At a Tropical Fish store in Milwaukee, Never knew Osram/Sylvania made a 400w Double ended Halide lamp...or any aquarium lamp for that matter. The store manager allowed me to take a peek under the cabinet at the gear, which ended up being an extremely noisy Cooper Lighting 400w HPS ballast. I asked why they have the MH lamps operating on HPS ballasts, the manager had no clue what I was talking about. This store sells these lamps for $129.99 each! ...ouch! Nice blueish/purple bulb though

American 400W pulse-start MH lamps, can be operated with an european 400W HPS reactor ballast (But Na-Sc pulse start lamps will suffer from color shifting on chokes compared to CWA ballasts), but not on the ANSI S51 HPS ballast.
So this lamp is probably overdriven. An overdriven blue indium iodide lamp, looks like a high kelvin MH lamp.
Most aquariums for sale in Israel, are marketed with blue lamps for aquariums, pink plant growing lamps, or just daylight halophosphors lamps, mostly 36W T8 or 18W T8. But there are also several aquariums for sale with LEDs and CFLs.

How much do you think this is likely to affect lamp life? I would expect a risk of arc tube explosion, especially with well-used lamps.


BG

I dont think the lamps are operated long enough for the chances of an arc tube rupture, these lamps are probably replaced once every 6 to 8 months to keep the highest lumen and spectral maintenance over the tanks for optimal growth of corals.

The lamp explosions are not results of lamp wear only, but as well manufacturing defect and mainly ballast failures.
Manufacturing defects are more likely with new lamps, ballast defects with old fixtures or humid environment...

I should imagine the ballasts are dipped in gunk (resin) as our transformers sometimes were when I worked at a transformer and electronics manufacturing plant. Either that or sprayed with varnish.


BG

I can tell you, than these coatings (not only the magnetics, but as well the PCB surface lacquer,...) do improve the humidity robustness, but only when the treatment is defect free. If there is any defect in it, it make the corrosion problem way worse, mainly in almost dry or dry environment: If the humidity manage to find it's way into some crack, it stay there forever, regardless of the outer environment and all the time degrade the thing. So with dry environment and not coated thing, the humidity quickly evaporate and so stop causing problem.

And because it is nearly impossible to check the quality of the coating (only on sample basis, so to check the general batch settings, but that won't remove random defects like surface having an accidental fingerprint on it, when someone pick the board by bare hands before the coating,...), many pieces with defects in the coating pass the quality check...

With really wet environment the coating/dipping does help, in dry environment it cause the failure rate to rise very significantly, even 2 orders of magnitude...

That is the reason, why most of the electronics use no coating at all, as most of it expect dry environment, where the coating would make the failure rate worse.

The transformers and some of our PCBs were dipped in resin, the transformers were then put in an annealing oven. The boards were suspended from drying racks with their wires. Some boards were sprayed (by hand) with conformal coating, however with either coating method some parts on the PCBs were masked (pots, LEDs, connections) .. this would eventually let moisture in!

All PCBs were inspected/tested/calibrated before being sent out.


BG

Many electronics are dipped with that transparent enamel and they are not failing much

But it cracks and it is never perfect. To count on it to stop moisture is futile. As evident from the new Elco/Electra A/C's - They pput the contrtoller circuit, enameled, near the radiator. Dead by corrsion in less than 2 months

The masking is not as big problem - it does not influence how the coat stick to the board.
But the fault start, when the board happen to be not perfectly clean when applying the coat. Then the coat layer allow the water to seep between the coat and board and do the corrosion damage there.

You may do your job well, but still there would be some fraction of faults in the production, this is not avoidable. But in order to ship quality products, you have to have an efficient quality check methods in place to remove them from the line. But if there is no check for some fault, you could be sure you would face quality problems with that fault.

The problem is, there are no known methods to really spot these faults on the manufacturing line. The fact, than the coat vs laminate bond isn't as strong as it should be is not visible, it does not immediately create any cavity, so neither any ultrasound test would discover it, it is still hermetic, so neither any immersion test would discover the fault.
And even the conditions leading to poor adhesion are not detectable: The greasy spot is not directly visible, unless some duct adhere to it. Contaminated lacquer mixture does not have to be spotted either.

These experiences came mainly from the automotive electronic industry, where both coated and uncoated boards are used on different places for already many years in quite difficult (for the electronic and even any machinery) environment: Inside the cabin (dashboard, central console,...) were corrosion problems virtually only with coated boards (well, count faults only during the car's warranty periods, as after that the makers do not get any detailed information about most faults anymore, so they can not track them)...


And by the way: My experience with car corrosion (mainly around voids) is, than the best protection is a hole, where the water and humidity may easily get out. And the same is valid for electronic...

I still prefer the magnetic autotransformer ballasts, I have a few that are upwards of 25 years old and still work like new with little buzzing/humming. The capacitor may have to be changed every 5-6 years but thats about it. Electronic ballasts die far too quickly from my experience, one I had lasted only a few months. Currently they're cheap junk IMO, at least the ones available here. And they really dont offer much energy savings unless you retrofit a few dozen fixtures, but then you're also getting less light too, so efficiency isnt really increased.

The manufacturers are depriving us of good quality electronic ballasts. It is totally possible to make a ballast that will last for many years, they just dont do it, they fight the magnetic instead

I think both are good as long as they are done well. And i have examples of either - an electronic that lasted 20 years in 24*365.25 use, and magnetics that were made so bad that they scorched themselfes and all the wiring due to arcing terminal blocks (cheap quality)

Each technology have their specifics. And the fastest way to kill the ballast is to not understand or ignore them.

Electronic is sensitive to moisture, high voltage (everything above 500..600Vpeak kill it, even when exposed for short time) overvoltages, high surrounding temperature and mechanical vibrations. On the other hand they are very tolerant to mild over/undervoltages (180..280V is of no problem even long term, even when it include the DC component). Their low mass allow to put fixtures on locations, where the materials have very limited structural strength.

Magnetic could well tolerate narrow overvoltage spikes (switching transients,...), but even 10% overvoltage kill them quite soon, when it is present for longer time (e.g. 220V rated ballast used on location, where most of the time is 240V on the input, e.g. close to the distribution transformer). And as they are heavy, some fixture holders may structurally fail quite soon...

Those are rough conditions - In plces where such conditions are expected, it;d be better to use magnetic

However, even electronics can be made to be better suited to those conditions :
 - Completely pot the ballasts in epoxy to protect from moisture
 - Include MOVs in ballasts to protect the semiconductors from minor surges

Do you see many manufacturers doing that ? No, they dont even do the basics right like limiting the inrush to the electrolitic capacitor, all to save on the cost of components

Potting the electronic could cause more problems than keeping it open.
It would be even worse than the lacquer coat, as the one-block material would like to crack:
When potted, the filling material would create a solid mass around the board. If that mass have different thermal expansion than other components, what would cause stresses between the different materials.
After few thermal cycles it would first delaminate the surfaces, so form gaps between different components.
And these gaps would get quite quickly filled by moisture, what would stay there long time and corrode whatever possible...
Then the stresses would crack the connection joints.

Good example are the electronic regulators for medium wattage portable gensets: The electronic board is potted into a cast aluminum "pan" using some black substance (appear quite indestructible...). You may splash the water on it, it keep working, so everything look perfect.
But in all units I've seen the regulator failed after about 2..3 years of light use (about 50..100 starts). And if you go through different discussions, these regulators are treated there as a "consumable material" (well, they cost less then the gas to fill the fuel tank for 8 hour run, so not as big deal, once you have the spares, the tools and the knowledge on how to replace it).