Hello All...

I may have brought this subject up before, but I am running a 250W HPS Luminaire, a GE M250R2 (M2RR25S0A2GMS3358).  I have seen a lot of these in the town I live in (a liberal mix of 85/15 between the M250R2 and Cooper's OVZ) and I have seen 3 or 4 of the M250R2 only achieving between 40-70% brightness.  These particular M250R2s appear to be between 15-20 years old, maybe older, but what causes the capacitors to eventually fail.  I see a luminaire of the same type, possibly same age that is operating at 100% brightness and can't figure it out.  Right now, mine (which is only 6 months old) operates fine, but how long, if ever, until the capacitor fails.  If it should happen, I would obviously just replace the capacitor instead of dropping $189.00 for a brand new fixture, but I seem to have seen this mostly in the M250R2 around here (like I said, mostly what is installed here with about 15% of the fixtures made by Coop) and none in any of the Coopers.  Is this perhaps a flaw in GE's M250R2? Or does it happen to any capacitor, anywhere by any chance?  I just need some answers because it is really bugging the cherries out of me. 

P.S.  If I had it my way, I would've just ordered a M250R2 with 250W MERCURY as opposed to HPS, but we all know that wasn't an option.  I ordered mine fresh from the Hendersonville factory.  It came with the fixture, lens, bulb, and 2 extra Yellow '25' NEMA labels.

I've seen Cooper Ovx failures resulting from cap failures. It's due to them using ge crapacitors.

Capacitor failure could be down to either heat from the choke added to the caps own natural heat and or voltage spikes partly from inductive kick at starting and stopping or from external sources

I've seen Cooper Ovx failures resulting from cap failures. It's due to them using ge crapacitors.

The cap in my M250 was made for GE by some Japanese company.  Would that make a difference?

In a CWA the capacitor is not exposed to any spikes or so, it is really the gradual wear.
The heat is very important factor, difference of just 20degC may make a difference between 5 vs 20 years of capacitor life. So just rather small temperature difference really means big difference in life.
The differences may come from the fixture design itself, but as well from the way, how the external environment influences the heat distribution within the fixture (if the wind redirect the air around the lamp and choke and only then reaching the capacitor, or the opposite; this may easily create a difference of 40degC, so 5 vs 40year life of the same components; even which side of the bowl is colder means a difference in the internal air flows even within a sealed fixture).
Then the capacitor design, it's dielectric loading play a significant role in the lifetime (there the magnetic saturation current tolerance is the most prominent factor with CWA).
And not negligible is the influence of defects and/or other imperfections inside of the capacitors.

Generally the oil filled types are quite robust (the liquid does not tend to form quantum traps as the solid dielectric does; better to say it is not so "easy" for them to form long chains leading to a short circuit failure in a solid state dielectrics), the new compact models are way more problematic. But a solid dielectric capacitor will never leak out any oily mess...

Thats part of the reason I got an oil filled cap to replace the crapacitor. I heard that those dry caps have a tendency of shorting and being crap in general. Plus, it was only rated for 280v, when the cap is on the 240 tap. The new one is a good ol aerovox cap rated at 450v.

It is very important to know, what the rating means. It is common (in Europe) to see a capacitor bearing three or more rating figures, e.g. 250, 350V and 450V, all as "VAC/50Hz". The point is, each rating belong to different use, so to different temperature range and mainly operating life (the 450V belong to just 500h rating, so just a motor start use; the 250V belong to 100khour life rating and 70degC ambient, so a power factor compensation for lighting, the middle one was for 10khours 85degC, so most likely motor run application), all on the same component. The individual rating just copies the performance of that component and it is on you (the one specifying it) to select the proper type for the given use.

So if you won't pay enough attention, the "450V" rating could well be actually less than the "280V" on another capacitor...
This separation isn't anything new, it was present even on old, clearly PCB oil filled components (there were even more steps, as the oil was more sensitive for an acute overheating)

Hmm, the fixture my cap came from was metal halide, and it was tested and certified working. :-?  Also noting that my fu tire has a heat reflection shield.

I really have no idea if I should get a different cap and what kind if any I should replace the current one with.  I will take some pics and post them on here shortly.

I wouldn't expect any reputable maker would ever make that error, it is more a case of some "quick DIY fix" repairs (e.g. "lets take the capacitor from that saw, I do not use it anymore").

So if the capacitor is coming from a light fixture, the voltage it was exposed to is for sure within it's rating for the lighting use.
The higher voltage ratings is usually required with CWA ballasts, where the capacitor is usually exposed to nearly double the ballast OCV, so with 250W HPS it will really lead to 450V or so.
As I understood, your ballast is HX and uses the capacitor only to compensate the power factor, there the rating corresponding to the tap is sufficient, so when the capacitor is coming from a light fixture, where it was exposed to 450V or so and you will use it only at 240V, it will be perfectly OK.

Well, when the capacitor fits into the fixture and does not come too close to become heated from the hot parts.
That could be another problem: The higher rated capacitor will be bigger, so it may not fit into the original place. And it may be, on some alternative place it get way more heat from te other components, it's life will become shortened just by that heat way more, than you gain by the voltage rating margin...

Obviously be careful to use correct capacitance - if you do not compensate the power factor correctly, the ballast primary will be overheating, but I guess that you already checked...

I've thought about the internal temperature of a dayburning outdoor fixture in a hot climate (like Phoenix in the USA) in the summer.  The ambient temperature is already 45C so add the solar gain and the rise due to the fixture running and the internals must get pretty hot.  Are they designed for this worst-case scenario or are all bets off about unit life when they run in the daytime?  It gets up to 40C sometimes in the summer where I live and I've seen street lights that have been neglected and dayburning for years.  There's one MV cobrahead like this that somehow escaped the LED changeover. Ballast coil, capacitor, and ignitor (for HPS and PSMH) temperature ratings all come into play.  It sounds like maybe the capacitor is the weak link, and the coil isn't anywhere near its max temperature rating before insulation varnish degrades.

Matt

Definitely no fixture was ever designed to dayburn for years exposed on the Phoenix summer sun, just forget that.
The design really count on the working photo- or other control system, so it assumes it gets powered only some half hour or so after the direct sun disappears from the sky and all the things already cool off a bit. Dayburning is considered as a fault, so to be sustained only short term (let's say few weeks accumulated overt the complete fixture life), but generally it is assumed the problem gets fixed before the affected fixtures accumulates any excessive wear from that heat.
Of course, the photocells must be designed to operate properly on that heat all the time (and there the components are mainly energized during the light, so during the daytime), but that is just some watt or so, so no significant heat contribution to the internal fixture temperature.

For the heat robustness, the capacitor is indeed the weakest part, even when it usually does not generate any heat by itself (capacitors are rather extraordinary efficient reactors; that's why they are so popular as cheepeese LED or fluorescent nightlight ballasting components). Therefore they should be installed in the coldest spot possible within the ballast box. It does not need much cooling, it really needs just the coldest place.

The core and coil has some heat limitations, but there is technically way easier to make the insulation so, it can endure that heat (you do not need to reach any high capacitance or so, so one of the easiest way is to make the insulation thicker, so less stressed by the electric field; plus you have way more freedom in the material selection than when you need to form a stable high value capacitor). So usually the magnetic part is the longest living component there. But it dissipates quite significant amount of heat, what makes it susceptible to an inadequate cooling or such design errors. But I won't expect that to be really the weakest point in any commercially made fixture.

While we're talking about caps, I saw this video and these ballasts came from NEMA heads. I noticed that one of the ballasts have a cap. These are 100w ballasts, and I've noticed that 100w nema heads produce noticeably less heat than 175w heads. I've been told that line material heads have had CWA designs, but the caps fit down the arm, but this cap doesn't appear to have the shape to fit in an arm. Any ideas, they both tested working and the cap was tested by the owner as perfectly normal.

You will see...
If iust works, it works, only no one knows how long life it still has till it fails.
The video remind me of one extra problem: The corrosion. Seeing the state of the ballasts and the capacitor can, I wouldn't be surprised the rust may eat through the capacitor can wall and so cause the oil to leak, so inspect that on your cap...

And about the cap placement: The arm is indeed quite nice place for such a thermally sensitive device, but as you wrote, it has to fit there.
But the question is, whether the lantyern used standard arm, or if it hadn't been mounted on some wider type.
Otherwise even in the head you may create rather cold places: Putting a metal deflector plates to redirect the hot air rising from the lamp may be sufficient to keep the capacitor colder (the flowing hot air drags the surrounding cold air with it and is it then only that colder air, which reaches the capacitor), still without any need to push it into an arm or so.
But even when operated hotter, just the fact the head is most likely the same design for both 100 as well as 175W "guts", I would expect the 100W equipment will have way more thermal margin, so run colder than rated for...
Anyway just inspect the components properly and then just try them. With an oil filled capacitor you have always take the eventual leakage into account, even when it will be a brand new component (it may be defective from the factory, so you have to ensure the eventual leak wouldn't cause any consequent damage), with dry type capacitor you won't have that problem at all.

I just received my cap. It measures 1.75" by 2.75", the same dimensions as the old one. It will fit just perfectly behind the heat reflector.