That's a head-scratcher for sure. But, I'd bet it's the heat, not the humidity, that is causing problems. Are all the lights coming off the same breaker in your panel? If so, I would suspect you have a loose connection somewhere in the wiring that gets worse with heat. Can you measure the voltage at the fixture when you have problems? I'd guess you're not getting a good 120v to the ballasts, which is why they are having such a hard time firing. It doesn't take much for a bad connection to bring down voltage to the point that ballasts would have a hard time getting the light to fire.

I'd start by checking voltage during the problem times, and then start tracing each light back to the breaker to see if you get similar voltage readings at each j-box. The only role that humidity could play, that I can see, is that it may have caused some corrosion that in turn made a connection sensitive to heat.

I have separate breakers for each set of lighting fixtures. They all exhibit the same problem. Like I said, this only happens when the air is humid. Definitely a head scratcher.

How about a dehumidifier?

You left out a couple of things. Long (6-8 foot) or short (4 ft or less) Are the circuits and fixtures grounded? Are the ballests mounted on any kind of standoff that might insulate them from the fixture housing? Any rust that might not allow for a ballest contact to the metal housing?

Try this. Take the covers off the lamps, if there are any, and rub one of the bulbs lightly in one direction after turning them on and they light. If the lamps come on when you do this it is USUALLY a grounding issue - not always.

Many of the ballests that will operate older style lamps are picky about being grounded to the housing, and having the housing fairly close to the lamp. The fact you are having the same issue with new ballests is sort of telling you the ballesta aren't the problem.

Ballests operate flourescent signs and outside lighting in downpours with near or at 100% humidity pretty reliably. And, ballests are pretty forgiving about voltage as they are designed to provide a constant voltage and current output over a good sized range

The newer ballests that work in plastic housings are obvious exceptions, as are the high output varieties that are often found in signs. And, for what it's worth, many of our standard, excellent quality made in the USA stuff is now made overseas and is no longer reliable stuff. I have a lot of new florescent fixtures go back for problems, and replacement ballests arent immune.

Running the dehumidifier only makes the basement warmer since it generates heat. I only run it when the A/C is on and the basement is really humid. That helps get the humidity out quick.

Bulbs: 4-footers

Circuits and fixtures properly grounded and tested.

Ballasts mounted in brand new sockets. No obstructions from making good contact as far as I can tell.

Almost Everything about the fixtures are brand new. Brand new ballasts, brand new 14 gauge NM source wiring, and brand new sockets. I did have to borrow some of the blue/red/yellow wire from the old ballasts in order to connect the new sockets because the wires on the new ballasts were not long enough. All splices were properly made and covered with wirenut. The bare ground wire of the NM cable is connected to another bare wire (via wire nut) which is securely screwed into the fixture.

I will try the rubbing test today.

But if this is a grounding issue, why does it only happen when the basement is humid and warm?

Last resort I guess I have to call an electrician.

Almost forgot...The lamps are less than a month old. I replaced all basement bulbs in July.

Just thought of something. What about the light switches themselves?

I have three separate switches in the basement. Switch A controls one fixture. Switch B and Switch C each control two fixtures. All switches are on separate circuits and separate breakers.

During these "episodes", Switch A sometimes won't start the one and only fixture it controls unless I play with the switch a little.

With Switch B (a three-way), at least one of the fixtures will turn on (usually) at the first flick of the switch. After playing with the switches a few times I finally get the second fixture to fire up. At first I contributed this to one fixture having a weaker ballast than the first but now I'm not sure. Sometimes neither will go on unless I play with the switch a little.

With Switch C (the one with the two fixtures that have new ballasts), I am always getting one fixture to go on (the one with the single ballast). The second one usually goes on after one or two flicks of the switch. Both of these fixtures have new NM 12-gauge wire that run into the same j-box where the power from the light switch feeds into. In the j-box, all blacks are connected together, all white together, and all bares are together as well as grounded to the box via a separate bare wire and screw. So if grounding was an issue, I think that both fixtures would exhibit the same problem. Only the one that has two ballasts is having the problem.

Well, you've replaced about everything else, so perhaps $4 more for good new switches would be a good investment.

These fixtures don't have starters, do they?

Last night I decided to take another look at the light fixtures that had the new ballasts installed.

First I opened the junction box to verify that I was properly grounded.

I measured ~113V across hot to neutral and the same for hot to ground.

Then I measured 0 V across neutral to ground.

My meter is not exactly the greatest. But it appears that the box setup is good.

I then did the rub test like I said it would and sure enough, just the lightest stroke made the bulbs ignite.

As I continued to inspect the two fixtures I noticed that there were some other BX cables in the vicinity of one of the two fixtures that were making contact with the fixture (on the top side of the fixture and on one of the lateral sides). I had never considered that these would affect the fixture. See photos below to understand what I'm talking about.

I also noted that the same fixtures was quite close to a copper pipe that carries hot water for my heating system when in use. I'm not sure how close that particular fixture was to that pipe prior to me removing the fixture for ballast replacement but it was a hair away when I reinstalled it and since I remounted it in the same exact screw holes its probable that it was very close then too. Again, I did not think anything of it when I reinstalled it in the ceiling this past weekend.

So, I started to insert wood shims to attempt to shield the BX cable from the fixture. I also cut out a piece of clear plastic flexible tubing (relatively thick) and shielded the pipe away from the fixture.

All of the sudden the lights appeared to go on with every flick of the switch. I figured I solved the problem. So I put drop ceiling light panels back in place and called it a night.

This morning I went to try it out and the problem was back. I know it sounds like it should have no effect on ballasts but the temp this morning in the basement was 77 degrees and humudity was at 77%.

When I contacted Advance Transformer on Monday, tech support told me that if I have a grounding issue with magnetic ballasts, high humidity will sometimes cause bulbs to not fire up.

Last night I figured I had removed any doubt that my fixture wasn't grounded properly. What else can I measure?

This still doesn't explain why two other circuits are experiencing the same problem. Everything was fine up until about 2 weeks ago. I haven't done any electrical work in my basement since March when I installed recessed lighting in front of my pantry.

By the way, in order to replace the ballasts I had to entirely remove the fixture from its mounting location in between joists because the screws (that hold the housing for the ballasts and wiring) are located on top of the fixture. I guess in the old days they didn't make pop-out housings like they do now.

{{gwi:2099548}}

{{gwi:2099549}}

{{gwi:2099550}}

{{gwi:2099551}}

{{gwi:2099552}}

The grounding issue you probably have has to do with the case of the ballest making good contact with the metal of the light housing. The other stuff, which is (or should be) grounded won't matter if it touches or not. Get something like a Dremel with a wire brush tool and remove paint on the housing and ballest where they touch (like at the mounting screw holes. Do the same thing where the ground wire and ground screw mount to the fixture.

There's no question that the ballasts are making good contact with the housing. In fact there are slots on the housing that exist for the purpose of "sliding" one side of the ballast in while securing it the other side with a screw.

As far as the paint finish of the housing (which is in perfect condition) affecting the ground, well that I've never considered. If the ground screw of the fixture goes through a hole in the fixture, then I see no need to remove paint surrounding the screw since the thread of the screw is making contact with bare metal.

I will however try to expose bare metal for the ballast to housing contact if it comes down to that. Thanks for the suggestion.

"As far as the paint finish of the housing (which is in perfect condition) affecting the ground, well that I've never considered. If the ground screw of the fixture goes through a hole in the fixture, then I see no need to remove paint surrounding the screw since the thread of the screw is making contact with bare metal."

You would be amazed how often this actually IS a problem. Get into automotive computer systems and this is the cause of some very, very strange problems. Auto manufacturers are now going to bare metal grounds using precious metal connectors. I wouldn't dismiss this lightly. It can cause some strange problems...which you definitely have.

OK, I will add that to my troubleshooting list. Good to know. Thanks.

BTW, I got a response from the ballast company. They suggested cleaning the bulbs with glass cleaner as a first step. They stated that humidity can cause a film to build up on the bulb. I will try this later today.

Could it be the fact that I use daylight bulbs? Do these type of bulbs take more starting power than cool white bulbs? I'm not familiar with chemistry and physics of fluoresent bulbs so I have no idea. Could it be that the older ballasts (as well as the newer ones) aren't designed for these types of bulbs?

Cleaning the bulbs did not help.

I have found ballast grounding problems from paint on the fixture. The "paint" may in some cases be powder coating which is typically thicker. I usually first suspect poor contact of the wire leads to the lamp holders and poor contact of the holders to the lamps. New ones does not rule out this possibility.

"I usually first suspect poor contact of the wire leads to the lamp holders and poor contact of the holders to the lamps. New ones does not rule out this possibility."

Wouldn't that mean that the lamps wouldn't work at all?

".....I measured ~113V across hot to neutral and the same for hot to ground..........My meter is not exactly the greatest...."

A cheap DMV will indicate voltage even when the connection is bad.......

The ground provided by BX is sometimes "iffy".....I think you have a high resistance connection to ground........

Try substituting a standard 60 watt light bulb for the meter in the above test....

The ballast housing needs a VERY solid connection to ground.
Use an external tooth lock washer between the ballast and the housing and tighten the screw hard.
The finish on the ballast and the fixture housing must be penetrated with solid metal to metal contact to ground the ballast.
Without a good ground the capacitance will be so poor the bulbs cannot strike.
When you touch the bulbs your body adds more capacitance to the circuit and the arc strikes.
If you actually have BX (pretty rare) it is a poor ground.
AC cable will have a bonding strip (thin aluminum is common) in contact with the armor and usually grounds fine if the correct cable clamps are used.

It may or may not be related to the symptoms, but I'm concerned about the fact that I can see some of the yellow and blue ballast wires going to the lamp tombstones in the photos. Those should not be exposed in this type of fixture. I think something isn't assembled right.

????? WELL ????????

Follow- up....

So here's what I did on the weekend. I took down all ceiling tiles that surround each of the five fixtures in my basement. The first thing I noticed was that (due to poor jobs by previous electrical contracts that pre-date my ownership of the home) a lot of the electrical wiring was not properly fastened (i.e. stapled) to the joists above. So after spending a few hours correcting this, it made the mess above the ceiling a lot easier to look at.

I then took a long good hard look at the three fixtures (that I did not change the ballasts for). There was nothing touching the fixtures at all. In fact all three fixtures were different types, probably because of the very low and varying clearance of the drop ceiling.

Then it dawned on me...in early July I replaced all 18 bulbs in all five fixtures. Maybe I got a bad batch ?? They all came from the same box and they were all PHILIPS daylight bulbs. So I started swapping bulbs between fixtures trying to figure out if a particular set of bulbs was causing the problem. I found that this worked to a certain extent. Some bulbs worked well in all fixtures, some worked well in one but not another. I couldn't determine for sure which were flaky and which were not.

Next step...I went to Lowe's and got a bunch of GE daylight bulbs. I put them into the fixture that had the hardest time going on. Presto, the new bulbs worked every time I turned them on. However all of the other fixtures that had the Philips bulbs all happened to start working at the same time too (probably because they were already warmed up) so I wasn't getting too excited.

I decided to call it a night at that point. For the time being I decided to leave the ceiling exposed until I finally nailed the culprit down.

Saturday I was away for most of the day but all lights worked fine in the morning and also when we returned in the evening. Sunday morning was fine as well. Sunday night was a different story.

We came home and I turned on the main lights (2 fixtures). NONE of the bulbs went on. I had to flick the switch a few times to get one fixture to finally go on. Eventually the 2nd fixture fired up. As soon as that happened I went to the other side of the basement to turn on the fixture that had the GE bulbs in it. It fired up on the first shot and all subsequent shots. Usually if the main lights don't go on, the fixture (that currently has the GE bulbs) won't go on for sure...but it did!

So the next step is to replace all bulbs with the new GE bulbs and give it a trial run for a few days. I'm not saying for sure that its the bulbs but I can't deny the proof at this point.

The only thing I don't like about the GE bulbs is that they don't instantly light like the Philips bulbs do...they sort of fade-in. The package of the bulbs mentioned something about EcoLux technology. Not sure if that has anything to do with it.

To be continued.....

oh, and by the way (in response to davidr)

That's how they sockets were wired before. There is no other way to wire them. The insulation is exposed but no bare wire is. I realized Friday night that these particular types of fixtures (the ones in the photos) are designed specifically to be installed above a drop ceiling grid. There are special brackets that allow the fixture to rest securely on the crossbars. They are not designed to be mounted with screws. I realized this when I was working on the troublesome fixture that now has the GE bulbs. In the two fixtures that I replaced the ballasts in, the brackets had been sawed off and the fixture had been screwed into pieces of wood that were secured in between the joists. Otherwise the brackets would have never cleared that copper pipe in the photo above. I will try to take more photos of the other fixtures tonight.

Tell you what I'd do, especially since you're contemplating replacing all the lamps. I'd rip out all the old-school magnetic ballasts and replace with modern instant-start, no-flicker electronic ballasts. Then, fit T8 lamps - more light for less power, and almost always a better quality of light.

I think that will take care of things once and for all.

FWIW, GE Ultramax electronic ballasts are DEAD SILENT - no hum at all that I can detect. They also seem to have a slightly softer start to the lamp than most electronic ballasts, which may extend lamp life.

PS - low-mercury lamps (T8 or T12) need a few hours of operation before they will reach full brightness - sort of a break-in period. After that they'll work normally.