Replaced capacitor in fluorescent light ballast, and the new one blew right away. Why?

Thread Starter

storckm

Joined Apr 6, 2018
9
My wife has a fluorescent light she uses to start seedlings for her vegetable garden. This spring, it stopped working, so I replaced the blown capacitor with a new one, same voltage and capacitance rating. Score one for helpful husbandry.

After a few days, the light stopped working, again. I assumed it was the other capacitor, since they are (I presume) similar in age. Imagine my surprise when I took a peek and found that the brand-new capacitor I had just put in place was blown.

Now I could simply have soldered the capacitor backwards. But I was pretty careful about that, and I think I put it in the with the same polarity as the one I took out--opposite to the one on the other side. This is my first foray into ballasts, and I'm wondering which of the following might be the failure's cause: (1) some rating other than voltage and capacitance is relevant here; (2) I simply had a bad capacitor; (3) some other component is bad and causing the capacitor to fail; (4) I put it in backwards after all; (5) localized quantum effects caused my circuit board to temporarily cease to exist, then resume existence with a blown capacitor.

I bought four capacitors from digi-key, so I can try again, but I don't want to just throw components at the problem. If it's any help, the twin T-8 fixture is branded "Lights of America," and probably originated at Wal-Mart (although she acquired it at second hand.

I also may have switched the wires to the pins at the near (to the ballast) end of the bulb, but I wouldn't think that would make a difference.
 

AlbertHall

Joined Jun 4, 2014
12,345
Yes that capacitor would go bang. It is an electrolytic and should only be used in DC circuits.
Is the original capacitor 47uF?
Is your mains supply 120V or 240V?
 

ebp

Joined Feb 8, 2018
2,332
Can you post some photos of the ballast? I'm guessing this is an electronic ballast where the cap is a filter after a rectifier. There are other possibilities, but it is guesswork as to what.

You didn't install it backwards or subject it to significant AC - it would have loudly announced its displeasure with either case within seconds.
 

Thread Starter

storckm

Joined Apr 6, 2018
9
Maybe. Although the original was rated 200V and I bought a 250V for greater longevity (or so I thought).
 

ebp

Joined Feb 8, 2018
2,332
Equivalent series resistance (ESR) and ripple current rating are the two most important capacitor parameters for this circuit.
What are the markings on the remaining original capacitor?
The chosen replacement looks like it should be suitable, but it would be good to check the spec's for the original if possible.

The circuit looks like it is probably a "half-bridge." The two capacitors would be in series, each charged to near the peak of the AC line voltage. The voltage of the two caps normally stays well balanced because of the way they are charged from the AC input (full wave voltage doubler) so overvoltage seems unlikely (200 V cap has about 15 volts of margin with 132 VAC line - 120 V plus 10%). I'm a bit suspicious that there may be something wrong with the diode closest to the capacitor. Diodes generally fail short-circuit, but that would detonate the capacitor almost instantly. I wonder if its reverse breakdown voltage has been degraded somehow, subjecting the cap to very large ripple current. It seems unlikely there is a problem with the transistor, since the lamps did light for a time with the new cap.

The two TO-220 devices will be bipolar transistors. The circuit is self-oscillating. The large ferrite toroid with the three windings of a few turns each is the base drive transformer. The things with the green tape on the cores at the left end are the output inductors (more or less equivalent to the large iron-core inductor of a magnetic ballast).
 

Thread Starter

storckm

Joined Apr 6, 2018
9
Those are transistors, E13007D. The original capacitors were marked CD11GH.

Ripple current is listed as 398 for the original, v. 980 for the replacement.

Never mind that. On the datasheet for the original, measurement was taken at 120Hz, replacement at 100kHz. At 100kHz, it will be 796 v. 980 mA; 398 v. 392 mA at 120Hz, and 716 v. 882 mA at 10kHz. Would that ~200mA difference matter, or should I try replacing the diode? Also, I'm not finding any ESR ratings for either part.
 
Last edited by a moderator:

Thread Starter

storckm

Joined Apr 6, 2018
9
Sorry, the ripple current rating is a maximum--higher rating will have more tolerance. No problem there.
 

ebp

Joined Feb 8, 2018
2,332
The high ripple current rating of the capacitor pretty much implies low ESR. I see no reason to doubt that the new cap is well suited for the task.

The transistors are MJE13007. It was originally designed by Motorola several decades ago but is still quite popular for this type of circuit.
 

ebp

Joined Feb 8, 2018
2,332
At this point I have no other suggestions.

I would try running the lamp for just a few minutes, then after disconnecting the AC power and allowing a few tens of seconds for the caps to discharge, very carefully checking to see if the cap(s) have warmed up more than just a little. There is some risk that the caps would hold enough charge to give you a bit of a shock and though not enough to be directly damaging there is always the risk of injury due to rapidly jerking your hand back and bashing it on something or whacking yourself in the face. Wear eye protection in case a cap blows.
 

MisterBill2

Joined Jan 23, 2018
18,167
Aluminum electrolytic capacitors do have a limited life and so it may just have been an early failure. But it is also true that in many cases that a part that fails was overstressed by another part that has failed. So I would check the various other parts with an ohm meter to see if a resistor has failed open or a diode has failed short circuited or a capacitor has failed as leaky or short circuited.

In addition, depending on the ratings of the grow lights, it may well be less expensive to replace the ballast with a new electronic ballast. I have found that purchasing them at Home Depot is the best way to get ones that are not some grossly inferior imports. Internet priced ones may be cheaper to buy but they may be of horribly poor quality.
 
Top