My October MR has a two page article on selecting resistor values for LED’s and 1.5 volt incandecent bulbs. Not a bad article, but the author did not mention the magic number for biasing LED’s. Namely 20 mA is the “do-not-exceed” current for any kind of LED. Anythng over 20 mA will fry the LED. 10 mA will make any LED glow plenty bright enough for model railroading purposes. Don’t bias a LED at 20 mA, leave yourself a safety margin.
The 20 mA limit is a heat limit. As current increases, the LED chip gets hotter. It has to shed the heat thru the plastic case/lens to the outside air. The amount of heat the case can handle is dependent upon the size, bigger cases have more surface area to exchange heat with the outside air. So the 20 mA limit is good on all brands of LEDs.
Which makes the selection of current limiting resistors more straight forward. To a first approximation (neglecting the LED forward voltge drop) your resistor is the drive voltage / 10. That comes out in Kohms. For example, with a 12 volt drive voltage, you need a 1.2K ohn resistor. Ohm’s law is supposed to work with volts, amps, and ohms. Old shortcut, if the current in in milliamps, the formula still works but the resistance comes out in K ohms.
If you want to be more precise, you can compensate for the forward voltage drop of the LED. If you think about it the resistor will have the drive voltage on one end and the LED forward voltage on the other end. So the resistor sees drive voltage MINUS LED voltage. So assume a LED forward voltage of 2 volts (very typical) then the voltage across the resistor is 12-2 = 10 volts. Divide by 10 mA and obtain 1 Kohm, not too different from the 1.2 Kohm obtained in the first approximation.
I only bother with the L ED forward voltage when it is large compared to th
Well, yes and no. There indeed are LEDs rated for a LOT more than 20ma. The ones I have for headlight replacements say 25ma, but with a 1K resistor on typical HO DCC voltages, they won;t exceed 10ma.
2V might be common for red, green, yellow, etc LEDs, but white LEDs are mostly 3.5V. Always read the specs if you want to compute that closely. If you stay well clear of the absolute limites, a little more or less current either way won;t hurt anything, but if you intend to drive the LED near its limits, you need to be much more careful in making sure a slight surge in power or something won;t cause the current to exceed the limits. However, for model railroad use, it’s pretty pointless to drive an LED used for a headlight at anywhere near the limit, they are generally FAR too bright.
Ah yes. There are exceptions to every rule. I decided not to qualify things with “typical”, “usual”, “most” , “generally” on the theory that readers understand that there are always exceptions.
Checking specifications is always a good idea, if you can find the specs. Lot of us are working with used LEDs scrounged out of broken equipment, or bought surplus and we lack even a maker’s mark, let alone a part number. With a part number you can google for the spec sheet, but lacking same, I have no idea how to find a spec.
But I can get no-spec LEDs to light up if I just bias them for 10 mA forward current (If) and hook em up right way round.
Randy already covered it, but to sum up, you method works as a means of protecting the LED, but may not get everything it can out of some LEDs that perfectly happy drawing considerably more than 20ma. Usually these specs don’t matter if your only concern is protection.
I use different LEDs inside structures to vary the light output for good effect. They run on a 3 volt circuit, so resistor values tend to go no higher than the double digits before things get too dim. If you stick in one with a higher value, say 50 ma draw, it can really confuse things for me. I build parallel/series circuits for the LEDs in large structures. Having one mixed in and setting it up like the others can then lead to a dim or no light situation, even though everything is hooked up and the resistance is like 22 ohms. There may also be a “master resistor” for the entire structure, then branch resistance is adjusted further with another resistor. The branch may not have enough current to drive the LEDs on it because of this sort of specialized situation. It’s only happened to me once, but had me scratching me head at first trying to figure it out as all the LEDs checked good when tested individually. You could run into it when lighting a passenger car or loco extensively, too, I suppose.
Generally, hooking up a LED in reverse won’t hurt it UNTIL you exceed a certain value. I think this is called reverse current, but it’s in the specs with any LED.
I did the below a few years ago. Been working with LED’s since 1972. On average, around 12 .5 volts from many decoders.
I have never had to calculate resistance as LED’s are current operated, not voltage operated. If not bright enough with a bunch in series, then to many for the voltage being used. I use an amp meter in series with the LED’s.
LED Resistor Values
Measured with 12.2 VDC supply using 20 ma max current LED’s. Values will vary a ittle depending on actual DC supply value.
Resistor Current
1000 ohm 9.0ma
750 ohm 12.45ma
680 ohm 13.12ma
510 ohm 16.25ma
I never go over 75 percent for any electronic devices.
By the way, there is a DCC and electronics forum here.
That’s why I just throw a 1K resistor on and call it a day. And tell other the same. I haven’t bothered with any of the superbrite LEDs yet, just ordinary warm white types for first gen diesels. I only had one where I needed to go bigger than 1K because it was too bright - a Stewart loco which had light pipes which must be much more optically clear than the typical plastic lens or light pipe. What I ended up doing there was adding the 1K in series with the stock 470 ohm (meant for DC operation) rather than replacing the 470 with the 1K.
The resulting forward currents for the various resistors are assuming a 3.5V forward voltage in the LED. Again, normal for most standard white LED. When I recommend 1K, I also usually mention the specific LEDs I use - I’ve bene using Miniatronics Yelo-Glow, in both 3mm and 5mm sizes. The color is nice and the way they are made with the filter just over the die inside the water clear case makes them look sort of like a light bulb screwed in the middle of the headlight. Yes, those 1000 for $2 ones from China on ebay are MUCH less expensive, I guess it’s no big loss if the color ends up not being right, so eventually I may try some. The Miniatronics are kind of expensive, but the color is right and Miniatronics items are some of the thing that the LHS actually stocks, so if I REALLY need some I can go get them.
I have been using the minitronics Tower LEDs which fit right into the headlight openings of most GP engines.
I have to ream out the light openings slightly and the the Tower LEDs fit right in - I super glue 2 of the Tower LEDs one on top of the other and they will slide right into the backside of the headlights.