LED vs. Transformer-type Pilot Lights 1

[peebee]

I had just about convinced myself that I wanted to specify LED-type pilot lights on all switchboards and electrical equipment as I assumed that LED's would have the longest life, when I came across some not very detailed or current literature that indicated transformer-type iuncandescent pilot lights may outlast and otherwise outperform LED's for the following reasons:

1. Transformers damp spikes & noise from reaching the filament, the transformers essentially run in saturation and the voltage seen by the lamps is very clean. LED's are typically directly connected via a resistor which will pass all spikes to the LED.
2. Low-voltage incandescents have short, fat filaments which have a much higher resonance frequency than most factory vibrations and therefore last a very long time.

Any thoughts on this? Any recommendation for LED, Tx, or other pilot lights? Which will provide the longest life?

 

[dpc]

The transformer-type lamps have a much better life-expectancy than resistor type, but I think this is due to the thicker filament more than anything else. They still burn out with regularity, and, of course, you can't use them for dc control circuits, only ac.

I have done several projects where the LED-type lamps were used. I never heard of any problems with LED failures. Each lamp generally has multiple LEDs, so failure of one LED would not really be a problem. The early LED lamps were just not very bright and this was often a problem in brightly-lit control rooms or outdoors in bright sunlight. New "cluster" type LED lamps are much brigher, but probably still not as bright as an incandescent lamp.

I've never heard of LED lamps failing due to voltage transients, at least not at an unacceptable level.

Seem to see alot of new big trucks with LED tailights.

 

[peebee]

So you think the LED light will last longer than the transformer type?

 

[tmahan]

Peebee, I can't address the issue of longeity directly because I flat don't know but that never kept anyone from posting before... ;>)

I know that at least two of the big European switchgear companies have been using leds for indication. I was on the phone with a service tech who kept telling me the the power off lead (as in dense, heavy metal) was not on glowing. I took a awhile to figure out what he meant by that.....

By the way the LED was defective or burnt out, the starter just didn't have power.

I have used LED exit signs in high vibration areas and they worked very well compared to the other types...

 

[wbd]

I would go with the LED's. I have used them for a number of years and have had almost no problems with burnout. The one time I did, the bulb had just been in service for about a week when it failed, so I atributed that to infant mortality.
wbd

 

[busbar]

I don’t have much experience with transformer-rated based-LED replacements, but with resistor-coupled versions for DC service in MV/HV switchgear.

Be careful—at one time my employer spent quite a bit of money on custom multi-chip red, yellow and green versions for use in nominal-48VDC service in a switchgear lineup. The based-LED manufacturer was very helpful with samples and all, but wanted an excessive incremental cost for a non-polarized version in case there were polarity differences at lamp bases.

Actual 24-cell DC voltage for flooded-cell float service is a bit under 54V and somewhat higher for ~24-48 hours if battery-string equalizing is needed. The integral series-resistor value and continuous watt dissipation was left to the LED-assembly manufacturer. They asked for and were furnished with and insisted they could easily comply with fairly detailed expectations for the lamp-replacement components. In order to try to match the light output of #1835 incandescent lamps, they selected a integral series resistor that ran the leds at somewhat above full current. The vendor insisted they had a superb track record with DC-service indicators in switchgear.

Unfortunately, during a visual inspection about six months later, when one was touched after removal of a diffuser cap, it shorted at the base terminals and tripped a bus breaker serving programmatic but non-critical circuits, causing all sorts of furtive, high-pressure ‘administrative’ jaw flapping.

Thermal deterioration of nonmetallic insulating parts of the based LEDs probably caused internal failure. It was expected that periodic removal of the based LEDs for inspection would be permitted for the replacement-LED assemblies.

OTOH, due to recent operating experience at sister facilities, there was significant reluctance to blindly reenergize the bus without a few hours of closely supervised, carefully documented offline testing, to reasonably assure that the trip was erroneous based-LED failure in the trip circuit and not driven by a true relaying-initiated problem.

 

[buzzp]

Series resistors with LED's do have some capability to limit transients because of the inductive properties of resistors (depends on type) and the current limiting of the resistance. LED's are the way to go over any other indicators I have heard of.

 

[dpc]

Yes, I think LED units should last much longer than transformer type indicating lights.

 

[jraef]

I agree with dpc. I have never heard of LEDs failing due to voltage transients within the "typical" range for low voltage controls. I also am suspect of the statement about filament thickness and industrial vibrations etc. For my money, hanging that wieght (the transformer) off of the back side of the pilot light has caused me no end of trouble in high vibration applications like rock crusher controls. LED pilot lights have very low mass and so do not "jiggle" as much. I switched to LEDs and my problems went away.
I have a friend working for a city trafic control division. They have been changing out traffic signal lights from transformer type LV systems to LED clusters. They did it for the energy savings, but he reports that his job has become incrementally easier since he no longer replaces light bulbs.
Go for the LEDs. Subvert the dominant paradigm... Think first, then act!

 

[peebee]

Many thanks to all.

 

[advidana]

Try old technology- neon lamps, they are not bright but I have seen them last for years and years without failing.

 

[ScottyUK]

Busbar's comment regarding innocuous faults like panel lamp holders momentarily shorting is worth emphasising. We experienced this on a generator AVR when the glass envelope of a filament lamp broke when being replaced, causing the filament supports to short together. The short blew a control fuse which prevented opening of the main field contactor, leaving the machine over-excited but off the bars. We had a great time proving that the resultant over-voltage did not damage the generator, and that over-fluxing of the generator transformer had not caused any problems.

Moral of the story: sub-fuse lamps if losing control power causes a dangerous condition to exist. It is cheap to do, and might save a very inconvenient situation arising. Designers - it is much easier for you to fit them from the outset than it is for us to retrofit later!



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

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