LED Resistor

[guavatone]

I was wondering what would happen if I wire up 2 LEDs using one resistor by wiring it this way:

+VDC >Resistor> (LED1 and LED2)

both going to ground on the other side

The resistor is Current Limiting and I was hoping to have the same effect as if both LEDs had their own current limiting resistor. Also saves some wiring.
These are just indicators and non-critical running off of 48VDC of a micpre

Thanks-

 

[Samuel Groner]

If the LED's are reasonably matched (i.e. same type and colour), they approximately share current.

I guess you want to switch them independently--note that if you switch off one the other will now pass almost twice the current and thus light brighter.

Samuel

 

[guavatone]

Thank You Samuel. So if I have the +V tied to independent Resistors then current and brightness should match? The LEDs are the same type//model. The LEDs are actually for Illuminated Push-button switches.

 

[bcarso]

At the same current the light intensity will be just a matter of how well the LEDs are matched. Since they have a very low impedance at typical operating currents---i.e., a big change in current for a small change in voltage---running them in parallel as you suggested initially makes the intensities dependent on two variables---impedance and luminous efficacy. Thus running in parallel is a bad idea if you want the tightest equality of intensity. And Samuel is of course correct that if you open-circuit one the other will roughly double in intensity if you are feeding a pair from the same current-limiting R.

If the source voltage is large compared to the LED voltage drop then your resistor behaves more like a constant current source, hence reducing the effect of the variations in forward voltage. Efficiency being of little concern here, that's a favorable situation.

Exercise for the reader: suppose you have no alternative but to send a single wire to serve the needs of n LEDs with anodes bussed together and cathodes switched in any arrangement to common. How would you assure constant brightness for any and all?

 

[guavatone]

[quote author="bcarso"]Thus running in parallel is a bad idea if you want the tightest equality of intensity. [/quote]

hmmm... I can't run them in serries since they need to be independently switched in and out of circuit.

 

[AnalogPackrat]

[quote author="bcarso"]Exercise for the reader: suppose you have no alternative but to send a single wire to serve the needs of n LEDs with anodes bussed together and cathodes switched in any arrangement to common. How would you assure constant brightness for any and all?[/quote]

You could bus the anodes with each LED paired with a resistor. An SPDT switch for each such pair would select the LED or resistor with the switch pole connected to an appropriate current limiting resistor to common. The paired resistor would be chosen to drop the same voltage as the LED when it is on. Cost is one extra resistor.

A P

 

[Svart]

Exercise for the reader: suppose you have no alternative but to send a single wire to serve the needs of n LEDs with anodes bussed together and cathodes switched in any arrangement to common. How would you assure constant brightness for any and all?

no mention of extra resistors was mentioned so my solution would be to stick an R on the cathode/low side for each LED assuming that the source is static.

there are a few more tricks too..

 

[guavatone]

Yeah, I paln to use the 2 resistor method. Both resistors are fed from same +v then to each LED/switch and then 0V.

 

[bcarso]

[quote author="AnalogPackrat"][quote author="bcarso"]Exercise for the reader: suppose you have no alternative but to send a single wire to serve the needs of n LEDs with anodes bussed together and cathodes switched in any arrangement to common. How would you assure constant brightness for any and all?[/quote]

You could bus the anodes with each LED paired with a resistor. An SPDT switch for each such pair would select the LED or resistor with the switch pole connected to an appropriate current limiting resistor to common. The paired resistor would be chosen to drop the same voltage as the LED when it is on. Cost is one extra resistor.

A P[/quote]

Sure. But what if you could only switch the raw cathodes of n LEDs to ground? A contrived situation I admit, proposed more for the puzzle value than any other good reason.

 

[guavatone]

I must admit I'm a bit lost on the idea of replacing cathodes

 

[Svart]

stick a constant current diode on each LED, high or low is another trick.

 

[bcarso]

The answer I was looking for is simple but has some subtle aspects, and depends on parts being well matched and roughly at the same temperature:

Drive everything with a fixed voltage, and derive that regulator reference voltage from a similar LED driven by a current source. Thus you have the right current for any given LED on the face of it, and you have compensation for ambient temperature and the self-heating of the components.

Because of the need that the parts be well-matched, it's not a scheme that I would endorse for production---especially if you ever had an LED fail and had to replace it with one from a different production batch.