Electronic Ballast Failure

[CBravo]

thread248-278233
We are still experiencing ballast failures even after installing 2 pole motion sensors on our 4 lamp T8 fixtures.

I'm thinking that because each ballast has two hots and no grounded supply conductors and the 2 pole motion sensor opens up both L1 and L2, the energy in the lamps has no place to disipate other than to circulate in the ballast internals.

So I'm thinking an MOV would be in order. So my question is;

2 MOVs per fixture, L1 to ground and L2 to ground.
or
1 MOV across L1 to L2.

Your thoughts please.

 

[itsmoked]

It doesn't appear you're doing anything obviously wrong. Do you have any VFDs running in the building area with the failing ballasts?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[CBravo]

No VFDs in the immediate vicinity. Several VFDs around 100 meters away.
We added a three phase transient voltage surge suppressor to the lighting panel as well and it seemed to reduce the failure rate, but didn't halt the failures.

 

[DRWeig]

The TVSS at the panel should be catching anything traveling from upstream. Are there circuits from the lighting panel that go outdoors? To the roof? Really close to the exterior structure of the building at the top? Is the lighting panel a great distance from the service entrance so that the equipment grounding conductor has some significant impedance out there?

I have experienced loss of a client's ballasts in these cases, in which lightning (even just nearby) was probably the culprit. Your MOV suggestion at the fixtures is worth a try. The possibilty of energy traveling from the ends of the circuits back toward the panel is real, so that the TVSS at the panel can't deal with them until they arrive there.

If you at least partially clamp the voltage between the phases and between each phase and ground at each fixture, it might save them.

Good on ya,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[DRWeig]

Oh, and after reading through your other post -- did you happen to get these ballasts and lamps for a really low price? There are quite a few el cheapo ballasts out there, and lamps as well, and some of them live up to expectations. Some of them do OK, though.

Your motion sensor switches cause no problems. They're just like light switches, and they have built-in time delays so they don't short-cycle. The fluorescent lamps do not store energy, so there is none to release when they are extinguished. They appear as a resistance to the ballast, albeit one that decreases after starting. Any damaging energy must be coming from the incoming power (or equipment ground, if it has some impedance due to distance from the service entrance).

Good on ya,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[CBravo]

It's a steel clad structure and there are no circuits that exit the building.The lighting conduit is run across the roof joists of course. But the area is at the farthest point from the main service. Around 500 feet or so and the 480 VAC feeders are run across the roof for about half that distance.

How would I go about measuring the ground impedance?

 

[DRWeig]

Ground impedance will still be low at that distance if everything is sized right and there aren't any loose connections along the way.

You mentioned that these ballasts are on a 208/120 service. Is there a chance that some device is switching off the primary of the transformer on a routing basis? Or is this addition subject to frequent power outages that don't happen on the origninal building? Interrupting the primary of a transformer can send a good bit of energy down the secondary conductors... Otherwise, I'm stumped. I would go from there to a power monitor if it's worth the effort (or just try the 3-MOV experiment).

I enjoy chasing transients. It can be frustrating, though.

Good on ya,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[Boof87]

Had a similar problem: a 10 storey office building with electronic ballasts in all the fluro's. Ballasts were blowing constantly (and randomly), we put surge suppressors on as you have and got the same results, a reduction but problem was still there.

After countless hours chasing transients etc. Starting dealing with the manufacturer trying to get information, found out only so many of these fittings could be installed on one mains Tx. Why? each ballasts produced a certain amount of a certain harmonic, the cumulative total was causing the ballasts to blow.

May be something similar? Maybe start asking questions of the manufacturer.

 

[itsmoked]

I can easily believe that. It's also a good fit for this case history.


CBravo; Get someone in to instrument and assess the harmonics on your distribution. It shouldn't take long as you have a static situation.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

Have a real good loook at the mains zero crossing: if you have bad harmonic problems then a multiple zero crossing can exist. If your ballast has an active PFC front end it tries to follow this waveform and the magic smoke escapes. This exact problem cost my former colleagues in IT a couple of big rackmount UPS units.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[DRWeig]

The harmonic issue does sound really plausible. Never would have thought of that! It's why I hang around eng-tips.

Let us know what you find.

Good on ya,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[OperaHouse]

Had a customer with a similar problem. He had about 100 photo eyes running on the same transformer. The internal power supply for these was phase triggered off the line so they all drew power at about the same time. I sold him a bunch of RC networks (.47uF 22 ohm) that he placed at each module and it solved the problem.