Current source OK until load disconnected

OK, reading on, you really do mean "current source".

It sounds like a poor design. I'm not sure when I would ever want to change a battery from a current source, and neither would I switch the load on and off from a current source (I'd switch the current source on and off.

OK, that's kinda reasonable, however unless you have a perfect current source (or at least a good one) it may not tolerate having the output shorted.

And you'll have a problem with the battery because you can't short it before you remove the battery (or you'll short the battery) and yo can't disconnect the battery first or the voltage will spike.

OK, I think this is a bad idea, but I'll give you my advice. I think the correct solution would be a different design.

However... The solution for the battery is to have a SPDT switch (or relay, etc.) that in one position charges the battery, and in the other position disconnects the battery and shorts the current source. "But wait" (I hear you say) what about the brief time while neither the battery nor the short is connected? Here's the trick -- place a high power zener across the current source. This zener must have a voltage higher than seen with the load connected. For the short period of time while the relay switches, it clamps the voltage across the current source. It's only going to be conducting for a matter of milliseconds, so it's OK to overload it, but perhaps not more than 10 to 20 times (e.g. use a zener rated at 5W for a 50W load)

For the LEDs, you can simply use a switch to short them out.

Remember, this advice deals with the problem as you stated it. I think there would be several far better solutions that would come to light if we knew more about the problem.

 

What current are we talking about? And what voltage across the lamp under normal conditions?

A simple solution may be to create a shunt regulator with a voltage a couple of volts higher than the lamp, placing it across the lamp. If you remove the lamp, the shunt regulator takes up the load. It would have to be rated for the full lamp power (and maybe a little more).

 

Why do you want to replace bulbs when the power is on, and without turning the others off?

Assuming they're LEDs, it's not going to be because they've failed.

If you have control over the current source, you should be able to prevent it spiking in voltage. Perhaps you can limit the rate at which its output voltage increases (without limiting the rate at which it decreases)

 

At the moment I'm confused as to which of these is the real problem:

1) Voltage spike from current source as load is removed (causing arcing perhaps?). The phrase "which blows it up" is used. Does this mean, the lamp, the battery, the constant current source, or something else? And how long does this take?

2) Other lights going off as one in a string is removed

Oh, I asked what current an voltage was normally required and I get an answer that the current varies. This means either a variable current source, several current sources, or something else. What else?

The lack of an ability to give us detailed information is making this like pulling teeth.

 

Not sure what these "taps" are, but maybe you can arrange the fittings for the lights to short out when the bulb is removed.

That an re-engineering the constant current drivers so they don't fail if driving an open circuit.

 

Chris, consider this a formal warning!

What is the first rule of "the secret section"?

 

The first rule of "Secret Section" is "Do not mention the Secret Section"

 

I guess the "secret section" has a backdoor. Someone published this on the internet. This could be used in parallel to the LED. If the voltage rises due to the removal of the LED or battery, the crowbar will short circuit the ring.

Only problem: how to turn off the crowbar after a new LED or battery has been inserted?

Anyway: a design that doesn't shut off when the circuit is interrupted (or short circuited) doesn't look sound to me. I fully agree with that part of the discussion.

Harald