Author Topic: Metal halide voltages  (Read 1032 times)
Shawn_Laughlin
Newbie
*
Offline

View Posts
View Gallery

Metal halide voltages « on: August 29, 2022, 09:29:57 AM » Author: Shawn_Laughlin
I’m trying to design/build a probe and pulse start metal halide light circuit and I’m confused on volt specs of different areas.
 
Do all metal halide lamps operate,once struck, on 120v ?

Is the output voltage of a 400w and 1000w metal halide ballast the same?

If both of those questions above are yes, is it just the ignitor for pulse start lamps that generates differing voltages to accommodate what the lamp needs to get it to start?

If I’m wrong with any of this or you don’t mind adding a few things to explain maybe what I could be missing I would greatly appreciate it?

Im building some underwater lights and since I live in the USA 🇺🇸 Superimposed ignitor are hard to come by and i believe that’s the only way I can remote mount a psmh lamp from the ballast at the distances i need to be able to use said lamp type. It wasn’t until I ordered ignitor a and lamps straight from the manufacturer that I noticed volts not being the same in some specs and now I’m question if What I got is going to work. The ignitor will be close to the bulb but it’s the differing volt specs that have stopped me from going any further.

Hopefully I haven’t confused the heck out of the people that read this. Lol . Thanks
 
Logged
Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: Metal halide voltages « Reply #1 on: August 30, 2022, 01:32:25 AM » Author: Medved
Dunno what level is the question.

All discharges enforce their voltage to the circuit, so to get stable and predictable performance, you need a ballast maintaining the designed current while letting the arc dictate the voltage it wants. The voltage is then the property of the lamp arc, based on how it is designed.

The ballast uses to be formed by a voltage source represented by its open circuit voltage and a series impedance, which then limits the current with the rated load voltage to the designed level.
Now for the discharge to operate in a stable and efficient manner, the ballast OCV can not be too high (the reactance would have to handle large V*A product, which uses to be the base for its losses), but not too low, as when is too much less then double the maximum operation arc voltage, the arc becomes unstable (assume AC feed). Or with some lamp (SOX, probe MH) the minimum has to be a bit higher to facilitate a reliable lamp ignition.

Now pulse start MH originated from Europe, where the 230V mains with a series choke, so the simplest and most efficient ballast form for a given frequency, it yields a ballast with the mains voltage, so the 230V, as its OCV. It was too low to ignite the then common probe start MH's without an ignitor device. So as the ignitior would have to be needed anyway, it was cheaper and yield more efficient lamps, when the lamp was designed without the starting probe and then use a HV ignitor for the initial ignition.

So because the OCV is 230V, most pulse start lamps were designed to match such ballast properties, so to have as high as possible arc voltage (to keep the ballast losses minimum), the lamps are designed with around 90..130V arc voltage. Higher power lamps use to have more stable arc (lower zero cross reignition voltage), so could be designed with higher arc voltages than the lower power ones. So a "70W" has its arc voltage around 80V, a 1kW could go as high as 130V arc.
That means the ignitor threshold voltage (the voltage the ignitor has to see in order to generate the ignition pulses, so the threshold used to determine the presence of the arc) for all of them at about 210V. That means technically there could be one ignitor type suitable for pretty much all the lamps, provided it can generate the 3.5kV pulses (the lower wattage lamps need), as well as support arc currents of up to 10A (the highest wattage lamps operate at). But such ignitor would be too expensive to use mainly with the low power setups, so usually the range is split to about 2..3 ignitor types, the lower power ones (usually 35..150W) and medium power (150..400W) and higher power ones (400W and above).

Related to operation at 120V: 120V is in any way too low OCV for the MH discharge to operate, you need a ballast with OCV at least 220V for practically all of them. Because there are no other lamps with such ballast properties (the power levels at the arc voltages and the OCV) in the US, you need to use the dedicated MH ballasts for them anyway. These use to combine the step up transformer (120V -> 220V OCV), a series ballasting reactance (it uses to be integrated into the transformer, in the form of a HX autotransformer), plus the ignitor (usually a 220V pulser connected to the tap of the ballast winding and so utilizing the ballast secondary as the 200->3.5kV step up pulse transformer for the ignition). Then you need nothing else than the ballast (complete with the pulser ignitor) connected to the lamp. No extra superimposed ignitor.

Now when you want to place the ballast long distance from the lamp, the common build in ignitor within the ballast won't be able to feed the high capacitance of the long cable by the high voltage pulse, then (per Code) you are supposed to use a ballast with its ignitor vbeing designed as a "long range", so suitable for the high cable capacitance load. Or you may just deactivate the ignitor within the regular ballast and then place an European style superimposed ignitor close to the lamp (unlike the ballast it uses to be small and light weight and dissipate very low heat, so it could be easily fitted into the fixture even when the complete ballast won't fit).
Then you select teh ignitor the same way as it would be selected in the European style 230V series choke circuits, so with the range covering the lamp type (important are the ignition voltage to be above what the lamp needs as the minimum and current rating above what the lamp current is). Given the ignitor type in your picture is rated at 1kW and has a pulse voltage of 5kV, it may work pretty much with all pulse start MH's.
Logged

No more selfballasted c***

WorldwideHIDCollectorUSA
Member
*****
Online

Gender: Male
View Posts
View Gallery

HID, LPS, and preheat fluorescents forever!!!!!!


Worldwide HIDCollectorUSA
Re: Metal halide voltages « Reply #2 on: August 31, 2022, 12:36:20 AM » Author: WorldwideHIDCollectorUSA
What Medved essentially says is that the ignitors on ballasts essentially have OCV requirements. This means that you can still use a European superimposed ignitor on a ballast with a 120v mains input voltage as long as it is an autotransformer ballast with an OCV of at least 200v. In my experience, I have used some European 220v-240v superimposed ignitors to run North American 400w M135/M155 pulse start metal halide lamps on North American 400w H33 mercury vapor CWA and HX autotransformer ballasts designed for a 120v mains supply with supposedly good results since the ignitor only cared about the ballast’s 220v OCV and NOT the 120v mains voltage alone.

You can actually easily buy European superimposed ignitors if you search “tanning bed ignitor”, “facial ignitor”, “tanning bed starter”, or “facial starter” on eBay.
Logged

Desire to collect various light bulbs (especially HID), control gear, and fixtures from around the world.

DISCLAIMER: THE EXPERIMENTS THAT I CONDUCT INVOLVING UNUSUAL LAMP/BALLAST COMBINATIONS SHOULD NOT BE ATTEMPTED UNLESS YOU HAVE THE PROPER KNOWLEDGE. I AM NOT RESPONSIBLE FOR ANY INJURIES.

sol
Member
*****
Offline

View Posts
View Gallery

Re: Metal halide voltages « Reply #3 on: August 31, 2022, 07:21:01 AM » Author: sol
(...) Or you may just deactivate the ignitor within the regular ballast and then place an European style superimposed ignitor close to the lamp (unlike the ballast it uses to be small and light weight and dissipate very low heat, so it could be easily fitted into the fixture even when the complete ballast won't fit).(...)

I have been running a setup exactly like this with a M98 ballast without ignitor in a box and the superimposed ignitor close to the 70W metal halide lamp. I've been using it nightly for 3-4 years now, with excellent results. There is about 10 feet of cable between the ballast box and the lamp although at the time I was considering putting the ballast further than that which is why I have wired it like that.
Logged
Shawn_Laughlin
Newbie
*
Offline

View Posts
View Gallery

Re: Metal halide voltages « Reply #4 on: September 01, 2022, 01:04:57 AM » Author: Shawn_Laughlin
Okay I read up on ocv and now I understand why the voltage is so high when there is no load. I was confused thought when my 400w probe start lamp fired off and I put my meter on the output and there were only 23v but, as it warmed up the voltage slowly creeped up to 115 i believe it was.

The 1000w ballast I have for my pulse start says lamp can be no greater than 15’ from the ballast not the igniter. So I just went for broke and mounted the ignitor like maybe two feet from the bulb with 40’ of 14/3 sjooow cord between them two and the ballast and to my surprise it fired off and looked to be running like it’s suppose to. Nothing unexpected happened in other words and all seemed to be fine.
My question here is would you expect the same results but the igniter might not last as long?
Or do the manufacturers just not test them out that far and give you what is normal for the distances between a ballast,ignitor and lamp on let’s say a remote mounted ballast on a light pole/street light?
I really need them to work with 75 to 125’ of cord and ima do my best to figure a way out.

The picture below is what I did.
Logged
WorldwideHIDCollectorUSA
Member
*****
Online

Gender: Male
View Posts
View Gallery

HID, LPS, and preheat fluorescents forever!!!!!!


Worldwide HIDCollectorUSA
Re: Metal halide voltages « Reply #5 on: September 01, 2022, 01:17:22 AM » Author: WorldwideHIDCollectorUSA
In North America, both 1000w M47 probe start metal halide ballasts and 1000w M141 pulse start metal halide ballasts have a much higher OCV compared to 400w M59 probe start metal halide ballasts in which the 1000w ballasts have an OCV between 375v and 440v and the 400w M59 probe start metal halide ballasts have an OCV between 270v and 330v. In that case, you might need to have a special 380v-415v superimposed ignitor for your 1000w M141 pulse start metal halide ballast such as this one due to the higher ballast OCV:

https://www.ebay.com/itm/295124139426
Logged

Desire to collect various light bulbs (especially HID), control gear, and fixtures from around the world.

DISCLAIMER: THE EXPERIMENTS THAT I CONDUCT INVOLVING UNUSUAL LAMP/BALLAST COMBINATIONS SHOULD NOT BE ATTEMPTED UNLESS YOU HAVE THE PROPER KNOWLEDGE. I AM NOT RESPONSIBLE FOR ANY INJURIES.

Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: Metal halide voltages « Reply #6 on: September 01, 2022, 02:20:16 AM » Author: Medved
The 1000w ballast I have for my pulse start says lamp can be no greater than 15’ from the ballast not the igniter.

With a semiparallel ignitor, the family common in the US, the ballast secondary winding is essentially a part of the ignitor (its HV step up pulse transformer), so distance "ballast to lamp" is essentially the "ignitor to lamp". What US folks use to call an "ignitor" is in fact not the complete ignitor, just its low voltage pulser part, it generates only about 200V pulses on its output. In order to form the complete ignitor (i.e. a device generating the high voltage), these 200V pulses have to pass a transformer to step them up to the required 3..5kV or so, we are talking about 15..25x voltage step up. In the US style gear the ballast secondary winding is used for that, therefore the ignitor pulser is always mounted with the ballast (the distance between the pulser and the step up transformer has to be really short, a foot is already a stretched maximum). The European style superimposed ignitors use a dedicated transformer device, which is together with the pulser assembled together in to what you then buy as the "superimposed ignitor" device.

Now for the distances what matters is the length of the loop between the pulser and the transformer primary (its inductance, as it has to carry high current pulses and the inductance would restrict that current; but that is within the superimposed ignitor or in the US the ballast assembly, so you don't have to worry about at all) and then the cable carrying the HV pulses (its capacitance). So with semiparallel (the US ballasts) or parallel (here belongs also all the lamps with an integrated pulse generators, like the selfstarting HPS, MV retrofits or so - there the ignitor is within the lamp assembly) ignitors that is the whole cable between the ballast and lamp, with superimposed ignitors it is the distance between the ignitor and the lamp (the line between the ignitor and ballast does not carry any high voltage, nor high current pulses)



So I just went for broke and mounted the ignitor like maybe two feet from the bulb with 40’ of 14/3 sjooow cord between them two and the ballast and to my surprise it fired off and looked to be running like it’s suppose to. Nothing unexpected happened in other words and all seemed to be fine.
My question here is would you expect the same results but the igniter might not last as long?

If the iginitor that was within the ballast is deactivated, it is perfectly OK, there is no reason why anything shouldn't last its normal lifetime.


 

Or do the manufacturers just not test them out that far and give you what is normal for the distances between a ballast,ignitor and lamp on let’s say a remote mounted ballast on a light pole/street light?
I really need them to work with 75 to 125’ of cord and ima do my best to figure a way out.

The picture below is what I did.

The cable length between the ballast and the superimposed ignitor is really limited only by its resistance, the same way as with any other electrical circuit. So if it would be unusually long, using a thicker gauge to compensate for the length will solve the resistance problem.

If you exceed the length of the cable carrying the HV pulses, the pulses will be attenuated, so they won't reach the rated peak voltage. It by itself won't be directly a problem for anything, except it will make the lamp harder, so less reliable to ignite. But that means the ignitor would have to make way more pulses till the lamp start will be successful, so the ignitor will wear out faster, so indirectly it may affect the ignitor life. But I guess more stress would come from the eventual hot restrikes, when the ignitor will be pulsing for 10's minutes continuously till the lamp cools down enough to be able to restrike. But definitely the lamp starting reliability will be suffering, maybe to the extend then some of the lamps may fail to start (so essentially reach EOL for this setup) with fewer hours than their normal rated life.

And for the completion if you make the wires between the semiparallel ignitor pulser and its transformer (ie the ballast coil) too long, it will reduce the pulses so affect the lamp ignition reliability as well, but as the pulse currents there would be reduced, it won't affect the ignitors life that much. But as I wrote, this uses to be of no problem, as all the wiring is either made by the ballast assembly or the ignitor (in case of the superimposed one) manufacturer.
Logged

No more selfballasted c***

Shawn_Laughlin
Newbie
*
Offline

View Posts
View Gallery

Re: Metal halide voltages « Reply #7 on: September 14, 2022, 09:36:29 AM » Author: Shawn_Laughlin
I stepped up my cable length with right at 93’ between the ballast and the igniter and then the short short distance between the igniter and lamp. Even with 93’ of cable it fired right up with ease with no flicker whatsoever just a smooth ignition is how I would say it looks.

That’s all great but for it supposedly being a 1000w lamp It’s not as bright as my Plusrite 01053 400w probe start. comparing it to the 1000w version of the Plusrite it’s like holding a match next to a bonfire. So I’m not sure if the lamp that is a sample from a supplier on alibaba.com is just very low quality junk or the plusrite ballast and parts arent doing there job?

I’ve had good luck with plusrite but You never know.
I read the volts between the x1 x2 and x3 leads and it was about the same at 95v.
But I can’t find info and what it should be so I’ll keep searching till I find.

I have another question regarding the gfci supplying the power to ballast if anyone cares to reply?
I have a voltage leak bc I’m reading 75 volts from water to ground but doesn’t trip the gfci. Shouldn’t it trip?
Logged
Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: Metal halide voltages « Reply #8 on: September 14, 2022, 04:49:03 PM » Author: Medved
For the cable it depends what is its cross section, so its resistance. But with about 30m I would not expect problems. To be sure, you can measure the voltage both on the ballast, as well as on the ignitor end of the cable. Would not expect more than a couple of V difference, but even 15..20V should not alter the lamp power in any significant way (it does affect the ballast input power, of course).

Can you measure the arc current? To check whether it is in the ballpark where it should be (Varc * Iarc product should be about 1.1x the nominal power; assume mainly the voltage is measured by a true-rms meter). If not, the ballast is the culprit, with CWA most likely the capacitor is worn out and so has lost part of its capacitance.

For the GFCI:
What you mean by "a voltage leak"? You measure a voltage (vs the ground) on a wire/structure not connected to ground which should be insulated? That is normal, mainly on longer wires, it is a capacitive coupling from the phase wire. When grounded, the resulting curent should be bellow 1mA, but because the V-meters tend to have 10MOhm input, you can easily see there half of the mains, yet it is of no problem.
In fact this is the reason, why the normal sensitivity of the GFCI is 30mA (some special ones have 6mA sensitivity, but these can not be used with too long cables, as the capacitive current from the long cable could be enough to trip them even when there is no fault, mainly when there is some high frequency disturbance on the mains; usually are restricted to serve just a single socket intended for a single small appliance, like a hair dryer or so in the bathroom).

But if the ground leakage exceeds the 1mA, you have a problem somewhere. A 30m cable can not generate that much of a capacitive leakage, even when at ~230V (the ballast OCV ballpark). And you should address that problem, as even when nothing else, if it is real fault leakage, it is after all a damaged insulation, so at least a fire risk even when all the grounding is well connected.
Logged

No more selfballasted c***

Print 
© 2005-2024 Lighting-Gallery.net | SMF 2.0.19 | SMF © 2021, Simple Machines | Terms and Policies