LED book...

[rrinker]

 No, when someone talks about a diode and Vr without posting a diagram of the circuit under test, I assume reverse voltage of the diode, not voltage across a resistor.   Hey, I was up til 1:30am on a conference call and I'm working on about 3 hours of sleep.

 I probably would have just measured the drop across the 1K, and also measured the actual resistence of the 1K especially if it was a standard 10% one.

I pretty much dount anyone would make an LED with the flags reversed. Mainly for the heat, like I said. We tend to think of LEDs as not getting hot, but at the junction they can be quite warm. But it's such a miniscule spot of heat, it's easily carried away by the substrate and large metal lead.

 What I haven't seen is a 2 lead bicolor LED, all the ones I've ever had were opaque so you couldn;t see the leads once they get inside the case.Is it two large flags, each with a diode, and a whisker from each flag to the opposite diode junction? I'd suspect this is more likely than there being the usual large flag at the bottom with one diode, and the other one upside-down on the small flag

 Another reason the actual dide junction wouldn;t be inside the case upside down - the light would all try to get out the lead side, not the top. The light comes from the diode junction, and you'd want that to point up and out, not down towards whatever board the LED was installed in.

                     --Randy

[chutton01]

rrinker

One lead longer is common but by no means universal. ANother common indication is a flat in the case. The problem of course with lead length is what if you are reusing an LED that you previous  wired and clipped all the wires nice and neat? If you can see through the LED by holding it up to the light, this is the most foolproof method

I absolutely agree the "longer leg" method is not foolproof, and I was just warning the OP about that. I think the flat spot convention on the LED base is common, but I am not sure if it for all intents and purposes universal.

I have yet to see an LED manufactured 'upside down' internally, probbaly because the small leg doesn;t have enough mass to heat sink the actual diode, plus trying to make a wire bonder that could do that would be an interesting task.

If this refers to whether my question of manufacturers making the LED internal structure different such that the larger "flag" is not the Cathode (must be the Anode then) for whatever reason, then that was kind of my question - how universal is that.

If you measured across the 10 ohm resistor, all you got was the voltage dropped by the 10 ohm resistor, not the LED.

I did, and that's exactly what I wanted as the 10Ω resistor was in series with the LED and the 1KΩ resistor (and yes, I could have measured across the 1K if my test setup was different), so the current thru that resistor was the same as the current thru the LED - just what I was looking for.  I also ramped it up to 8 LEDs in parallel, and got about the same value.

And with a pure DC source like a battery, where are you getting a Vr value from? There is 0 volts across the LED backwards if it's running off a battery.

Correct, and since I plan to use DC power for these lights, I am not certain why I would need to worry about this theoretical "back voltage" you speak of.
Are you conflating NPN/PNP transistor operation in amplification/oscillating circuits with this fairly simple DC diode circuit? Back emf?

 

 

 

[rrinker]

One lead longer is common but by no means universal. ANother common indication is a flat in the case. The problem of course with lead length is what if you are reusing an LED that you previous  wired and clipped all the wires nice and neat? If you can see through the LED by holding it up to the light, this is the most foolproof method, I have yet to see an LED manufactured 'upside down' internally, probbaly because the small leg doesn;t have enough mass to heat sink the actual diode, plus trying to make a wire bonder that could do that would be an interesting task.

 If you measured across the 10 ohm resistor, all you got was the voltage dropped by the 10 ohm resistor, not the LED. And with a pure DC source like a battery, where are you getting a Vr value from? There is 0 volts across the LED backwards if it's running off a battery.

 

                    --Randy

[chutton01]

BroadwayLion

The Negative pole has the BIG flag (inside the LED) and the short leg outside)

The Positive pole has the LITTLE flag and the long leg.

LION always looks at the flags rather than the legs. BIG = Negative. I think I have that written on the wall in the train room, but not in the computer room.

So far it does seem that for Christmas LEDs, the Cathode (Negative for LEDs) is indeed the large "flag" or "triangle" - this image may help.
What I'm wondering is how "universal" is this across discrete non-SMD LEDs in general? Reason being I clip the legs to be the same short length (so that identifying feature is gone) for size reasons, and I don't recall if the LEDs I scavanged even had that little flat side on the base.

[BroadwayLion]

Mark R.

 

 

 

BroadwayLion

[snip] ....The LONG lead goes to the NEGATIVE side of the circuit, (I use GROUND for this), the shorter lead goes to the + dc voltage.

Watch out for LIONS.

ROAR

 

 

 

 

The long lead is positive, not negative ....

Mark.

 

 

 

Correct. LIONS are dyslexic and ALWAYS mix up any binary problem.

The Negative pole has the BIG flag (inside the LED) and the short leg outside)

The Positive pole has the LITTLE flag and the long leg.

LION always looks at the flags rather than the legs. BIG = Negative. I think I have that written on the wall in the train room, but not in the computer room.

ROAR

[chutton01]

CSX Robert

 

rrinker

Now you are too low. You aren't calculating or measuring things correctly. When the LED is operating normally in a circuit with a power supply, resistor, and the LED, it's the forward voltage that matters...

 

 

No, he did measure it correctly (I=V/R, current = volts/resistance), although there was no need to add the extra 10 ohm resistor because he could have just measured the voltage drop across the 1k resistor and calculated it from there..  The 10 ohm resistor actually reduced the current draw slightly, but it would have been neglegible in this case.  In fact, this is the same thing as inserting a multimeter set to current in the circuit - the meter measures the volatge drop across a low value resistance to determine the current draw.

Well, I have a MSEE, so I believe I know what I am doing (I also learned practical electrical wiring from dad, so I can work on house wiring - yes, to code - but then again any model railroader worthy of the title should be able to do that).

BTW, the reason for the 10Ω resistor in the test setup is simply because the setup has a number of parallel LED legs (about 8) so I could try different resistor combos to see what happens in the real world, and I wanted one single point in series for measuring overall current (for this single LED measurement, I disconnected the other wires and had one circuit loop only). This is because I plan to create lots of standard LED+resistor combo, soldered together, so I can put them in structures in a modular fashion using the two brass "rails" suspended across the ceiling method that's fairly common nowadays (I vaguely have plans of using two brass tubing with holes drilled in that the LED/R combo drops in for easy soldering, but I'm sure the real world will prove this impractical or too pricey).

The annoying thing is I am using my good multimeter and got what seems to be a rather low measurement - it wasn't like I was using my Harbor freight dollar special one (actually, that one IIRC was a freebie).

[CSX Robert]

rrinker

Now you are too low. You aren't calculating or measuring things correctly. When the LED is operating normally in a circuit with a power supply, resistor, and the LED, it's the forward voltage that matters...

No, he did measure it correctly (I=V/R, current = volts/resistance), although there was no need to add the extra 10 ohm resistor because he could have just measured the voltage drop across the 1k resistor and calculated it from there..  The 10 ohm resistor actually reduced the current draw slightly, but it would have been neglegible in this case.  In fact, this is the same thing as inserting a multimeter set to current in the circuit - the meter measures the volatge drop across a low value resistance to determine the current draw.

[rrinker]

chutton01

Well, as promised, I checked my scavanged LEDs in the test circuit.
V_R ≈ 28.8mv
R = 10.0Ω
Therefore, current for 1 LED ≈ 28.8/10.0 ≈ 2.9mA, and a thousand LEDs (easy proposition on a large layout) would be 3A, which is not bad at all.

For comparison, I also measured an old school late 1980s Red LED. It pulled only 3.9mA, which I found very surprising, I figured on a magnitude higher.

Circuit was 10Ω resistor (to measure across), 1KΩ resistor, LED, 9V battery - and an old knife switch I found among my electric part box (for on/off control in style) - all components snuggly plugged into my late 1980s Archer^® Experimental Breadboard

 

 Now you are too low. You aren't calculating or measuring things correctly. When the LED is operating normally in a circuit with a power supply, resistor, and the LED, it's the forward voltage that matters. This is going to be around 3.5 volts for a white LED, unless they are really wierd ones.

 Now we need to start witht he absolute basics, before even Ohm's Law. This is Kirchoff's Law. Simplified, what Kirchoff's Laws state is that components in series add voltage. So if you put 12V into the circuit, and the LED 'consumes' 3.5 volts of it, what's left is 'consumed' by the resistor. In this case, 8.5 volts (12 - 3.5). Kirchoff's Laws also state that loads in series share the same current. An LED does not draw a fixed current, but a resistor is a linear device and does. So if we want the LED current to be around 10ma, we also want the resistor current to be around 10ma. Mow we have 2 out of the 3 values for Ohm's Law to come into play - we know the voltage (8.5) across the resistor, and we know the current, 10ma.  Solving for R, we get R=E/I, 8.5/.010, 850 ohms.  Resistors are not available in every single value, so you pick the closest standard value, or larger. If you use a 1K resistor and work backwards, you get the 8.5ma I mentioned previously.

 So hook one up to a 12V power supply with a 1K resistor. Measure the power supply boltage. The use the meter in volts mode and measure the voltage across the LED, it should be somewhere around 3.5 volts. Do the same across the 1K resistor - it should be (power suppy voltage) - (voltage measured across LED). Set the meter to ma mode and insert it in the circuit (anywhere is fine: ps - meter - led - resistor - ps, ps - resistor - meter - led - ps, any combo is all the same per Kirchoff's Law) and it should read 8.5ma or so, depending on the actual ps voltage.

      --Randy

and one of these days I will remember to copy this and paste it on my web page so I can just post the link instead of retyping effectively the same thing every time I explain this.

[Mark R.]

BroadwayLion

[snip] ....The LONG lead goes to the NEGATIVE side of the circuit, (I use GROUND for this), the shorter lead goes to the + dc voltage.

Watch out for LIONS.

ROAR

 

The long lead is positive, not negative ....

Mark.

[chutton01]

Well, as promised, I checked my scavanged LEDs in the test circuit.
V_R ≈ 28.8mv
R = 10.0Ω
Therefore, current for 1 LED ≈ 28.8/10.0 ≈ 2.9mA, and a thousand LEDs (easy proposition on a large layout) would be 3A, which is not bad at all.

For comparison, I also measured an old school late 1980s Red LED. It pulled only 3.9mA, which I found very surprising, I figured on a magnitude higher.

Circuit was 10Ω resistor (to measure across), 1KΩ resistor, LED, 9V battery - and an old knife switch I found among my electric part box (for on/off control in style) - all components snuggly plugged into my late 1980s Archer^® Experimental Breadboard

[richg1998]

CNWman

Hi, I'm new to the site. Can anyone suggest a good book on the use of LEDs in an HO scale layout (mostly related to scenery such as signals, structure lighting, etc.)?

 

Along with all you have been told, I have used 3mm, green/red bipolar LED's for single target signals. Used resistors but operated off of 5 volts DC. I used TTL circuits maybe ten years ago for 14 blocks, plus sidings on a club layout.

 I tried to use the bipolar LED's on AC for yellow but all I got was a small green spot and red spot deep inside the LED. Never took time to see what frequency would give me yellow.

The bipolar LED works very well with stall motor turnout controllers like the Tortoise.

The LED's do work on DCC also. My NCE Power Cab has a red LED with 1k resistor on the DCC circuit coming out of the panel. No prtective diode either.

Rich

[GGOOLER]

here is a good site for a led calculator just plug in the values hit enter and BOOM. there is a circuit schematic with res. values.

http://led.linear1.org/led.wiz

later

g

[BroadwayLion]

CNWman

How is it that they would not work on AC if that was how they would work on a Christmas tree? I don't understand.

 

The Don't. *YOU* are using AC, they are not. There *are* recitiers in and of themselves. There are secrets inside of they mystery warts on the chain. They entire circuit needs to be taken into consideration, not the individual LEDs. Once you take the LEDs out of the set, YOU are now engineering a new circuit. Do it the way that we have suggested, and they will last a good long time.

ROAR

[BroadwayLion]

CNWman

LION you've been a great help to me in answering my nubie questions about LEDs. I bought some LED Christmas light sets at Home Depot just after the Holidays... they were closing them out so I got a good deal. I'm now awash in little LEDs! Time to do something with them. What is the best way to power them... by the AC accessory terminals on one of my power packs, or maybe buy a little DC starter transformer on E-bay and use that strictly for power? Won't the lights last longer on DC power? Thanks- KD

 

Basically, they will only work on DC, If you look carefully at those Christmas Sets they will have a mystery enclosure of some sort, which you cannot open. There may be rectifiers and resistors in there. If you do not see such a device, then there are little itty bitty resistors and or diodes attached to the LEDs directly withing the lamp sockets.

I presume that you are opening the lamp enclosure and removing the individual LEDs.

All you need is a little DC wall-wart, anything between 6 and 18 volts will do. Since you do not know the actual values of the LEDs, just use the trusty 1000 Ohm resistors and see what happens. I use these for 12 - 16 volts.

The LONG lead goes to the NEGATIVE side of the circuit, (I use GROUND for this), the shorter lead goes to the + dc voltage.

Watch out for LIONS.

ROAR

[rrinker]

 Yes, early LEDs weren;t nearly as efficient and needed more current to even light up - if you go below a minimum current level, an LED won;t even light up. Most red LEDs are rated higher than 20-25ma, and to get maximum brightness you need to run near the maximum current. The white LEDs we use for lighting and headlights are generally too bright at full brightness, at least for scale appearance, and so cutting them back well below the maximum is what we do to make them look good. There are high brightness white LEDs rated for 100ma or more. Those sort of things are for specific applications and also usually need to be attached to some sort of heat sink because despite the higher rating, they still just have a hair-fine diode junction at the heart, and it's hard to get lot of heat away from a very small area. If you've ever blown up an LED - sometimes it literally blasts a small chunk of the epoxy casing out. Enough heat was generated in that tiny space in there to vaporize whatever it could, which causes a massive pressure increase, like a popcorn kernal.

                                  --Randy

[rrinker]

 In the Christm,as lights they have LEDs wired back to back, so one lights on one half of the AC and the other lights ont he other half of the AC. An individual LED will light on AC, but it's getting excess reverse voltage half of the time, which exceeds the rating of most LEDs. ANd it's only about half as bright because what it's really doing is blinking on and off 60 times a second. A lot of people can actually see that, especially if there is a fluorescent light as ambient light. It's like using a CRT monitor with the refresh rate set to 60Hz, it flickers annoyingly to most people, moving up to a 72Hz refresh rate gets it out of sync with the office lighting and it no longer appears to flicker.

 DC is the best way to drive LEDs, it's what they were meant for.

                       --Randy

 

[chutton01]

rrinker

 Running a white LED with a 1K resistor on a 12V DC power supply results in about 9ma per LED. 4.25ma with a 2K resistor. Not sure how you got 35ma per, the LEDs are typically only rated for a max of 20-25ma and at 35ma they wouldn't last too long.

12V - 3.5V LED forward voltage (typical white LED) =8.5V 8.5V /1000 ohm resistor = 8.5ma.

Probably going by faulty memory, that's how I got 35mA. I have the LED test setup at my workbench, I'll run it again tonight, using the incredibly complex method of measuring the voltage drop across a resistor in series with the lit LED, and dividing as I = V/R.  If it really is single digit mA , I won't be too surprised.
Oddly, a number of sites on-line state average current draw of 30mA or so, but these might be older sites.  IIRC old school LEDs from the 1980s had somewhat hefty current draws.

[CNWman]

How is it that they would not work on AC if that was how they would work on a Christmas tree? I don't understand.

[rrinker]

 Running a white LED with a 1K resistor on a 12V DC power supply results in about 9ma per LED. 4.25ma with a 2K resistor. Not sure how you got 35ma per, the LEDs are typically only rated for a max of 20-25ma and at 35ma they wouldn't last too long.

12V - 3.5V LED forward voltage (typical white LED) =8.5V 8.5V /1000 ohm resistor = 8.5ma.

                --Randy

[chutton01]

CNWman

LION you've been a great help to me in answering my nubie questions about LEDs. I bought some LED Christmas light sets at Home Depot just after the Holidays... they were closing them out so I got a good deal. I'm now awash in little LEDs! Time to do something with them. What is the best way to power them... by the AC accessory terminals on one of my power packs, or maybe buy a little DC starter transformer on E-bay and use that strictly for power? Won't the lights last longer on DC power?

You have to run the LEDS you salvaged from Christmas Tree lights on DC only - AC will not work. Indeed, I have been tested my reclaimed LEDs with a plain old battery, as I am trying to figure the best combo of resistor/LED (looks like the old standard - 1KΩ - is the best, although they work OK on 2KΩ). LEDs are in parallel, I'm going with the now standard two "rail" (brass pieces) along the structure interior roof, LEDs & resistor attached across the rails, cone pointed down, as needed.  LEDs as stated are voltage directional, I believe the larger triangle within the LED is the Cathode, and this page seems to confirm that.

I'm mostly wondering about current draw, I believe each LED is running 35mA; with 100 LEDs, that's 3.5A, which is not insignificant.

[CNWman]

LION you've been a great help to me in answering my nubie questions about LEDs. I bought some LED Christmas light sets at Home Depot just after the Holidays... they were closing them out so I got a good deal. I'm now awash in little LEDs! Time to do something with them. What is the best way to power them... by the AC accessory terminals on one of my power packs, or maybe buy a little DC starter transformer on E-bay and use that strictly for power? Won't the lights last longer on DC power? Thanks- KD

[BroadwayLion]

The platform is indeed a piece of cardboard. It is resting on a strip of ceiling tile. I uses an aul to drive a hole through the platform and the table top which is also lightweight tile material. The Stirer goes clear thorugh, with the end and the wires coming out the bottom of the table, and connected to the bus wires. I then can adjust the height of  the light fixtures.

ROAR

[rrswede]

I understand the pointed pointless point. Do you use superglue to glue the pointless tops to the LED or something less permanent?  What is the base material the stirrers are attached to? The edge looks like a squashed piece of corrugated paperboard that has been painted.

swede

[BroadwayLion]

rrswede

Now I'll get busy and see how many Christmas LEDs I can realistically use on the layout. I truly appreciate your assistance.

LION has more than 1000, and is still putting them on the layout. Each of about 200 signals has three LEDs, Each of the station platforms is well lighted, some only an inch apart. Station platforms on the Route of the Broadway LION are four feet long. (Subway trains you know.)

These were made with a coffee stirrer (the kind with two chanels in them) an LED, and a thumbtack, painted to match the posts and with the point cut off.

(I found the point to be pointless, you know)

ROAR

[rrswede]

Mike and Roar, just for experimental purposes, I hooked up a 9 VDC wall wort with a 1000 ohm 1/2 watt resistor in series with a Christmas LED, then 2 LEDs hooked together + to - with the resistor and AC power and then built the circuit Mike suggested using AC power, all three worked just fine. Now I'll get busy and see how many Christmas LEDs I can realistically use on the layout. I truly appreciate your assistance.

swede

[rrswede]

Thanks for the response, ROAR. Yes, lighting in passenger cars as well as other cars was a concern on my part. The schematic that Mike had previously posted and included in his response to me substantiated what I suspected was the layout utilized in the after market universal AC/DC LED hookup.

Will be back, and again, thanks. 

swede

[rrswede]

Thanks for the response, Mike. I don't utilize DCC, but based on what I have read, it is sort of modified AC and I believe you are correct, this approach will work. Your diagram and part call outs will get me trying this approach as well as the wall wart one. I really enjoyed going through the entire thread on Night Lights. I don't believe I will get to the level of lighting sophistication on my layout as you and others have displayed but I have bookmarked this thread and will be following it. Who knows, lighting may become as important to me as the trains and their movement around the track are.

swede 

[BroadwayLion]

Yes, LION uses LEDs in passenger cars. Since cars can travel either way through the layout, you need to put a full wave rectifier in the circuit. The AC side of the rectifier connects to the track power, and then you will always have + and = dc inside of your car.

LION is still working on a circuit to keep the lights lit while the train is stopped. Him will report on that when (not if) he is successful.

ROAR

[mlehman]

swede,

I'd think circuits suitable for DCC would work with AC. Many people think they're the same, but they're not. But for something like lighting with LEDs, I'm pretty sure a circuit I used in my passenger cars will work. If you need something less than 12 volts, the spookshow website link  I posted earlier shows a simple adjustable DC output circuit that could piggyback on top of this one. It's about a 3rd of the way down the page and is drawn with markers.

http://cs.trains.com/mrr/f/88/t/213765.aspx?page=2

It's set up to provide power to the 12 volt dc LED strip light segments, but with a voltage regulater can do other voltages.