I've been thinking of hanging a couple of high output strips lately so I pulled out one of my F96T12 HO strips and threw a cord on to test.

Well it started dim and flickery after a while it did warm up some but not nearly to its potential. So I grabbed some test equipment and started checking. Line current settled down to about 1.75A, on a ballast thats supposed to draw 2.14!, and lamp current was a measly 500ma. Power factor seemed a little low as well, in the low-mid 80%'s, but I don't really know what to base that off of.

So what gives? The ballasts are late 90's Advance RS-2S110-2-TP.

Lamp current gotta be 800mA for F96T12. If the lamp looks dim then 500mA is likely what the current indeed is. If you can access the capacitor which is in the ballast (access as in - connect to its terminals from outside), measure its capacitance. Decaying capacitors do cause stuff like this to happen

I was worried it could be the internal cap, having read about others with similar issues. So even if i dig in and find the cap how would I know what the correct value is? Perhaps just try different values and see how lamp current reacts?

Trying random values is not the answer. The problem with this way is, the reactance of the capacitor and ballast coil subtract from each other and dont add up. If you happen to put in there a capacitor of a value such that it and the coil cancel out, you basically connected the lamp to power without a ballast..

As capacitors slowly degrade  by dropping in capacity, i think it'd be best to try to increment the capacitance in small steps (no more than 0.5 uF at a time) (can be done by splicing additional capacitors in parallel with the existing one) and measuring lamp current after each step. Use capacitors of appropriately high voltage rating and application type. 450V motor run capacitors gotta be ok for this

Thats sort of what I was thinking when I said trial and error, start with something close to what the old cap measures and work up.

Any way to know for sure that it is the cap before tearing into it, or is it one of those thing that you just don't know until it open?

If and when I do tear into it I'll be sure to keep track of the results.

Sounds like a cap, but just to be sure, are you using known good lamps? Sometimes they will go mercury starved. Trying to replace the cap really isn't worth it, especially in a late 90s Advance ballast. If they were my lights, I'd just replace the ballasts altogether.

I did try several different lamps. Would the lamps effect the lamp and line current that much though?

I'm not planning on changing many caps either but I thing it would be interesting to take at least one apart just to see whats going on. Even if it doesn't get put back into service.

I don't think bad lamps would affect the current that much, but it's worth a shot. And yes, trying to replace the cap would give you some experience repairing ballasts if nothing else.

Just a little update.
I opened up one ballasts and put it in the oven at 350*. After 30-40 minutes the tar was soft enough to easily dig out with a small putty knife. I was able to find and remove the capacitor. It measured 2.1 uf, I have a bucket full of capacitors that I will dig though and see if I can find some slightly larger values. I may have to use several capacitors in series to get a value that low since most of the ones I have are quite a bit larger in value, many came from an old 1Ø-3Ø converter. Once I find the correct value I will get a single capacitor of the correct value.

Surprisingly, aside from the time in the oven it only took 20, or a little more, minutes to get this far.

I will update again when I have some results with some other capacitor values.

Most of the capacitors used over here (as a separate "can" installed in the Fluorescent lantern) are of the Polyester film type

Most of the time they degrade inside without any visible damage (i mean, when you open the Plastic can and look at the roll of foils, it looks intact and perfectly fine). Only sometimes you would see thin Black lines along the metallised ends (near where the conductive layer burned off). In your case with the thing dirty in remains of Tar you would most likely not see anything

Sometimes they fail catastrophically, and then you can see a bulge or in the more extreme cases the thing is melted or a Pharaoh's Snake of melted Polyester is growing out of it

I dont know what sort of capacitor is used in the US potted ballasts (Polyester or Paper) or how it is packed (just the roll of foils dumped in the Tar ? Or in a can ?)



If you allready opened it, dont pot the new capacitor inside. Take the wire of the inner connection from the coil to the capacitor to the outside, and be free to connect different capacitors to it

Yes I think the simplest way to do this would be to simply cut the wires going to the existing capacitor and extend them outside the ballast, leaving the old capacitor in place but disconnected. It would not be very invasive as it would only require removing a 1/4-1/2" layer of tar to expose the wires.

I did diagram were all the wires connect as well as the location of components inside the ballast.

I'll see how this goes as an experiment and decide weather I purse it further.

In the good ol days the caps were oil filled. Now, some ballasts (esp the cheaper newer residential magnetic ballasts before the ban) use crappy dry film crapacitors. I've tried remote mounting the cap of my 1968 ge bonus line when I was fixing the wires but that plan fell flat on its face when I discovered the windings were toasted by 34 watt lamps.

Another quick update.

I did a bit of testing this morning and came up with some good results. By wiring a hodge podge bank of capacitors I was able to make some values that got the current close to where it should be.

Here the set up


The old capacitor measures 2.1 uf and was obviously low.
After some experimenting I found some workable values
-4uf    lamp current 775ma, line current 2.33A
-3.7uf  lamp current 725ma, line current 2.21A
-3.5uf  lamp current 675ma, line current 2.10A
-3.35uf lamp current 625ma, line current 2.01A

The Input current for this ballast is listed at 2.14A so it seems I must compromise on the lamp current. 3.5uf will get me the closest. Unfortunately that is not a standard value and the voltage that will be across the capacitor is over 500V! 525 to be exact, Most run capacitors are only rated 370 or 440V. I can run two capacitors in series to solve the voltage rating problem, but I will still have trouble getting the value right. Closest I can come up with is two 7.5uf 370 or 440v caps in series which will get me 3.75uf at 740V or 880V. I feel that value is to high especially since most caps have a 6-10% tolerance. Hopefully some searching will unearth a solution.

Here is another pic that shows the original capacitor from the ballast in the foreground, the rest of my pile of capacitors that were not being used to test the light (small cans are 10 and large cans are 55 mfd and yes they are old and most likely filled with PCB's, at least one has the number 1955 on it, though I'm not sure the they are that old so it may not be a date, It can be seen in the pic) Also shown is another small experament I did a while back, I have searched for a cheap way to re-seal/varnish transforers, so that is a H39 ballast out of a nema head. It has been varnished with Polyester resin. I could not find any info regarding the max operating temp after it had cured so I thought I would simply try it. I did two ballasts and have put a few hours on the one not shown and it seems to be holding up OK but more time will tell. I'll try to post it in it's own thread once I have ran it enough to find the result.

Are you sure the capacitance of your capacitors really is whats written on them ? Used capacitors may be a great deal off



With capacitors in series, you dont have control over the DC voltage of the center point between them (and it does not have any effect on their AC functionality)

So one which have a little internal leakage will tend to discharge to 0V DC (i.e. the voltage across it will be symmetrical sine wave), while the other might charge (i.e. the voltage across it will be sine wave riding on some non zero DC level). And then the peak value of that voltage (when the sine is at its max in the same direction as the DC bias) might still be too much for the capacitor...

DC might be introduced into the system when a lamp is slightly rectifying (but probably not a lot of it), or significantly if a lamp is EOL

Resistors across the capacitors will provide better control over how the DC splits between them



Or try kV-range capacitors out of microwave ovens..

Are you sure the capacitance of your capacitors really is whats written on them ? Used capacitors may be a great deal off



With capacitors in series, you dont have control over the DC voltage of the center point between them (and it does not have any effect on their AC functionality)

The values I listed were measured, not calculated. Although the individual values were close what was measured.

You are correct, a permanent bank should have resistors to keep the voltage balanced between each capacitor when they are used in series. This set up was only for testing so I did not bother to dig out and connect any resistors.

I do have some HV capacitors that I removed from microwave ovens. I am not sure what the values are though I'd have to look at them.