Best way to wire multiple LEDs to single function output?

[RBrodzinsky]

What would be the best way to wire two SMD LEDs to the same function output of a decoder?  Serial w/ 1 resistor; parallel w/ 1 resistor; parallel w/ individual resistors?

I am installing a LokSound into a Kato N-scale Amtrak P42 Genesis. There is a way to mount the red class lights, on both ends, so would like to put them onto function outputs.  But, since there is no light pipe for them, each end of the loco will require 2 0402 red LEDs

[crusader27529]

Parrallel with one resistor is not a good idea. Individual resistors with a parrallel connection is fine, but requires more current.

As long as the LED type is the same (color, specifically), a series connection with one resistor is my preferred way to connect multiple LEDs. Even if the color used is different, multiple LED with one resistor is fine, as long as the blue output is a high enough voltage. Some decoders supply approximate track power levels, but some supply only 5V.

Do a Google search, but I use this site: http://ledcalc.com

I've connected a dual headlight and two red marked lights to a single function output without any problems.

[peteski]

The serial connection of LEDs and a resistor would be the simplest and lowest component count. But with 3V dropped by each LED I would limit the series connection to two or possibly 3 LEDs and a resistor. Mind you though that if you also run this loco in DC then the LEDs will stay dark until the track voltage goes over the total voltage dropped by each LED plus couple of volts. And id one of the LEDs fails (opens) in the string, they will all go out. Not that this happens very often. This option also consumes the least amount of current.  Thsi assumes that the "blue" wire is the full rectified decoder voltage (around 12V) in DCC.

The most versatile and reliable would be to  have the LEDs hooked up in parallel, each with its own series current limiting resistor.  In this setup the LED would light up on much lower throttle setting (track voltage) in DC operation. And LED defect would not affect others.

But LEDs are very reliable so I would not worry about  LED failures.

[RBrodzinsky]

Thanks, I was leaning serial, and think I will go that way.  I think ESU decoders put out ~12V rectified in N (will double check)

[peteski]

Just to make your life easier, here is an example:

Lets chose sample values.
Vblue (decoder's blue wire voltage) = 12V
Vf (white LED forward voltage, approximate) = 3V
If (forward current chosen for the LEDs - the brightness)
R = current limiting resistor

For example, if you want to pass 5mA (0.005A) through 3 series-connected LEDs:

R=(Vblue-Vf-Vf-Vf)/If  which is (12-3-3-3)/0.005=600 ohm
That  is not a standard resistor value. We use the most common 5% tolerance resistors and here are the standard values. Standard values would be 560 ohms or 620 ohms. So the If will be a bit higher or lower than 5mA.

If you are curious, rework the equation to find the current with a known resistor value: If=(Vblue-Vf-Vf-Vf)/R.

If you want to only have 2 series-connected LEDs with 5mA passing through them then:
R=(Vblue-Vf-Vf)/If  which is (12-3-3)/0.005=1200 ohm (or 1.2k ohm). That is one of the standard resistor values.

[Big Windy]

You could also use a CL2 constant current led driver in an SMD Package. Constant 20 ma in a series circuit. Red LEDs usually only have a 2 volt forward voltage/drop. As long as you connect it to the full rectified voltage of 12 volts. If the function is led ready you will have to go parallel.

[peteski]

You could also use a CL2 constant current led driver in an SMD Package. Constant 20 ma in a series circuit. Red LEDs usually only have a 2 volt forward voltage/drop. As long as you connect it to the full rectified voltage of 12 volts. If the function is led ready you will have to go parallel.

Theoretically any parallel LED connection without resistor in each LED branch can result in unbalanced brightness.  If all the LEDs are identical part binned from he same production batch then this method will work ok.  And yes, red LEDs only need about 2V to turn on but white LEDs (headlights) need around 3V each.

As mentioned by Big Windy a constant current device is also a valid option, but in this case it is uneconomical. Where a resistor costs pennies (or even fraction of pennies), the constant current device is much more expensive (and usually larger physical size).  The constant current device works quite well (although in addition of the voltage needed to turn the LEDs on it needs few volts itself).  Best use for this type of circuit is in DC locos, since it will keep the LED in constant brightness through the entire throttle range (voltage range). But because it needs few volts to turn on, the headlight will not come on until the loco is already running at some speed.

I stated that using the constant current device for DCC functions is wasteful since the voltage inside the decoder is around 12V and pretty constant throughout the model's operating range.  A simple resistor will work just as well (and is much cheaper).

[craigolio1]

I've run LEDs in parallel many times, with only one resistor and havn't had problems yet.

I think the issue is the if you use smallest possible current limiting resistor, and one LED fails, not the remaining LED receives to much and out comes the magic smoke?

I find full brightness LEDs, especially red markers, to be far too bright so I use resistors many ohms larger than needed. If an LED fails the other still won't receive even its own full rated current.

Craig

[peteski]

I've run LEDs in parallel many times, with only one resistor and havn't had problems yet.

I think the issue is the if you use smallest possible current limiting resistor, and one LED fails, not the remaining LED receives to much and out comes the magic smoke?

I find full brightness LEDs, especially red markers, to be far too bright so I use resistors many ohms larger than needed. If an LED fails the other still won't receive even its own full rated current.

Craig

Yes, you can do it, and it is often done - many battery-operated miniture LED Christmas lights have all of the LEDs wired in parallel, and even industrial or household LED light sources (like light bulbs) have pairs of parallel connected LEDs in the housing.  Like I mentioned earlier, if the LEDs come from the same production batch and from the same bin series then this does work well. But as you stated, if one of the LED fails, the current which was passing through the failed LED will now pass through the remaining LEDs. And as you said, depending on the design parameters, it might create over-current condition for the remaining LEDs.

Still, that is not the best practice. Do to their electrical characteristics, it is best to have controlled (limited) current for each LED. Of course series connection by design controls the current to all the LEDs in the chain.