Fluorescent ballast repair

[Peabody]

I have a houseful of, actually nine, flourescent ballasts for T12 F40 pairs, dating back to 1972.  They work fine during the winter, but in warm weather they don't turn on very fast, sometimes taking several minutes to come on, if at all.  But they aren't actually dead because if things cool down they work fine again.  Putting in new tubes doesn't change anything.

I'm embarrassed to say that I don't know what's inside a fluorescent ballast that's almost 50 years old, but since an increasing number of my ballasts have exactly the same problem, I have to assume that there's a particular part or type of part that deteriorates over time and causes this effect.  Actually, I'm a bit puzzled by the temperature profile I see, because I would have expected that it would be less likely to start when cold, but it's actually just the opposite.

So having nothing better to do, I thought I might try to repair them.  Replacement would cost about $25 each, and switching everything to no-ballast LED, which is what I would actually have to do, more than that.  So if I can repair the ballasts, I'd like to do that.  120VAC here, if it matters.

It's been demonstrated scientifically that if someone who doesn't know anything about classical music asks you to identify a classical piano piece he has heard, 72.97% of the time the correct answer is Beethoven's "Fur Elise".  Similarly, if an electronic circuit's performance begins to deteriorate, but does not completely fail, I think the reason is usually "capacitors" with perhaps even higher odds than Beethoven.  Is that the explanation here?  I mean, other than gradually failing solder joints, it's hard to imagine what else it could be, particularly when multiple ballast examples show exactly the same symptom.

Does anyone here have any experience looking inside ballasts?  Are they repairable (within reason)?  Is there anything else in there besides capacitors that's likely to explain what I'm seeing?  I guess I'm kinda assuming I will be able to get inside these ballasts, and that they aren't potted.

The other 27.03% of the time it's split between Mozart's Turkish March and Debussy's Clair de Lune, which might correspond to diodes and switching transistors.  But nobody ever hears Brahms, and it's never a resistor.

[rhodges]

What the heck, I will throw in my opinion until someone really knowledgeable chimes in. Your ballasts are probably just inductors and the problem may be the "starters". The starter is (was) a small cylinder with two electrical lugs. When you first apply power, the starter gives power to the tube filaments so that they can get the gas ionized. Then after a few blinks, the tube is ready to go and the starter shuts off. I think there may be a tiny heater wire and thermo switch in the starter.

Hypothesis: If a starter is defective, maybe its thermo switch can work when the ambient temperature is low, but a higher temperature might make the starter "think" the tube is already going.

[EOC_Jason]

Those old magnetic ballasts probably aren't worth the effort. Every one I've seen has been potted in like tar or some other substance, so cleaning that out alone would be a major PITA and I don't know if it could be done without damaging anything...

[edavid]

Look again, Amazon has a 10 pack of T12 electronic ballasts for $90, or singles starting at $11 each.  Well worth it to get rid of ballast buzz.

[Gregg]

I'm a big fan of Fulham Workhorse ballasts; they seem to be the best reliability and quality for the $. 
As previously stated, most old magnetic ballasts are potted in tar and not worth the effort to even try to repair and even if you succeed, you'll still have an inefficient buzz box ballast with line frequency strobe effects.     

[bob91343]

I have gone through taking apart tar potted inductive components.  Yes it's a pain and a very messy one at that.  But that's what you go through to get inside.

However, as someone mentioned, it's just an inductance and is unlikely to be the cause of the problem.  The starter is more often the source of trouble.

I have a box of starters I have had for decades.  I presume they are still good.  If I remember correctly there is some kind of gas filled tube in there that breaks down and allows the ballast to provide an inductive kick sufficient to strike the ionization.  I never really understood fluorescent lamps.

My workbench lamp has a manual starter.  You hold the button down until the tubes glow and when you let up it lights properly.

[Peabody]

Thanks very much for the replies.  Since posting I've done more research, and I now have a different picture of what my options are.

First, you certainly can get replacement T-12 ballasts, and even entire fixtures, at a reasonable price, but it's not clear how much longer T-12 tubes will be available, or at what cost.  What I read is that T-12s aren't being made or imported anymore.  So that's a question going forward.  KInda like what I face with my R-22 AC compressor.

At the same time, the cost of LED replacements has come down a good bit since I last looked.  And it appears that what makes sense is to switch over to direct-mains-powered LED tubes.  While they make LED tubes that work with ballasts, I don't know why I would want to continue to be subject to ballast issues.  What makes more sense to me is the ones that run directly off AC, with line and neutral at one end, and dummy pins at the other end.  And it appears that my existing T-12 tombstones are even the right type for these LED tubes.  So I would just remove the ballast and toss it, and rewire the tombstones directly to mains.

I find this at the Home Despot:

https://www.homedepot.com/p/Viribright-4-ft-40-Watt-Equivalent-Ballast-Bypass-Cool-White-4000K-Single-End-Power-T8-LED-Tube-Light-Bulb-12-Pack-750329-12/302327728

And that's before my 10% veterans discount.  It would do my entire kitchen.  Then I might pick something warmer for the bathrooms.

What do you think of this option?  It seems like a fairly low-cost way to put all the obsolescence issues behind me, with really easy conversion, and maybe save a few bucks a month on power going forward.

[james_s]

Are these rapid start or are they old enough to be preheat start with the separate little can starters?

In the US, fluorescent ballasts for lamps larger than 24" are autotransformers. The older preheat start type are typically a lead-lag configuration that puts a capacitor in series with one of the lamps so they run out of phase to reduce strobing. Rapid start ballasts run the lamps in series with an isolated aux winding that provides cathode heat in the middle ends of the two lamp string.

The most common cause of fluorescent lamps not starting when using rapid start ballasts is that the reflector is not properly grounded. This is particularly problematic during times of high humidity.

I use T8 tubes and electronic ballasts in most of the fluorescent lights I have, but I do like the old preheat start T12 lights. The ballasts are less efficient but they tend to run the tubes a bit hotter so they're nice and bright.

[Peabody]

Are these rapid start or are they old enough to be preheat start with the separate little can starters?

Mine don't have the separate can starters, so I assume they are rapid start.  Attached is a pic one one of them.

The most common cause of fluorescent lamps not starting when using rapid start ballasts is that the reflector is not properly grounded. This is particularly problematic during times of high humidity.

I would never have guessed that one.  And I assume the ballast case has to be grounded to the fixture too, presumably through the "U" at one end where the screw goes.  I thought the ground was just for safety, but I see now that the ballast itself says the tubes need to be near ground.  Well, easy enough to make sure the ground is there.  Maybe that will be the solution.  Thanks very much.

[PaulAm]

I ran across a deal for direct wire LED tubes a while ago and put them in my workshop.  Wish I had bought more, they're great.  I think it was around $6-7/tube.

I bought some LED tubes on closeout at one of the big box stores that worked with a ballast.  That was a waste - they only work with some ballasts and are not compatible with the ones I have.  Stick with the direct wire types.  Those tubes are going in the donate pile.

I've also started replacing the HIDs in my shop with LEDs.  Those can get pricey.

[edavid]

I ran across a deal for direct wire LED tubes a while ago and put them in my workshop.  Wish I had bought more, they're great.  I think it was around $6-7/tube.

I bought some LED tubes on closeout at one of the big box stores that worked with a ballast.  That was a waste - they only work with some ballasts and are not compatible with the ones I have.  Stick with the direct wire types.  Those tubes are going in the donate pile.

Well, I had the opposite experience.  I first tried some direct wire ("bypass") tube replacements, but they had unacceptably slow turnon.  I'm glad I figured that out before replacing all the tombstones in the fixtures (a common requirement for the bypass products).

Then I bought some cheap Philips ballasts and ballast compatible tube replacements, and the results were excellent.

Both types of tube replacements were "Utilitech" brand from Lowe's, and cost about $4 each, but I'm sure they have different models by now.

[james_s]

Yes that says RS, so rapid start. It's an autotransformer ballast with the lamps in series. The yellow wires come from an isolated winding that provides cathode heat, the pairs of red and blue wires come from the ends of the secondary and have a tap to get a few volts between each pair to heat the cathodes (filaments) in that end of the tubes. It's a high power factor ballast so there will also be a PFC capacitor in there and usually there's a small capacitor across one of the lamps to aid starting. The whole thing is potted in tar to make it quiet so unless you want to go through the very messy and stinky process of melting out the tar there is not much you can do.

RS ballasts rely on having a grounded plane near the lamps for reliable starting since they lack the inductive kick produced by a conventional ballast and starter arrangement. Capacitive coupling helps to ionize a small amount of the argon in the tube at which point the lamp will ignite. You may find that passing your hand over a stubborn tube will cause it to light up, that's a pretty good indication that you have a grounding problem.

[oolloo]

I'm not sure why you would want to use fluorescent lamps.  4 years ago I replaced all my lamps with ballast compatible led T12's  , fixtures  (they're 40 years old ) home depot was not sure, they  told me the led tubes aren't compatible , because they're not as thick.  That was in 2015

Tried to sell me all in one led soldered fixtures $200 each (Canada same $ to usd as Au. )

At Costco Canada I
Spotted a 2 tube led pair,  ballast compatible  lamps for $17, it  consumes almost half the electricuty, brighter and instant on.  Your thinking "but my ballast is the issue"

I know,  there are dozens of videos on YouTube regarding eliminating the ballast to save exactly half the electricity .  LED doesn't require it.

My tubes have paid for themselves many times over , $9  A piece vs $2-4. Up front   I bet they're cheaper now, even dollar stores sell led light bulbs now.

[6PTsocket]

I have gone through taking apart tar potted inductive components.  Yes it's a pain and a very messy one at that.  But that's what you go through to get inside.

However, as someone mentioned, it's just an inductance and is unlikely to be the cause of the problem.  The starter is more often the source of trouble.

I have a box of starters I have had for decades.  I presume they are still good.  If I remember correctly there is some kind of gas filled tube in there that breaks down and allows the ballast to provide an inductive kick sufficient to strike the ionization.  I never really understood fluorescent lamps.

My workbench lamp has a manual starter.  You hold the button down until the tubes glow and when you let up it lights properly.

I have a magnifier circline lamp clamped to the bench. When the 3 position rocker switch (off, on, push to start) failed and it was hard to get a replacement I installed a starter and a regular on-off switch.Works great.

Sent from my SM-G900V using Tapatalk

[bob91343]

Very nice!  Can you give details as to just how you installed a starter?  I find the push-and-hold method a bit tiresome and wouldn't mind having it start automatically.

[soldar]

The most common cause of fluorescent lamps not starting when using rapid start ballasts is that the reflector is not properly grounded. This is particularly problematic during times of high humidity.

I can confirm this. Many people do not know this.

I am switching over to LEDs as I change lights here and there but I still have a bunch of old ones with inductor ballasts and they work well.

The old system of ballast and starter worked pretty reliably and if it wasn't working the ballast was the least suspect and the problem was much more likely to be with the tube or with the starter. And the starter could be replaced with a momentary switch for testing.

Cheap inductor ballasts got hot and wasted energy but in general they were OK. Another problem was reactive energy and you would need capacitors to counteract this.

[soldar]

Very nice!  Can you give details as to just how you installed a starter?  I find the push-and-hold method a bit tiresome and wouldn't mind having it start automatically.

The starter is just a switch which is normally open when the tube is on but will close momentarily when it is off.  It is very ingenious. Have a look here https://home.howstuffworks.com/question337.htm

[james_s]

Very nice!  Can you give details as to just how you installed a starter?  I find the push-and-hold method a bit tiresome and wouldn't mind having it start automatically.

You wire it just like any other preheat fluorescent lamp. The starter connects to one pin on each end of the tube. The ballast is a choke coil in series with one of the AC line wires feeding the remaining pins. The ballast is a series impedance just like the resistor used to ballast an LED. The starter is nothing more than an automatic switch that shorts across the tube allowing the cathodes to draw power and heat up, then the starter opens and you get a spike that strikes the arc.

You should be able to just follow the wires from the switch, disconnect the two to the start section and connect them to the starter.

[james_s]

Their description of how a starter actually works is wrong. The starter is a normally open bimetal switch in a glass tube containing argon or neon. When power is applied a glow discharge strikes inside the starter and that heats the binetal strip causing it to close. This shunts the current away from the discharge in the starter so it goes out while the cathodes are heating at the same time. The starter now cools off and the contacts pop open, creating the spike that strikes the tube. The running voltage of the tube is lower than the striking voltage of the open starter so nothing more happens.

That is until the lamp is worn out and the cathode emission drops causing the lamp voltage to rise. This is why worn out tubes often cycle and flash on preheat systems. This is also why the starter has to be matched to the tube type, there are two common types, one for shorter tubes on choke ballasts and one for longer tubes on autotransformer ballasts (in 120V land.)

[schmitt trigger]

I used to work for a ballast manufacturer.

An electronic ballast will provide 10 to 20% efficiency gains. Also, because they operate at high frequency, the 50/60 Hz flicker is completely gone. And you have no ballast buzz.

As others have stated, the ballasts are filled with sand-filled tar. Excellent environmental protection and noise damping, but extremely difficult to remove. And you will (mostly) only find an inductor. Some two-tube ballast do have some paper capacitors, to create an arc-current phase shift, in an effort to reduce powerline flicker. These capacitors, like all capacitors which carry significant amounts of current, will degrade over time.

What I find puzzling is your report that at colder temperatures the ballasts work better. Usually is the other way around.

[james_s]

What I find puzzling is your report that at colder temperatures the ballasts work better. Usually is the other way around.

Colder temperatures typically mean lower humidity, this is particularly true in some parts of the world where summers are extremely humid. High humidity bleeds off charge and makes it harder for the lamps to ignite. Due to the multiple factors involved, warmer days may exhibit the problem more than colder days, especially if the lamps are indoors where they may never get cold enough to have cold weather starting troubles but they do see the high summer humidity.

[steve1515]

I've been thinking about upgrades to my old T12 and T8 lights also. The choice that I've ended up with is get new electronic ballasts and go all T8 lamps or go with direct wire LED tubes. The problem I always noticed was that the florescent lumen value was always more then the equivalent LED.

For example...
The regular T8 linked here is listed as 2600 lumens.
https://www.homedepot.com/p/Philips-32-Watt-4-ft-Alto-Linear-T8-Fluorescent-Light-Bulb-Cool-White-4100K-2-Pack-479667/303811442

The LED replacement linked here is listed as only 2100 lumens.
https://www.homedepot.com/p/Philips-4-ft-32W-Equivalent-Cool-White-Linear-T8-LED-Light-Bulb-472878/303810126

It seems that every time I start looking at LED replacements, I find that equivalent LED lamps don't have the same light output. Or at least that's what it looks like to me when reading the specs. Why don't they make LEDs that have the same amount of light output? Or, is this not how the lumens value works?

[soldar]

Their description of how a starter actually works is wrong.

Yes, I noticed that too. The first explanation is wrong but there is another one below which is correct.

[schmitt trigger]

The important parameter to compare is the lumens per watt.

And the color rendering index -CRI.

The original LED lamps had poor CRI. My wife made me change the lights back to CFLs in the vanity mirror lights, as the makeup would look odd.
Nowadays, newer and more expensive units actually have a very good CRI.

[Peabody]

@schmitt trigger:

What I find puzzling is your report that at colder temperatures the ballasts work better. Usually is the other way around.

Yes, that was my reaction.  But it's absolutely predictable, for multiple ballasts - now up to four that behave exactly that way.  They misbehaved late last summer, worked perfectly all winter, and are now acting up again.  I have a drop ceiling, so the light fixtures are in the relatively unheated/uncooled area near the attic, and thus closer to the outside temperature.  I don't think it has anything to do with humidity.  We've had a week of rain, but with a bit colder temps, and all the lights work fine.  But when it warms up, they'll go flaky again and I'll be able to test the grounding fix.

@steve1515:

The explanation I've seen is that in the case of LED tubes, all of the light is directed down into the room, whereas a significant portion of the light from a fluorescent tube goes up into the fixture and has to bounce back down, with considerable loss.  So the net effect is that the LEDs actually produce as much room light  even though they are rated at lower light output.  The videos I've watched on Youtube seem to confirm this.

The big issue, as schmitt trigger says, is the CRI, which I had not fully considered.  Apparently LED tubes with low CRIs can make things look really goofy.  But I don't know what the CRI of a cool-white fluorescent tube is.