The problem is with my HPS floodlights (all of them)

They ahve a GRP enclosure which had tiny holes in it, and i sealed them by coating the entire internal side of the cover with RTV silicone. I tested the floodlight with a hose at high pressure, and no water got in at all

However, when it is raining the floodlight absorbs moist air from the surroundings. Then the moisture condenses on all internal surfaces - that includes the ballast and makes it rust

It also condenses on the windings of the coil and i think this can be a problem if there is a hairline crack in the coating of the ballast's wire (its in the place where the wires are bent out of the coil to the connection, so i have a reason to think there might be tiny crack in the place where the wire is sharply bent)

I thought to paint the ballast to protect it (it allready began rusting), but i think that just paint can overheat on it when working, and can thermally isolate the ballast causing it to overheat, and i am unsure if it is ok to paint on the windings

How can i protect it otherwise ?

Normal "sealed" fixtures do not work, as when the air inside heat up, it pushes some air out of the fixture and when it cool down (mainly during the rain), the internal pressure drop and suck in the wet air (and frequently even the water). I've seen many "sealed" fixtures with the water pool and heavily rusted components inside...

So you have two options:
- Make the sealing capable to hold the pressure (either overpressure or vacuum), so no air/moisture could go in/out of the fixture when the temperature cycle. My experience is, then this does not work for long time, mainly if the fixture is regularly used during the rain.

- Ensure all the electrical installation is under some kind of "roof", so the water can not drip on it during the rain. But otherwise keep the fixture well ventilated. The trick is, then the changing air dry out the fixture very quickly, so the rust does not have much time to build up, mainly when the fixture warm up during use. When such fixture is regularly used for long time (like street lights,...) the heat generated by it's components keep all of them dry. This is the reason, why open fixtures endured for so long, way longer then "better protected" newer "sealed" ones...

They are in occasional use - maybe few days in a month. Most of the time they are disconnected

The water on the compnents looks like spray of small drops, so i guess that at least some of the water is not condensating on the roof and dripping down, but condensating directly on the components

The ballast core allready is a bit rusted, and when i touch the rusted spot i have rust left on my hand - which mean that it is actively deteriorating - and i'd like to stop it

Whenever there is sun for a while after rain, i can see that the glass of the floodlight is getting foggy areas on it from condensing moisture, and when i open it the same thing is everywhere not just on the glass

It is impossible to seal this floodlight, and impossible (without pretty much making holes in it, which i am not going to do) to provide good ventilation. Yet i see ones of this model working for many years without any problems (they are mid 70s), so i guess that it can be solved with the floodlight as is

Whenever there is sun for a while after rain, i can see that the glass of the floodlight is getting foggy areas on it from condensing moisture, and when i open it the same thing is everywhere not just on the glass

I think this is clear poor ventilation. And it is this moment, when the corrosion progress the fastest...

On what places they are used? Windy place help, as the wind help to replace the air inside, so help to dry it out. On the other hand "protected" spot mean no wind, so the moisture stay inside for longer time.

Keeping the fixture ON for longer time help a lot - maybe that is the biggest difference (the "other" fixtures you compare with are ON each day for the whole night, so they dry out by the regular heat)

My experience really is, then the best corrosion protection is the hole, where the water (and moisture) may go out quickly.

Otherwise what do you want to do? Pressure sealing is impossible, bigger holes you do not want to make, only look more carefully on the "other" fixtures and try to search for "five differencies", why your piece deteriorate quicker then the others.

Try spraying the ballast with silicone oil, it will repel water and prevent rust, but will not evaporate off. Not Q20, WD40 or such but a silicone mould release agent.

I don't think the silicone is the best choice - as far as I know the silicone oil absorb quite lot of water (for an oil).
The "WD40" or so are indeed not much useful, as they evaporate rather quickly.
But some higher temperature parafine would work well, but you would have to heat up the ballast in order to let the parafine to soak in all the pores and so don't allow the water to enter.
If you have access to a vacuum pump, impregnating with shellac (or similar resin) would help. But it have to be dried by a vacuum, to ensure it will dry out from the solvent and still remain properly soaked by the resin...

I dont have access to something like that. The maximum vacuum i can get is with a wine bottle pump

Will ordinary paint or glass paint work ? My concern is mostly the ballast overheating with the thermal isolation the paint provides

If the paint is not in an excessive layer, it's thermal resistance would be nearly negligible. But the question is, if you would be able to cover all surfaces, include those in the cavities without the vacuum.

The only option is solvent-free, low viscosity lacquer. I know internal voids in car's body are coated with some kind of wax by submerging the heated metal works into bath of melted wax. The trick s, then such coat solidify and form a protection layer without the need of any material (solvent) leaving the place. So the coat is even without ay bubbles and quite strong for the material used.

There are no much deep cavities in there - most of them are still full of the original vacuum potting material of the ballast

Is it ok to apply window sealing RTV silicone direcly to the coil in a small spot, where the wires are exiting from it ?

It is a spot on the side of the coil, where each drop touches multiple layers of the coil, so the isolation between the wires is stressed at the voltage between distant layers of the coil (possibility to breakdown)

And the wires from the coil are exiing there and are sharply bent (esp. the one from the inner layer and ignitor tap) and i suspect that there may be hairline cracks in the isolation of the wire in the bend, so the drop might even get to the copper through the crack

Do not use silicone sealers, unless they are rated for electronics use. The most common silicones use an acetoxy reaction to cure, which releases acetic acid ( the vinegar smell) that corrodes copper very quickly. Spray it with a thin coat of clear varnish, but check first on an unimportant area that it does not react with the existing paint.

Is turpentene-dissolvable paint ok ?

 i know they have insulating varnish that you can dip the ballast in.
and it will help and pentrate the inerworking of it. and it will fill any cracks about to become shorted turns.
and it also help on rusting of iron core.

Is turpentene-dissolvable paint ok ?

Test by applying a drop to somewhere like the side, where it is covered by the original paint, but is not going to worry you if it bubbles. If the paint goes soft, or bubbles then it is not going to work. Best would be to use some clear spray, or just use superglue to make a clear overlay, though you will get a white bloom around it.