So I have this 100 watt Metal Halide that I put together using of an old Mercury fixture about 2 years ago.
This is the equipment that I bought http://www.1000bulbs.com/100-Watt-Metal-Halide-Ballast/38884/

Tonight the light is dead. I replaced the bulb, nothing.
It will power an Incandescent bulb . It will even power a 100 watt Mercury bulb just fine.
Im thinking the ignitor or capacitor ?
Does anyone know how to test these items so I know which to replace.
Or am I better off just replacing everything with higher quality parts?

Thanks

MH will not light in an MV ballast. its the other way around. MV will work on a probe start MV ballast.

The MH MIGHT light up, it wont work well cause MH uses a slightly higher voltage to start up.

So MV can turn on with the higher voltage, while the MH cannot start with lower voltage.

MH will not light in an MV ballast. its the other way around. MV will work on a probe start MV ballast.

The MH MIGHT light up, it wont work well cause MH uses a slightly higher voltage to start up.

So MV can turn on with the higher voltage, while the MH cannot start with lower voltage.

Ummm yes I know this. Please reread my question

It might be or bad ignitor or a short (or reduced dielectric strength) inside the ballast winding.

Some trouble-shooting questions:
It is HX with capacitor for phase compensation (parallel to primary) or CWA (cap is in series with "cold" end of secondary)?
If you tried it with a MV, did you measure the lamp current? Should be ~1.2A
Did you measure the OCV? Should be at least 220V for HX and >250 for CWA
If you power it without a lamp, what is the "sound" of the ballast?
Does it smell with ozone?

If the voltage and current are OK and the ballast only hum or is quiet at all, the cause is likely in bad ignitor.
Too low OCV mean interturn's short. The consequence is the ballast get hotter then usually.
If the ballast without the lamp "creak" (like some sparking), the ignitor is good and the fault is in the winding.
If both ignitor and voltage/currents are OK, then there is some "low breakdown" damage, yielding some sparking. Consequence is some ozone emission. It might be only too small clearence on the ballast output, slight bending of leads might solve the problem.
Otherwise such isolation issue might be in the ignitor as well...

Yes,try a replacement ignitor,if you don't have a spare, here's one for 35W to 150W PSMH .

It might be or bad ignitor or a short (or reduced dielectric strength) inside the ballast winding.

Some trouble-shooting questions:
It is HX with capacitor for phase compensation (parallel to primary) or CWA (cap is in series with "cold" end of secondary)?
If you tried it with a MV, did you measure the lamp current? Should be ~1.2A
Did you measure the OCV? Should be at least 220V for HX and >250 for CWA
If you power it without a lamp, what is the "sound" of the ballast?
Does it smell with ozone?

If the voltage and current are OK and the ballast only hum or is quiet at all, the cause is likely in bad ignitor.
Too low OCV mean interturn's short. The consequence is the ballast get hotter then usually.
If the ballast without the lamp "creak" (like some sparking), the ignitor is good and the fault is in the winding.
If both ignitor and voltage/currents are OK, then there is some "low breakdown" damage, yielding some sparking. Consequence is some ozone emission. It might be only too small clearence on the ballast output, slight bending of leads might solve the problem.
Otherwise such isolation issue might be in the ignitor as well...

Thanks for the info.
I figured the ignitor is bad. It use to make a cracking sound like a spark when the power was interupted. Now it doesnt do this.
I went to Home Depot and bought a new 100 watt Metal Halide fixture. I put my original bulb into the new fixture. After having to replace the socket so that a non-safety exploding type bulb would fit. . The bulb cycles on and off. This is probably what kiled the ignitor in the first place.

Can a cycling bulb really kill the ignitor?

Can a cycling bulb really kill the ignitor?

Yes, at least some designs, the ignitor has a significant power dissipation when active and the question is, how are affected components dimensioned. Usually only short burst is anticipated during design.
The situation might be worse for cycling then missing lamp, as when the lamp cycle, it heat-up the lantern space, so the ignitor's ambient. Given ignitor components temperature rise, they might exceed their abs-max temperature.
On top of this the ignitor has limited life of pulsing burst operation.

To fight with this problem, higher quality ignitors use auxiliary timers doing at least one of two things:
- When the ignitor is bursting more then few seconds, it stop doing so for given time (~5minutes for HPS and 15minutes for MH). This causes the ignitor does not burst, when the lamp cool down to restrike temperature. Side effect the lamp is not reheated by tiny sparks when the arc could not yet establish.
- The cumulative burst time is limited (to ~15min for HPS), so when the lamp start to cycle, the time is eaten-up by few cycles and then restrike attempts are stopped till power is ON. Power OFF-ON cycle reset the timer.

More intelligent ignitors (or electronic ballasts) do monitor voltage pattern, so are more accurate to distinguish between real "cycling" lamp failure and other reasons for lamp extinguishing (like power micro-cuts, brouwn-outs, etc...)