CFL startup current spike: welding my relay closed?

My garage door opener has two “standard” light bulb sockets. I don’t recall for sure, but I think each is rated for at least 100 watts. Some years ago I replaced the original incandescent bulbs with a couple of 150-watt-equivalent CFLs. According to the packaging, the actual power consumption for these CFLs is only 43 watts each - so according to the socket rating, there shouldn’t be a problem with heat building.

Power to these sockets is controlled by a relay, in turn operated by the opener’s internal circuitry. You can turn on the lights manually with the switch by the garage’s house door, or the lights come on when the big door is opened (and then they are supposed to shut off a few minutes later by timer).

The interesting thing is that sometimes the relay sticks closed, such that when the timer expires, the lights stay on indefinitely. The only way to kill the lights in this situation is to manually cycle the relay using the switch by the house door. I can hear the relay cycling when I do this, but it usually takes 2-3 dozen cycles before the relay finally opens and the lights go out.

Is this related to the start-up current for the CFLs? How big is that current spike? Basically I’m wondering if a new relay (of the same specifications as the original) will fix this behavior, or if it’s inevitable with these CFLs (unless I install a much larger relay).

CFL bulbs can have a rather large inrush current. I don’t have any exact numbers for you, but something on a rough order of magnitude around 20 to 25 amps for maybe 300 to 500 usec is probably in the right ballpark. The duration is short enough that it won’t overheat your socket, but it’s enough to weld your relay contacts.

Maybe you can use an inrush current limiter to protect the relay contacts?

If I limit the inrush current, will that interfere with the startup of the CFLs?

How would I select an appropriate current limiter for this application?

Might it be easier to just get a honkin’ big relay, one that’s rated for 10-15 amps?

If it is sticking, it is likely worn out. Incandescent bulb inrush is usually estimated at 10X nominal for 100 milliseconds. This is quite brutal in comparison to capacitor inrush of a CFL. BUT, winter starting might do some odd things. CFLs seem to be Muntzed to the last penny.

43 watt CFLs?? so like 200W equivalent each ? That is some light right there.

Inrush current limiters (NTC thermistors) need tens of seconds to cool before the next turn-on. They are sized by nominal current.

As noted by others, what is occurring here is “contact welding.”

It will probably work for a while. And then will fail prematurely.

The three most common materials used for contacts are:

  1. Pure silver.
  2. Silver Cadmium Oxide (AgCdO).
  3. Silver indium tin oxide (AgInSnO) or silver tin oxide (AgSnO).

More info here.

Each has its pros and cons. As for welding, #3 has the best weld resistance, followed by #2. #1 (pure silver) is not a good choice for contact material when welding is a concern.

Look up the specs for the current relay and find out what the contact material is. If they’re pure silver, I would replace it with a relay that has contact material #2 or #3. (And of course, make sure the voltage & current rating for the contacts meet or exceed those of the existing relay.) If the current relay has contact material #2 or #3, replace it with a relay that has a significantly higher current rating for the contacts.

And in case this isn’t obvious, the control voltage for the coil in the replacement relay must be the same as the current relay. It would also be good if the steady-state current draw for the coil is the same or less.

Can someone tell us how a 150W-equivalent (or 200W-equivalent) LED device would compare, with respect to the questions in this thread?

CFLs are already pretty much incredible pieces of junk, a short duty cycle application like a garage door opener seems to be a good way to get a really short lifetime out of them.

LED’s would be ideally suited to this kind of use. They fire up instantly and require no special fitting. Prices are coming down all the time now.

And no ghastly start-up current spike (which seems to be the OP’s problem here)?

LED lamps have a driver circuit which can have a significant start-up current (inrush). It depends on how that driver circuit is designed. Some of them have a fairly mild inrush and others have a much larger inrush, comparable to that of a CFL.

Problem is that nobody (yet) makes a 150-watt-equivalent LED bulb.

This. I have an LED torchiere lamp in my living room that uses six Cree 30-watt LED modules (yes, 30 watts actual power draw), with each one powered by its own Flexblock. All six Flexblocks are supplied by a single 240-watt power supply, and they have an appreciable inrush current at startup. If I turn the lamp on with the dimmer set at a middle or high power, the power supply interprets the large inrush current as a short-circuit and goes into a “hiccup” mode, where it turns off the output for a couple of seconds, and then turns it on again to see if the “short” still exists. The hiccup cycle repeats until I turn the dimmer down to a low power setting so the startup spike is below the power supply’s threshold; after that I can turn it to max brightness without any issue.