70-watt MH temporarily stopped working--overheated ballast?

[Travis Evans]

Quick question: Do general-purpose-lighting HID ballasts typically have
a built-in thermal protection device of some sort like some fluorescent
fixtures have?

I was operating a 70-watt metal halide ballast (traditional
iron/magnetic type; not electronic) with the proper lamp, but in an
enclosure where it may have built up too much heat and run hot (the
metal box in the luminaire was intended for a physically much smaller
70-watt HPS ballast), though I don't know for certain if this was
actually the case. After several hours of apparently normal operation,
the lamp abruptly extinguished. At first I thought it was a power dip,
but the lamp still would not start even after about 20-30 minutes.

After a while, I opened the luminaire and looked things over. There
didn't seem to be any obvious signs of heat-related physical damage,
though there was a rather strong odor that seemed a fair bit stronger
than normal compared to my previous experiences with this stuff (though
I have no idea whether this would actually be normal or not). After
doing a few electrical continuity checks and after another hour or so, I
reassembled and carefully reapplied power for a few seconds. The lamp
then started and appeared to operate normally.

I'm sure there could have been another explanation, but I can't think
of anything else that seems particularly likely.

 

[Andrew Gabriel]

Quick question: Do general-purpose-lighting HID ballasts typically have
a built-in thermal protection device of some sort like some fluorescent
fixtures have?

I was operating a 70-watt metal halide ballast (traditional
iron/magnetic type; not electronic) with the proper lamp, but in an
enclosure where it may have built up too much heat and run hot (the
metal box in the luminaire was intended for a physically much smaller
70-watt HPS ballast), though I don't know for certain if this was
actually the case. After several hours of apparently normal operation,
the lamp abruptly extinguished. At first I thought it was a power dip,
but the lamp still would not start even after about 20-30 minutes.

After a while, I opened the luminaire and looked things over. There
didn't seem to be any obvious signs of heat-related physical damage,
though there was a rather strong odor that seemed a fair bit stronger
than normal compared to my previous experiences with this stuff (though
I have no idea whether this would actually be normal or not). After
doing a few electrical continuity checks and after another hour or so, I
reassembled and carefully reapplied power for a few seconds. The lamp
then started and appeared to operate normally.

I'm sure there could have been another explanation, but I can't think
of anything else that seems particularly likely.

If it's an internal ignitor lamp, the ignitor has a thermal trip
to stop it trying to ignite a hot lamp (which is unlikely to work,
but would wear out the ignitor). You have to wait for the lamp to
cool down before the ignitor will cut back in.

That doesn't explain why it extinguished, but it might explain why
it won't restrike until the lamp has cooled.

If you want quicker restrike (although it's still not instant), you
need to use an external ignitor and a lamp with no internal ignitor.
The external ignitor generates a higher voltage spike than the
internal ignitor can (which just uses the back-EMF from the ballast).

 

[Travis Evans]

If it's an internal ignitor lamp, the ignitor has a thermal trip
to stop it trying to ignite a hot lamp (which is unlikely to work,
but would wear out the ignitor). You have to wait for the lamp to
cool down before the ignitor will cut back in.

That doesn't explain why it extinguished, but it might explain why
it won't restrike until the lamp has cooled.

If you want quicker restrike (although it's still not instant), you
need to use an external ignitor and a lamp with no internal ignitor.
The external ignitor generates a higher voltage spike than the
internal ignitor can (which just uses the back-EMF from the ballast).

My system (ANSI M98) does use an external ignitor, and normally, the
lamp only needs around 5-10 minutes to cool down enough to restrike,
which is why this event seems so strange. Then again, I've seen MH
lamps do some somewhat odd things at times. I unfortunately don't have
a spare lamp of this type, or else I would have tried that to narrow
things down.

Since I did mess around inside the fixture in the meantime, I suppose
that some connections had somehow come loose and were fixed when I
reconnected wires, but IIRC they seemed good and tight in the first
place the last time I was inside there.

 

[Travis Evans]

I was operating a 70-watt metal halide ballast (traditional
iron/magnetic type; not electronic) with the proper lamp, but in an
enclosure where it may have built up too much heat and run hot (the
metal box in the luminaire was intended for a physically much smaller
70-watt HPS ballast), though I don't know for certain if this was
actually the case. After several hours of apparently normal operation,
the lamp abruptly extinguished. At first I thought it was a power dip,
but the lamp still would not start even after about 20-30 minutes.

Okay, now I have no clue what's going on. I swapped the original HPS
gear back into the luminaire. Now *that* is acting fishy, whereas it
didn't before. I moved everything--socket, wires, ballast, ignitor,
lamp, photocontrol. The only things that are the same are the luminaire
itself and the power cord and switch. And now, the HPS gear works fine
for about two hours then *it* abruptly shuts off for a long time. (The
MH gear, which I put back into its original luminaire (of a different
style and design) has acted completely normally so far.) And I have
verified that it's not something obvious like the photocell seeing too
much light.

I guess the good news is that that presumably doesn't leave very many
things left to be the culprit...

 

[TKM]

Do you have a meter that will check lamp voltage and whether or not the
igniter is working?

It sounds to me like there is more than one problem. The ones that are
likely are (1) an HPS lamp reaching end-of-life (lamp voltage goes up beyond
what the ballast can supply) (2) an igniter with an intermittent problem
such that it doesn't always try to start the lamp when the lamp goes out.

What's the fixture like -- especially the reflector optics. Are the optics
designed for the HPS lamp being used?

Terry McGowan

 

[Travis Evans]

Do you have a meter that will check lamp voltage and whether or not the
igniter is working?

It sounds to me like there is more than one problem. The ones that are
likely are (1) an HPS lamp reaching end-of-life (lamp voltage goes up beyond
what the ballast can supply) (2) an igniter with an intermittent problem
such that it doesn't always try to start the lamp when the lamp goes out.

What's the fixture like -- especially the reflector optics. Are the optics
designed for the HPS lamp being used?

The fixture and lamp are fairly new--bought about five months ago in
May. This is a replacement fixture, though, because the last one did
have an ignitor that failed after around half a year (which in turn was
a replacement for one whose ignitor failed about half a second upon the
very first switch-on. ) So I suppose flaky ignitors can't be
completely ruled out, though this ignitor I think happens to be a
different brand, and this unit was purchased several months after the
first two.

This fixture was (supposedly) designed for HPS since that was what it
came with originally. It's a wall-mount unit with a flat accordion-
folded aluminum reflector behind the lamp and a rectangular plastic
diffuser over it. A while back I temporarily replaced the internals
with a metal halide ballast/ignitor/lamp as an experiment. This worked
fine until a few days ago, where I saw the behavior I described. Now
that I've swapped back in the original HPS equipment, that is showing
the same symptoms, and the MH stuff (moved back to its original fixture)
is not. It's as if there's something about this fixture in particular
that's causing the problem with both setups.

The cycle seems to be roughly 2-3 hours on, 30-40 minutes off, which
makes me want to think it's heat-related.

When the fixture dies for a while, there still appears to be continuity
between line (120v) and the photocontrol, at least, which seems to
confirm that the cause is internal to the fixture.

First I will try eliminating the photocontrol completely and see what
happens since that's easy to do and my understanding is that they can be
fairly failure-prone. Is there any chance that excessively high ambient
temperatures can affect photocontrols (defective or not)? Do any of
them happen to use a bimetallic element or something that could open for
a while regardless of ambient light if they get hot enough?

Failing that, I'll have to start looking at the lamp, ignitor, and
ballast more closely.

 

[Andrew Gabriel]

First I will try eliminating the photocontrol completely and see what
happens since that's easy to do and my understanding is that they can be
fairly failure-prone. Is there any chance that excessively high ambient
temperatures can affect photocontrols (defective or not)? Do any of
them happen to use a bimetallic element or something that could open for
a while regardless of ambient light if they get hot enough?

Some use a bimetallic element for generating the on/off delay
and hysteresis, to prevent rapid switching. I don't know if
external heat can significantly impact the operation of them
(I've never used one of that type - they're not very energy
efficient).

You could bypass it permanently on, and see if that makes the
problem go away.

Failing that, I'll have to start looking at the lamp, ignitor, and
ballast more closely.

Be careful probing around inside with it connected to the mains.
Besides mains voltage itself (and I presume some form of step-up
transformer/ballast in the US?), you also have the high voltage
kick from the ignitor.

 

[Travis Evans]

Some use a bimetallic element for generating the on/off delay
and hysteresis, to prevent rapid switching. I don't know if
external heat can significantly impact the operation of them
(I've never used one of that type - they're not very energy
efficient).

You could bypass it permanently on, and see if that makes the
problem go away.

The culprit appears to be the photocontrol.

Last night I removed it from the circuit and let the lamp run
continuously for about 6-7 hours. It stayed on without incident. I
again ran it for several hours today without the control, again with no
problems.

Then I disconnected power, opened the fixture, and checked continuity at
the black (120VAC in) and red (120VAC out) leads on the photocontrol.
Without power, it was closed as expected. I reconnected it and briefly
confirmed that it appeared to operate normally--the lamp went off when
light hit it and on when it was covered, as it should. Then I let it
run with the photocontrol for a while, and again, after about an hour or
two, it shut itself off.

Right after it shut off, I physically disconnected power, opened the
fixture, and periodically checked continuity across the photocontrol.
It was open, and it stayed open for several minutes, even without power.
It took a while before it finally closed again.

I repeated the above tests just to be sure, and got the same results.

In my exeprience, these controls normally only take a few to several
seconds to close when power is removed or light is prevented from
falling on the sensor. Apparently, when the fixture is on long enough
for this photocontrol to heat up sufficiently, it opens and stays open
for a while for some reason. And it would appear that the other
photocontrol used in my MH gear (which is the same brand, I believe)
does the same thing when used in this fixture. Whether this is a defect
or a safety feature, I have no idea.

The photocontrols appear to operate normally in my other (intended to be
MH) fixture, which is of a different style and design. Maybe that
fixture positions the photocontrol in a more thermally isolated region
from the other heat-producing components. In the "problem" fixture, the
ballast, ignitor, photocontrol, lamp socket, and wiring are all mounted
within close proximity to each other inside a small, closed metal box
directly above the lamp. I would imagine that the internal temperature
could become rather high after a few hours of operation.

Though this wouldn't explain why I never encountered this behavior in
the past... I can't think of anything that would currently be different
internally or externally to the fixture that would have caused a change
in the thermal characteristics. Again, this is even when using all of
the components that came with the fixture, wired and mounted as they
were when originally shipped.

I'll probably just leave the photocontrol disconnected. I always keep
it covered and switch manually, anyway, since photocontrols aren't
particularly well suited for my particular application.

 

[Andrew Gabriel]

The culprit appears to be the photocontrol.

If it's in a NEMA socket, you could just buy a better type.
In the UK, bimetallic ones did exist (and may still), but there
are much better semiconductor types with accurately defined lux
level switching, hysteresis, delays, micropower standby, etc.

However, if you aren't using it, I would suggest disconnecting,
as I think the bimetallic ones use a few watts during daylight.

 

[Travis Evans]

If it's in a NEMA socket, you could just buy a better type.
In the UK, bimetallic ones did exist (and may still), but there
are much better semiconductor types with accurately defined lux
level switching, hysteresis, delays, micropower standby, etc.

However, if you aren't using it, I would suggest disconnecting,
as I think the bimetallic ones use a few watts during daylight.

I'm not sure what type this is, but I do measure a watt or two of power
draw when the photocell receives light but the lamp is deenergized. In
darkness, the power usage drops to zero before the ballast powers up.
It makes a faint click when switching on or off. Almost all of my
photocontrolled fixtures are like this.

This particular one is a small plastic box with a threaded tube for the
photocell, which sticks out of a hole in the chassis and is held in
place by a plastic ring which screws on. There is no socket, but the
leads connect with wirenuts. This is the same type used by two of my
other fixtures (one of a different brand), so it seems standardized.
But as you say, I don't currently have a use for it, so I'll just leave
it removed since I have the fixtures connected to power switches.