I recently got a haul of 4 late-60s Canadian General Electric M1000 fixtures and 2 Powerlite B2228 fixtures. (3 M1000s and 1 B2228 usable, the rest were just good for parts). They all came with nearly new Philips India 1kW MH lamps. Unfortunately upon initially testing these fixtures, I discovered that the capacitors in EVERY fixture were gone to the point the Power Factor was 0.15 and current was 20 amps. Obviously quite dangerous and unsustainable.

Being in Canada, the M1000 fixtures all have CWI ballasts. For power factor correction, they used 2 12.5uF capacitors in parallel to create 25uF, with the lamp in series with the capacitors (as usual). There is also some other 2.35uF cap connected to two wires out of the ballast. Not sure what that one is for. Would have to draw a diagram to fully explain it. I replaced the two main run capacitors with a new 25uF capacitor and left that other 2.35uF cap in place (which to be clear is NOT in series with the lamp whatsoever).

Now on to the first test with the new cap. Since I am currently away from home, I only have these 6 nearly new 1kW MH lamps to test these things. As shown in the video below, this is what I get out of it.

(On Odysee since YouTube hates me) - https://ody.sh/uEd40rWE7m

As you can see, the lamp starts up just fine, with a normal WARMUP power factor, and good current draw. This lasts for about 45 seconds right up until the lamp pressure increases and the mercury arc gets tighter..... aaand it cuts out and cycles.

My questions are: is this normal? I've heard of MH lamps being run on M1000 ballasts in the past, but it seems like these ballasts just don't like them. How were these run in the past with MH lamps if this is the behavior I get? How the heck did the previous owners of these lamps run these things with MH lamps AND with caps that did nothing? (Like 20 amps ballast current kind of nothing)? Everything seems normal now up until the lamp randomly and abruptly cuts out.

I can't test this with a mercury lamp until I get home unfortunately, but yeah. Just curious to see if this is normal behavior and if I should be looking for other issues before I spend $32 per new capacitor to get the rest of them working.

Thanks everyone!

Having a large capacitance (if I got your schematics right) connected parallel to the lamp is not the best idea, may kill the electrodes rather quickly! Even if someone found it is okay-ish with mercury lamps which do not de-ionize fast on polarity reversal, definitely a bad idea with MH lamps...

As a side note many years ago I found out that with old and sputtered mercury lamp that is stuck in a glow discharge and does not ignite properly, briefly (!) connecting a capacitor of a few microfarades across the lamp forces it to transit to the arc.

I recently got a haul of 4 late-60s Canadian General Electric M1000 fixtures and 2 Powerlite B2228 fixtures. (3 M1000s and 1 B2228 usable, the rest were just good for parts). They all came with nearly new Philips India 1kW MH lamps. Unfortunately upon initially testing these fixtures, I discovered that the capacitors in EVERY fixture were gone to the point the Power Factor was 0.15 and current was 20 amps. Obviously quite dangerous and unsustainable.

Being in Canada, the M1000 fixtures all have CWI ballasts. For power factor correction, they used 2 12.5uF capacitors in parallel to create 25uF, with the lamp in series with the capacitors (as usual). There is also some other 2.35uF cap connected to two wires out of the ballast. Not sure what that one is for. Would have to draw a diagram to fully explain it. I replaced the two main run capacitors with a new 25uF capacitor and left that other 2.35uF cap in place (which to be clear is NOT in series with the lamp whatsoever).

Now on to the first test with the new cap. Since I am currently away from home, I only have these 6 nearly new 1kW MH lamps to test these things. As shown in the video below, this is what I get out of it.

(On Odysee since YouTube hates me) - https://ody.sh/uEd40rWE7m

As you can see, the lamp starts up just fine, with a normal WARMUP power factor, and good current draw. This lasts for about 45 seconds right up until the lamp pressure increases and the mercury arc gets tighter..... aaand it cuts out and cycles.

My questions are: is this normal? I've heard of MH lamps being run on M1000 ballasts in the past, but it seems like these ballasts just don't like them. How were these run in the past with MH lamps if this is the behavior I get? How the heck did the previous owners of these lamps run these things with MH lamps AND with caps that did nothing? (Like 20 amps ballast current kind of nothing)? Everything seems normal now up until the lamp randomly and abruptly cuts out.

I can't test this with a mercury lamp until I get home unfortunately, but yeah. Just curious to see if this is normal behavior and if I should be looking for other issues before I spend $32 per new capacitor to get the rest of them working.

Thanks everyone!

curious one, usually the main reason a Metal halide lamp cant be used on a mercury ballast from a North American point of view is simply insufficient OCV to strike the lamp when cold, which is clearly not a problem here since the lamp does strike up fine

the way they are cycling almost is as if they are being used on a H34 ballast rather then a H36 ballast


from the capacitor values you give, it does sound like a H36 ballast, but have you double checked that?


if you have volt-meter I would also messure the OCV of the ballast with no lamp installed, it should be in the 400V-500V range


do all 6 lamps exhibit the same across all 4 GE M1000's? or have you only tried 1 lamp and 1 fixture? I almost wonder if the lamp itself has a broken weld thats acting like a bimetallic switch as it warms up, or something along those lines...


hope something in this helps (and nice hault BTW 1Kw Cobra heads are always very neat )

Another thing that I have noticed when comparing North American 1000W H36 mercury vapor ballasts with North American 1000W M47 probe start metal halide ballasts is that both of those ballasts often have nearly the same OCV, which may explain why the 1000W M47 probe start metal halide lamp strikes when power is applied. Interestingly, I have seen that in some 80s GE catalogs, they do mention that 1000W M47 probe start metal halide lamps can operate just fine on 1000W H36 mercury vapor HX autotransformer ballasts or 480V series choke ballasts at around 950W as long as the temperature is above 50 degrees Fahrenheit. It is certainly quite odd to see that the lamps are cycling though.


I think another factor that may cause the lamp to cycle is the presence of a “peaking capacitor” in the fixture since many installation instruction manuals for installing constant wattage ballast compatible high pressure sodium and metal halide retrofit lamps often require the removal of the “peaking capacitor” as in the case with GE’s I-Line metal halide retrofit lamps and Iwasaki’s Sunlux Ultra Ace high pressure sodium retrofit lamps.

In that case, I am pretty sure that the 2.35 microfarad capacitor that you are talking about is the peaking capacitor in the fixture. I highly recommend removing it and testing the fixture without it. Note that removing that 2.35 microfarad capacitor should not cause any harm to the fixture, but make sure to keep the 25 microfarad capacitor in the circuit as it is a series capacitor.

Bit of an update here: I finally got around to testing these fixtures with a PROPER 1KW MV Lamp last night... and what do you know... it worked flawlessly. I DID get one of the 1KW MH lamps to start up on one of the fixtures though, and it ran up FAST, pulling just over 1150w and 10 amps on the ballast in the process. Only ever worked once with only one specific lamp so not sure what that was all about.

Anyways the 1kW MV lamp struck right away and warmed up slowly and controlled. Peaked exactly on spec at roughly 1050w (with a line voltage of <110v mind you because of the long extension cord and remoteness of my storage place). Just glad they all work. So this all must be a thing with these Canadian CWI ballasts and these crappy Philips India 1kW MH lamps. So a bit uneventful but interesting nonetheless.

Just glad I was able to save these and get them working. I'm sure if someone else (that didn't know about lighting) were to have bought them, they would've ended up in the dump.

Here's a vid of one running: https://ody.sh/jdCpm0eA3f

Have you tried running the 1000W M47 probe start metal halide lamp in the lantern WITHOUT the 2.35 microfarad capacitor? If so, I would love to see the results.

No, since I know these all work now, I'm not going to try that. It's clear these things are only happy with MV lamps (as they were designed for) so I'll only be running MV lamps from now on. That and the 2.35uF caps are soldered in (without spade terminals since GE liked to do that) so it would be a real PITA resoldering the wires back on.

Admin edit - Unnecessary quote removed

Thanks for letting me know. I was wondering if you heard about “peaking capacitors” needing to be removed when retrofitting certain constant wattage mercury vapor ballasts to constant wattage ballast compatible high pressure sodium and metal halide retrofit lamps?

Haven't heard of that one. I wasn't even aware what that small capacitor was for before you let me know! I'd love to find a proper wiring diagram for these fixtures as well, coils and all. I could just draw one up as well but it's hard to see the individual taps on the ballast. All that's in the fixture is a small diagram for 120/240v wiring. I'll be sure to try my 940w HPS retrofit lamp in one of these someday too.

I should also mention that one of these fixtures (now the parts unit) came with a fairly modern Sylvania India 1000w MH ballast (probably because the old ballast baked itself from the dead caps). In that fixture, the old 2.35uf cap was still in there, but just disconnected. Out of the 4 fixtures I got, 3 remain completely original, minus the new capacitor in each fixture. That fourth one now has no ballast, lens, or gasket. Just parts.

Admin edit - Unnecessary quote removed

 Some older MV ballasts just don't seem to like MH lamps ! My 1000w Revere Cat Eye won't run anything but MV . I tried both a Probe-start MH and my Iwasaki 940w Retrofit lamp and both do nothing but sit there .

Could those philips india lamps be H34 spec? To my knowledge india uses the European style 1000 watt mercury lamps which have identical parameters to North American H34 lamps.

As far as I know, the only metal halide lamps that I know of to be electrically interchangeable with North American 1000W H34 mercury vapor lamps are the Philips HPI-T 1000 lamps and Iwasaki’s M1000A metal halide lamps along with Venture’s HIT1000W/U/EURO lamps.