Basics resistor/voltage question

[voldemort]

I'm trying to work on connecting a few of my lights on my layout.   

Do I need to run separate wiring for the 5v, 12v 15v items on the system with a transformer for all (or a buck converter).

OR if I run a circuit of say 14v and run in series with a resistor, will that decrease the voltage down so I can use a 3.6v LED on it without a converter?

I've never really understood resistors.  I would have thought they used up power and got hot.  I don't know if they decrease current or voltage primarily. 

Thanks for the help

[woodone]

You might want to look into some voltage regulators. They are made in different voltages. Also different amp outputs.

[mmagliaro]

When you are powering LEDs, you must resist the temptation to think about "decreasing voltage".  The whole idea is that with an LED, you need to limit the current through it. 

If you have a 14v supply, and all you intend to run off it are LEDs (as opposed to mini incandescent lamps), there is no reason at all that you can't just put a resistor in series with each LED.  In fact, that is how you should do it.  I would not try to use a voltage regulator to make a 3.6v supply, or put a big resistor off the 14v supply to limit the current to all the LEDs.  In the first case, say you make a 3v supply.  That should theoretically not cause too much current to pass through an LED that is rated for 3.6v maximum forward voltage.  But if you ever connect a different type of LED in there (that has a lower forward voltage, like 1.4, or 2, etc), you'll still blow it up, and it's easy to forget.  Plus, not all LEDs have the same forward voltage spec, nor do they all light up equally at the same forward voltage.  Much better to just put a resistor in series with each LED.

In the second case, (putting one big resistor off the 14v supply and connecting all the LEDs to that), you are now limiting the TOTAL current to ALL the LEDs thorugh one resistor.  It would probably have to be a big power resistor that will make a lot of heat (like burn-your-finger heat), and any time you add or remove an LED, the current draw will change and all the LEDs will go brighter or dimmer.

The calculation for a single resistor works like this.
3.6v forward voltage drop means that from a 14v supply, if you put a resistor in series with the LED, the resistor will drop 10.4v.
You need to limit the CURRENT through the LED.  Typically, 10ma is a good number for individual LEDs with leads that you might be putting in for building lighting (as opposed to tiny surface mount ones you might use for a loco headlight).
V = I x R   (voltage = current x resistance)... Ohm's Law
10.4 = .010 x R   (.010 is 10 mA)
So R = 1040.
Therefore, a 1000 ohm or 1200 ohm resistor should do the trick.  You would put one of these in series with each LED.
And if you use another type of LED with a different forward voltage spec, you would calculate a resistor for that one
just as I did here.  I would tend to always start with a slightly higher resistance value.  No need to push an LED to its limits, which will shorten its life and not really give you much more brightness.

[voldemort]

that is exactly what I needed to know. Well explained

Thank you!

[mmagliaro]

that is exactly what I needed to know. Well explained

Thank you!

You're welcome!  BTW, I just want to make absolutely sure I was clear that you will be putting a separate resistor on
each LED.  So if you have 50 LEDs spread throughout a bunch of buildings, each one gets its own resistor in-line with one of its leads (it doesn't matter which).  This may seem somewhat tedious, but it's really not bad to just solder a 1/4 watt or 1/8 watt conventional size resistor in-line with one of the leads on each one.  The resistors are cheap.

I should add that the real reason I have agita about using just a voltage regulator to strictly the control voltage below the LED max rated forward value (i.e. we could set up a regulator with a 3v output to drive 3.6v LEDs) is that if anything ever goes wrong with that regulator and it exceeds 3.6, all your LEDs will be toast in an instant.  I feel much safer having current-limiting resistors in there.

[peteski]

While not critical for the resistor calculation, the Vf of white LEDs is rarely more than about 3.2V (especially when we used them at very low currents like 5-20mA). It often is slightly under 3V. I would just use 3V as a standard Vf value for white LEDs.

Max's explanation about worrying about the current is right on target.  Most modelers seem to assume that LEDs are "voltage devices" like light bulbs, but they are really "current devices" which just happen to develop the forward voltage across their terminals when current flows through them.

[videobruce]

Not to get OT, but I have (or had) 3 strings of LED's under my table and one string almost completely died (which was removed). The 2nd string most of the LED's at the powered end are dead (groups of three). The rest of the string is ok for the most part. The 3rd string almost all are ok but that is a little further away from the power source which is a measured 12vdc from a switching supply (not a old school 'brick') that the strings were 'spec' for..

Is it the LED"s themselves that are dying or is it the resistor in each group of three that is the problem?  This happened over a period of time (a few years), 99% of the time the LED's were off!

[peteski]

It is likely the LEDs.  Resistors very seldom go bad.  If you have an ohmmeter (multimeter), you could check the resistors in-circuit.  But I could be wrong.  The resistor vaue is stamped on them.  Something like 241 or 331 for example.

[nickelplate759]

Interesting that they are failing in the pre-defied groups of three.  Are the three LEDs in series?

[peteski]

Interesting that they are failing in the pre-defied groups of three.  Are the three LEDs in series?

Depends on the strip.  One of those strips has a single resistor visible per segment of 3 LEDs.  Those are in series and likely run off 12V.

The other strip has a segment that has 3 LEDs and also 3 resistors.  in that strip each LED has its own resistor. In my experience those are designed for running off 5V, but I suppose there might be a 12V version too (with higher value resistors).

It seems very unusual (especially on the strip where each LED has its own resistor) for 3 in a segment to fail the same way. It as if there was a  some odd problem with the copper traces on that segment to affect all 3 independent LED/resistor combos.

Not so unusual for the 3 series-connected LEDs to fail affecting all 3 LEDs. There seem to still be some very small amount of current flowing through the failed segment of both strips since the LED dies still emits some light.

EDIT:  actually it appears that the strip with individual resistors per LED uses dual-die LEDs.  There are 2 light dots visible in each LED housing.  Those are internally connected in series, so they would need around 6V to light up.  With individual resistors they are likely running on 12V.

EDIT:  These LED strips are mass produced and often not the best quality.  I have seen cold solder joints on them, ,but that usually causes a open circuit, not a faint glow.