P=225mW, quit close to 1/4W!
        However, I think the resistor died of electromigration due to the high voltage.

      2. Joe

        Resistor R1 is a classic “1/4W” size (which are sometimes rated at 1/3W) and looks to be carbon film, similar to the old Philips CR25 type. Looking at a similar existing type, Royal Ohm CFR0W4, we see that it has a maximum working voltage rating of 250V (and a maximum overload voltage rating of 500V). With C2 constantly discharged, the voltage across R1 will typically exceed 300V, so that’s why it failed.

  2. Hartmut

    Once I found a similar failure in a switching power supply.
    My Tip: Use two resistors in series!

  3. fred g

    Hi, Nice site and info! I have a lamp with the same ballast. I checked continuity on the fuse and it’s open. Do you happen to know the rating of the fuse I should use?
    Thx, Fred


      I am interested in learning how to construct and repair a
      electronic ballst. Any body can help me in providing such
      training information

  4. Ernest

    I just finished repairing my sister’s lamp just like this one. Identical board. Thanks Nuno for the great photos and schematic. I could not have done it without them.

    I had several failed components. Bias resistors R3 and R4 in the transistors were open. Replaced with two 22 ohm one half watt resistors in parallel, giving me 11 ohms, which turned out to be OK. One had to be standing upright, the other “piggy back.” ( I also had to use a PC drill to enlarge the holes in the board.) Replaced open emitter resistor R5 with a 3 watt wire wound 2.5 ohm RCL type 7745 and the open R6 with a measured 1.8 ohm 3 watt of the same style. I made my resistors measurements with a Heath Kit IM-2215. That’s what I found for resistors without digging too deep in my this and that boxes. After the 11 ohm resistors were installed, a friend came by with some 1/2 watt 10 ohm resistors.

    The transistors were replaced with E13005’s. They have a very good power dissipation. During some testing and probing and slipped and made a spark. Out went the light!.. Fortunately, only the fuse blew. I had to remove it. It was marked 2 AMP. I replaced it with one of the mini’s that I have, a one(1) Amp fuse. I have had the lamp operating for about 15 or twenty minutes, but I intend to test it longer as this circuit warms up those transistors. It’s all assembled now, and it is one of the floor type lamps.

    I checked the transistors with an analog volt ohm multi meter backed up with my EMC-212 transistor tester. The junctions were shorted out.

    My friend had given me a couple TE-13007-2 transistors which I had previously installed. Since I had not replaced R6 yet, they did not work. But, I think they will, and I may have to use them in another unit my nephew has which is identical. It should be much easier, lets hope. I sure struggled with this one. I’m putting this in the thread, in hopes that it may help others. So, don’t be afraid to substitute the resistor values as I did. Regards, Ernie K1WJK.

  5. AintBigAintClever

    I’ve just fixed exactly the same problem in a Kosnic KCF24R7S 24 watt halogen replacement CFL. One 510K resistor in the startup circuit measured around 810K, the other was open circuit. Replaced each one with series-connected 220K and 330K resistors, now working fine.

  6. Rick

    Found C3 exploded in our ballast – literally. Is that really a 1Kv capacitor? There is nothing to be read on the original – that whole exploded thing…


Leave a Reply

Your email address will not be published. Required fields are marked *