How to protect Metal Halide Bulb from burning

[Md Ahsanul Haque]

Hi
I have around 50 metal halide spotlight installed to light up a big area. All the lights are looped from 4 (all lights divided into 4 groups) main lights connected to 32A breakers. But after a while the bulbs are broken. I checked all the cables and there is nothing wrong with the cables. I am guessing too much current going in them. How can I protect these lights without rewiring. Can I use resistors to limit the current flowing in them?
Regards
Ahsan

 

[Md Ahsanul Haque]

Sorry forgot to mention these lights are 250W/ 2.15A rating

 

[davenn]

hi there
welcome to EP

just confirm that they are operating on 115VAC mains system ?

Dave

 

[Bluejets]

My guess is excessive voltage drop.
You'll find there is a range they work in...outside this and you get problems.

 

[Colin Mitchell]

A resistor such as 3R3 or 4R7 at 10 watt will increase the life enormously.

 

[Gryd3]

A resistor such as 3R3 or 4R7 at 10 watt will increase the life enormously.

250W rated lights at over 2A result in an effective bulb resistance of ~60Ω.
Adding an additional 3.3Ω will decrease the current by about 5% and will cause the resistor to dissipate ~12Watts which is over the 10W rating you recommended.
Using the larger value resistor will reduce current flow a little more, but will cause the resistor to dissipate ~16Watts.

Number may be slightly off, but the results are still above 10W depending on the use of 110 - 125V...
This should not be a recommended solution until the current setup is looked at for other possible problems.

Can the op please clarify the use of any other equipment that may be connected to the lighting?
http://en.wikipedia.org/wiki/Metal-halide_lamp#Ballasts

 

[Colin Mitchell]

Use a 2R2 at 10 watt will make an enormous difference as the starting resistance of the lamp is about 10 ohms.

 

[Gryd3]

Use a 2R2 at 10 watt will make an enormous difference as the starting resistance of the lamp is about 10 ohms.

Metal Halide requires a striking voltage of 5K+ V depending on lamp temperature.
I don't know what kind of lamps you thinking about, but adding a resistor here is bad idea without knowing the whole story.

Resource

Since the discharge reacts to increasing lamp current
with falling voltage (which would cause the current to
rise indefinitely until the fuse blows or another part of
the circuit fails), the lamp current must be limited by
a ballast during operation.

I would like to know if the op is using an external ballast or a light with one integrated.
I would also like to see a diagram for how they are wired.
50 Lights on a 32A breaker does not compute, unless the op has multiple breakers in use.

 

[Colin Mitchell]

Disregard my posts. I was thinking of the Halogen lamps.

 

[Bluejets]

Metal Halide requires a striking voltage of 5K+ V depending on lamp temperature.
I don't know what kind of lamps you thinking about, but adding a resistor here is bad idea without knowing the whole story.

Resource


I would like to know if the op is using an external ballast or a light with one integrated.
I would also like to see a diagram for how they are wired.
50 Lights on a 32A breaker does not compute, unless the op has multiple breakers in use.

I agree......
These lamps will draw maybe 4 to 6 times full load during the initial strike.
There are times on large installs where I have had to fit a small PLC which would go through a start sequence, limiting the start of each group to a particular timing.
In that arrangement, it didn't matter which start buttons were pressed and then the start was on a priority basis with programmed delays.
Supply authorities can get a bit cranky if one loads up their distribution network like above , not to mention the problems occuring here with the lamps.

 

[Md Ahsanul Haque]

Sorry if i was not clear before so here is the spec again. Each of these metal halide bulb is around 2.15A/250W/220V rating. They are connected through ballast integrated inside the housing. No external ballast is being used. They are connected to 220VAC and no not all of them are connected to a single 32A cb. They are divided into 4 groups and each group with 12 lights maximum. As the main supply is 3 phase 415VAC i divide each phase for 12 lights except the last phase have 2 groups of total 17 lights. So you can say each phase is 12-12-17 lights and each coming out from a 32A cb.
About the connection let me explain:
From 4 32A cbs, 4 lights are connected from the 4 groups. The next light from the same group is looped from the previous light and so on. I checked there is no problem with the cable but many lights are burning down vary fast. Some of the ballast also broken after 4-6 months of using.

 

[Bluejets]

What size cabling supplies the lights and how far from the main distribution switchboard are the lights, especially the last fitting.?
Still sounds like excessive voltage drop to me.
A diagram of your complete layout also ...thanks.

 

[Md Ahsanul Haque]

You have to keep 2 things in mind when you see the attach picture. Each line (R,Y,B) consist of 3 core 2.5mm cable (Live, neutral and earthing). The main supply (bottom left) is coming from 4 cb, each 20A 2 pole. I also use 1 ELCB of 40A rating and one 3 pole mcb of 32A rating. other than internal ballast for each lights there are no more circuits within the system. Previously I was using higher rating cb, but due to this burn out light problem I reduce the cb rating to the current one mentioned above. Is there any possible solution to stop these lights breaking down in regular intervals?

 

[(*steve*)]

Are the lights that fail predominantly at the end of a long run, or at the beginning?

Are the failures concentrated on those lamps between a particular pair of phases?

Have you operated just one of these lights from a single phase source to verify that it too doesn't fail fairly quickly? (Is it possible that they're poor quality lamps with a short lifetime?)

Reducing the circuit breaker current is just going to result in the breaker tripping. It won't reduce the current.

 

[Bluejets]

Give answer to #12 and you might get some help.

 

[Md Ahsanul Haque]

Steve: I can not get any pattern of the lights burning out. Lights from all the three phases are damaging and in a random pattern. Sometime its not the lights but the ballast unit damage. I had to change a new ballast in order to run the lights. So I cant give any conclusive answer to your question. I am using Kingston metal halide bulbs for this area.
Bluejets: I thought I already answer your question. The cable layout is attached to my last reply and the cables are 3 core 2.5mm size. I already mentioned the db panel layout in my reply.

The distance for the 1st light (others looped from the first light) from the db panel is around 40m to 60m.
I dont know if this information helps or not but these lights are used to light up the AIS (Air insulated switchgear) area of my company. The total area is around 100m X 200m approximately. And some of the lights are closer to 275KV CB for the transformers. This place has a very high humidity in the air.

 

[Bluejets]

It's still difficult to follow your layout from what you describe which is why I asked for a layout diagram which I haven't seen yet.
However, if you are running 40 to 60 metres to you first fitting as you say, and then have at least 12 fittings per circuit, and then looping out a further distance, all on 2.5sq mm cable then of course you will get problems.
This is not even considering the circuit with 17 fittings.
The voltage drop over this route length with this loading would be enormous.
I could work it out but i would have to know the exact layout.
Just at a wild guess I'd say you would need circuit cabling somewhere around 16 to 25 sq mm ...not 2.5 sq mm.

 

[(*steve*)]

I would try operating a single (or a small number of) unit in a similar humidity environment to see if there is something about the bulbs. Comparing this against a dry environment would be useful.

Have you contacted the manufacturer? A good manufacturer would be concerned and/or be able to offer advice.