I've been working on a few Metal Halide lamp igniters and attemping to repair them.

Why not just replace them? I priced them out first and about fell over. 3/4"x3"x3" potted device. 350.00 bucks. So I did some digging real digging.
I dug out the potting on the igniter to see what 350 bucks was buying. About 20.00 worth of parts is what I was getting.

Have Id a good portion of the items on the board, but I'm stuck on one or two. Tried my best to map out the circuit and made a schematic as best I could.
Will attempt to attach schematic and a few photos of what I'm stuck on.


At this point in time I cant 100 percent identify these. They test with an ohmmeter as open or very high in resistance in both directions. They are about the size of a 1/4 watt resistor. They also seem to have some type of color coding on the leads which may be seen in the photos.


Kinda thinking they might be varistors but I have no idea for certain. If I'm succesfull in attaching my drawing they ID as Z1 and Z2 near the spark gap portion of the drawing.
The input to this ignitor is a pulsed Square wave from an electronic ballast. Voltage into the ignitor is in the area of 100 volts. The output of the ignitor is approx 4 kilovolts. The circuit only needs to function a few seconds to strike the lamp. I expect that is the function of the thermistor near the start of the circuit.

Thanks to any and all help.

 

Looking at the circuit, they could be high voltage diodes, which would read open at the low voltage from the meter. They would form a voltage doubling rectifier.

Looking at the pictures, one of them seems to have a spiral pattern within which might suggest a resistor (or, low probability, an inductor), but that wouldn't make sense in that circuit.

I have never come across components with colour coded leads.

 

The color on the leads is new to me also. Some of the color was damaged when I dug out the potting, but traces still remain. I also broke one of the devices in half when removing the potting. With magnifcation the inner construction at the break is a square with a reflective silver gray metallic look to it. Very possibly semiconductor material. Not sure if i could get a photo of that or not.

Diodes along with the caps do make sense maybe more so than varisters. If they are diodes why not mark them like any other? Not sure if it makes an difference, but these are made by a company in Europe.

The doubler is a great idea and makes sense. Worthy of exploration.

 

The metallic looking square does sound diode like.
Do you have a PSU, say 30V?
You could confirm that they are diodes by connecting them in series with a resistor of around 10k across the PSU. Try it with the 'diode' both ways round and measure the voltage across the diode. If you read the supply voltage one way and less than that the other way it is a diode.

 

Working on doing just that or very similar.. The diode testing. I may end up have to open up a working igniter to test know good devices. Those two could well be the issue.

Assuming a doubling circuit also helps the spark gap make sense to me. Correct me if I am wrong but the input to the doubler would be AC and the output DC with a big ripple in the waveform.
The spark gap could then turn the DC into a high frequency AC for use by the output transformer.
Not recalling if I was able to determine the spec on the spark gap or not. It was made by epcos, but don't recall coming across the part number on their site.

I have four maybe five of these to fix. Lots of research so far. The diagram has helped me tenfold. As have you sir.....thank you, thank you

 

they dont have color coding on the leads doh.
You should crack open one of them or see if you can scrape off the white material.

A diode typically has small metal plates welded and a resistor has a spiral track

 

There is definitely some type of color on the leads. It's on opposing ends and only one shows it in the photos. Black and either yellow or green. I might try and break some of the coating off. It's ceramic like so I doubt it will scrap off.