I am running a CMH on a 120V HX remote  ballast with ignitor disabled because I have a superimposed ignitor with the lamp. Ballast to lamp is about 15 feet and ignitor to lamp is about 18 inches. It starts reliably but the ignitor hums as if it were pulsing continuously even as the lamp has struck and warmed up. The buzzing stops after 20 minutes which is the timeout built in. Is this normal ?

try to change ignitor, maybe it is defective.

That’s one of my hypotheses. I’ll try a different lamp first because I will have to order another ignitor. In any case, this is not normal behaviour for ignitors.

There is sometimes cases that a lamp may trigger the ignitor to pulse despite it is still working. HPS lamp that about to cycle, my trigger the ignitor when they are about to cycle. MH lamps may do this if they are near EOL as well. I've seen at Hecht park, a MH lamp with a bulged arctube that triggered the ignitor: https://www.lighting-gallery.net/gallery/displayimage.php?pos=-139211

I can see that worn lamps could do that however this one is new. It only has about 25-50 hours at the most. I tried with a new lamp and it still does it. The ignitor and ballast are new as well. Next is to verify that the connections are nice and tight. Barring that, I’ll have to get a new ignitor.

I imagine that this behaviour is at least somewhat harmful to the lamp...

That sounds normal tbh, when i first got my SON street lantern i asked the same question but after working repairing street lights daily they all seem to Buzz when the lamps fully run up.

This is the first time I've dealt with an ignitor without the ballast nearby. I must have mistaken the sound of the ballast and the ignitor when lit. I would imagine that it would be better if it kicks out after the lamp strikes, but it doesn't, obviously.

There is another effect that could cause the buzz:
Superimposed ignitors contain a pulse transformer, with its secondary, with high turns on it, wired in series with the lamp. That means the arc current creates very strong magnetic field in the igniter, driving its ferrite hard into the saturation. And driving into saturation may cause quite strong magnetostriction effects. Because the driving current is mains AC, it crosses zero where the core returns fromthe saturation and short while after that it gets saturated again. That causes pulsed movements with rather high harmonic content, so could easily make very similar noise to the ignitor is firing, although actually it is not.

You may verify that by disconnecting the ignitor Neutral connection during the lamp running. With the Neutral disconnected the igniter is clearly not firing, but the current is still flowing through the transformer secondary. Then the resulting noise can not be anything else except the magnetostriction from the operating current. If the sound does not change with th Neutral being disconnected, you are c,early dealing with the magnetostriction and you could rest safely, the thing is working as it is supposed to without any firing.
If the ignitor is firing all the time, it may wear off prematurely, so better to replace the bulb in time. Plus make sure you are using a correct ignitor type for the lamp (some HPS ignitors may false trigger on the zero cross reignition overshoots of the MH).
The fact it stops buzzing after 20 min could be the ignitor warming up and changing its acclustic properties (the buzz stops spreading from the core), the test with disconnected Neutral may reveal the truth.
But be cafeful with messing up with the live electricity...

Thank you.

I have paid attention more closely on this and it seems that the buzz only starts about 5 minutes after the lamp has started. There is a very brief buzz when first turned on, which would be normal as the ignitor is doing its job, followed by silence and then buzz for about 20 minutes and then silence until the next time.

Next test, when I have time, is to do a hot restrike. If in fact the heated up ignitor is enough to stop the buzz, during a hot restrike, the coil would not cool down completely so the buzzing should be less.

As for the neutral disconnection, I intend to try that some time by wiring in a temporary switch. That way I can do the test more safely.

I will also double check that it is in fact rated for MH. I doubt it would be rated for standard MH only and not compatible with CMH...

One more thing of note is that when this buzzing is going on, the lamp operates normally. No flicker is observed, save for the flicker normally produced by a magnetic ballast.

If your ignitor stopped buzzing for a while during the hot restrike, so it is an anti-cycle ignitor, that don't pulse constantly during a hot restrike.