This is the schematic of an electronic ballast model 234SLESW designed for running a pair of 4 foot T8 or T12 fluorescent bulbs. It claims to have an Energy Star rating when used with F32T8 bulbs. I picked one of these out of a dumpster and it was broken so I had to draw up a schematic. I figured I'd share it with you. It should be fairly easy to construct if you needed a light, efficient, flicker free ballast for your shop or garage or heck, even your living room!
WARNING: This circuit is AC line powered and contains voltages and currents which can KILL if you are not careful. Charged capacitors will SURPRISE YOU! They can hold a lethal charge for hours! If you don't know much about working with line (and higher) voltages or if you aren't crazy (like me) then DO NOT attempt to construct this circuit. I CANNOT BE RESPONSIBLE if you electrocute yourself to death! That said, let's have some fun!
Parts List
R1, R6 470K Non-Flameproof
R2 22R
R3, R5 1R
R4 30R
R7, R9 30R 1/4W
R8, R10 22R
R11 47R 1/4W
CO1, C7, C8 470n, 250V
C1, C4, C5 100n 63V
C2, C3 47u, 250V
C6 2n2
C9 - 12 33n, 630V
C14 3n3, 2KV Ceramic disc
L1 See Notes
L2 See Notes
L3 See Notes
T1 See Notes
D1 - D5 1N4007
Q1, Q2 PHE13005
Q3, Q4 SS8050
DIAC 32V, DB3
RV 270V 10% Zinc Oxide Varistor ZOV-07D271K
F1 2.5A
All resistors 1/2 Watt and flameproof unless otherwise specified. All capacitors polyester unless otherwise specified.
Eventually, the core of T1 will saturate. At this point, current in T1b will decrease, slightly turning off Q2. This will decrease the current to T1c and in turn T1b and so on until Q2 is completely off. The collapsing flux in T1 will cause the voltage in the windings to reverse, turning on Q1 via T1a in much the same manner as Q2 is operated by T1b; the induced voltages in T1a and T1b being opposite.
Once Q2 is on, the top 1/2 of the circuit operates in much the same way as the bottom 1/2. Current flows through T1c and the bulb filaments and filament capacitors in the opposite direction and also through C8 instead of C7. Current is also flowing through the base of Q4 through R4, C6, R3, Q1, and back through the supply rails, thus clamping the base of Q2.
Eventually, as the filaments of FL1 and FL2 warm up, the mercury inside becomes conductive and current begins flowing straight through, mostly shunting the filament capacitor circuits. However, enough filament current still flows to keep them warm. Also, HV transients produced by the inductors help break down the mercury, along with C14 which is tied to the chassis ground.
Mains supply is filtered by RV, L1, and CO1. RV also soaks up any back EMF from L2, L3, and T1 in the event that it makes it back through L1. A voltage doubler consisting of D1, D2, C2, and C3 provides around 330V for the rest of the circuit. C1 and the DIAC provide the starter pulses to get the oscillatios going. D4 and D5 protect Q1 & Q2 from back EMF from the coils. Q3 and Q4 clamp on the bases of Q1 and Q2 in such a way that both are never on at the same time, preventing a dead short.
This circuit uses an interesting method of driving the bulbs. Instead of a normal split supply, where the load, in this case everything between T1c and C14, is connected to the center of the split supply; we have it connected between two capacitors. This novel approach allows the load to swing at full rail voltage while at the same time preventing any strong DC currents from flowing. This protects the entire load circuit in the event of a short circuit failure of one of the power transistors.
The circuit is basically 2 parts. The first part is a high frequency oscillator which is everything to the left of T1c. The second part is the inductive current limiting section which is L2 and L3. Even though L2 and L3 are physically quite small, at the high frequencies (ultrasonic) generated by the oscillator section, they provide enough reactance to absorb just the right amount of current for the specified bulbs. This is why electronic ballasts are so great.
Unfortunately, I could not measure the inductances of the inductors and T1 as I cannot afford an LCR meter. Does anyone want to donate one to my cause? I can tell you the sizes though which really means nothing if you were to try and build this thing, because your ferrite cores may have a different Q factor than the ones used here. T1c is 9 turns of #28 or 30 wire. A tap is then taken and T1a is wound in the same direction and is 4 turns of the same wire. T1b is also the same size wire and is wound in the opposite direction of T1a and T1c. T1b is not connected to the other windings in any way except inductively. L1, L2 and L3 I do not know the number of turns because I don't want to take them apart if there's nothing wrong with them, but L2 and L3 are identical and are wound with what appears to be #30 wire until it fills up the bobbin. For L1, it would probably be the easiest to use a premade line filter from a broken switching power supply like from a TV and its associated power factor correction capacitor(s).
All drawings were made with XCircuit
(opencircuitdesign.com/xcircuit), a platform independent, X-Windows
application.