Wire... How small is too small for LED's?

It looks you bought it, so give it a try.   Hook one up for a test and let set under power for a few hours.

LEDS draw very little current, so the chances are this will work as long as the run is not too long.   

You can't see the wire that would be too small for one LED!

For #34 wire, two conductors one foot long each supplying the LED, the voltage drop would be .1 volts.  That's wire 0.00630 in diameter, about the size of your hair!

Thanks guys, I'll go try it out and see how it goes. What value resistor do I need for this? I'll have to get a bunch.

What is the supply voltage? If the LEDs are meant to operate with 20 ma, then if the supply is say 18 vac, the resistor should be about 750 ohms in series with each LED. If many of these LEDs are in series, then I would use the GRJ Lighting Regulator from Hennings. You could use one regulator to supply each LED, but that gets costly. The benefit however is that you can dial in the brightness you desire. Alternatively, if each LED is a single and you want to use a simple resistor with out the regulation benefit, you could use say a 500 ohm resistor and a 500 ohm trim pot in series with each LED and dial in the brightness to some extent. This could all be laid out on a piece of perf board.

The resistance in series would have to dissipate around 0.3 watts so an 1.8 watt resistor and a tiny trim pot would be fine.

Somebody check my work, I'm watching the political news while I'm writing this .

If you're running on 18 volt track power, the LED is actually seeing half of that, remember it's a diode.  Of course, I like to add a protection diode running on AC.

I'd actually recommend you use at least a 1k resistor for signal LED uses.  It drops the current to the LED, and they're still plenty bright.  You can also use smaller resistors, for 1k with 18V AC or 12V DC, all you need is 1/4W resistors.

Chuck, what's a 1.8W resistor?

See I figured that would happen. I think that's when the political candidates were talking! 1/8 watt (next key over to the right of the dot...

Chuck, if the resistor has to dissipate .3 watts, an 1/8W resistor would be a tad light in any case, I think running it at twice it's rating makes it a smoke resistor!

I was completely off on the whole post. The answer would be 12 v times 20 ma or .24 watts. Still too much for 1/8 watt. Thanks for all that. And if using 18 vac then a diode in series with the source?

Actually, you have to take into account the voltage drop across the LED as well, so the power would be a little less.  However, I like to run stuff like markers and signals at lower current levels anyway, so I normally get away with 1/4W resistors with plenty of margin in power handling.

Good morning guys. Last night I did manage to get some leads of the tiny wire soldered to an LED. That will get easier with practice. By the time this project is done, I should be quite proficient.

Let's talk operating voltage. Sorry Chuck, but you have made an assumption that the signal system will be associated track voltage, not the case. The signals will be driven by C/MRI which will provide the ground side of the circuit. Operating voltage can be anything I want, I'm leaning toward 5V. I'll have to see how that looks. I seem to have a large stockpile of 120Ω resistors leftover from enterTRAINment. I'm thinking that they were intended for use with an LED occupancy display we had planned, but never built.

So, is that value appropriate for that voltage? This math stuff is where my electronics skills break down.

I typically use 220 Ohm resistors for green and red LEDs off of 5VDC, though you may want to use a slightly lower value on the green than red for equal brightness.  The slight difference doesn't really bother me.  

Another option that may be easier if you have a large number of LEDs to hook up, which I bet you do, Eliot, would be to use the C/MRI to switch the ground side of the LEDs, and use a  constant voltage supply on the high end, eliminating the need for a resistor on every LED.  While a constant current supply is better for LEDs in general, going the voltage route would allow for any number of LEDs to be connected to the supply.  (Well really more like 50 LEDs from a 1 amp supply)

JGL

JGL, remember these LED's are single package, bi-color, so it's a polarity flip does the red and green. I have ordered 18 of these cards from Chubb. It's his Searchlight Signal Driver card.

There 12 separate circuits on this card, that drive12 individual, 2 lead, bi-colors. Each circuit uses 2 output bits and is able to provide an adjustable AC signal to create yellow. There's a trim pot in the circuit that allows you to favor the red for a more amber look, probably something like 60/40. It is possible that the resistors are built into the board.

The large pads on one side will take their signal from the C/MRI output bits, the ones on the other side will go out to the LED's. I'll know more when I get them in hand and start assembling them.

There are already massive 5V power supplies for the C/MRI on each of the 3 motherboard panels. I could either draw off them, or I have more and could set up a dedicated one. I will need to power other 5V circuits as well.

A link would be helpful Elliot.

I assume this is what you are looking for John.

Chubb's SSD card

Edit: I just dug a little deeper and found this, FULL DETAILS

Just curious Elliot. How are you stripping this wire? I suspect its not easy without precision strippers. If it was magnet wire you would only need a match.

Pete

I'm still playing with that Pete. I think it said the insulation was Teflon. It can be done with an X-acto knife very carefully. The soldering iron runs less risk of damaging the conductors, then scrape the melted bits off with fingernails. When I go into mass production, maybe one of those little alcohol burners, and just run a batch.

Can you burn off Teflon? I used to strip Teflon over solid wire with a needle nose by crushing the end of the wire. The Teflon would just split in halves. With stranded wire, it probably won't crush very usefully, but it's worth a try.

I have plenty Chuck, I'll give that a try. BTW, this IS the same wire that We Honest includes with their signal kits. So it turned out to be exactly the stuff I was looking for. For some reason, the spool seemed smaller.

I took a couple 120Ω's and put them together. Time to fire up the test unit!

Big_Boy_4005 posted:

I have plenty Chuck, I'll give that a try. BTW, this IS the same wire that We Honest includes with their signal kits. So it turned out to be exactly the stuff I was looking for. For some reason, the spool seemed smaller.

I took a couple 120Ω's and put them together. Time to fire up the test unit!

Who is We Honest? on the bay? I don't see them.

Oh I see them now with a Google search. eBay search didn't work.

Elliot, forget the resistors.  From the documentation.

No external resistors or potentiometers are needed as everything required is built into the SSD.

However, look closely at the documentation as there are assembly options based on what you're connecting.  Also, there's an interesting statement about the inputs...

These inputs must be provided from an open collector output (e.g. an output pin on a DOUT32 card).

How are you driving those inputs?

Then there's this comment...

Note: The quoted prices for all circuit boards provided by JLC Enterprises are for the fabricated boards only, that is, without any electronic parts. All circuit boards come unassembled and less components.

Yes I read that too. Really nothing needs nothing to be decided with that board.

Elliot, if the insulation is teflon just be aware the fumes from burning teflon are toxic. I suspect that small amount can blown away with a little fan though. 

Pete

John, the outputs from C/MRI are what they call "current sinking". I verified that when I checked the transistor being used, it's a 2N3904 on the output card. Is that the same as "open collector"? If that's correct, then I have that covered. When I get the cards, I'll order the components.

Pete, thanks for the heads up on the Teflon. I'll watch out for that.

The test unit has been running for 20 minutes. Haven't started a fire yet. 

It's only "open collector" if it's wired that way in the device.  I'm not familiar with how they're wired.

OK John, now this is the stuff I don't totally understand. What I do know from the enterTRAINment days is that the C/MRI outputs when active provide the ground side of the circuit. I was able to deduce this when I hooked up the turnout relays. There is a diode across each coil. Hook them up backwards and the fuse pops.  The relays are wired with the plus side in common and the ground switched. For now I'm just using micro switches to control them, but it should be a simple change over to C/MRI by swapping one wire per turnout. 

I also looked up that transistor on the output card, it's an NPN. I hope all the puzzle pieces are fitting together on this.

The verdict is in on the original question: Is the wire large enough to power the LED? I used about a 16" pair of leads, figuring that would be about the longest run, on a signal bridge. I let it run for 2 hours on the bench. No problem!

I guess the answer is yes. Thank you gang!!! This is just one of the building blocks for my signal system.

Just because I have an answer, doesn't mean the conversation has to stop.

Is this the wire? Not exactly cheap at $81, but looks interesting. 1000 Meters is a LOT of wire.

Black 7/0.05 teflon coated copper Ultra slim Wires Dia. 0.28 cable

There was never a doubt the LED would run with the wire.  Like I said, I doubt you can buy wire too small, at least not generally.

As for the C/MRI outputs, they sound like they might be Open-Collector, but since I don't know for sure, I hesitate to say "no problem".

Yeah Chuck, you found it. Wasn't a cheap spool, but there's a lot of it! They sell pre-cut lengths, but they aren't long enough to go where I need to go, so I went bulk. I might have just gone into the wire business without even trying. It's about 4¢ per foot if anyone needs some.

John, I'm so glad to have you, and the others, to walk me through some of this technical stuff. I see there might be a use for a scope in this C/MRI business after all.

I'll throw in that it seems likely that the 2n3904's on the C/MRI would just be open collector, but it's just as plausible that they have pull up resistors installed.  You could run a quick test with a volt meter between ground and an output.  Should read about 0 when the circuit is "on" and float near 0 when off if the collector is open.  If you get something like 5V, then there is a pull up.  You can convert this to an open collector with something like a 7407 non-inverting buffer IC.  

JGL

Norton posted:

Elliot, if the insulation is teflon just be aware the fumes from burning teflon are toxic. I suspect that small amount can blown away with a little fan though. 

Pete

That was my thought too, how are you going to strip the small wire?

When I worked for the civil service/Navy we used strippers for doing 30ga. solid wire field changes on UYK-7 computers.  The strippers had jaws that held the wire so it wouldn't slip and cutters that did the stripping; as you closed the handles the jaws clamped the wire and the cutters stripped the insulation in one motion.  You DID NOT want to get any skin caught between the moving parts cause once you started you had to complete the process before the thing would release.  They worked very well but being Uncle Sam I expect they were Mil-Spec and expensive.

Still if I did this all the time I would find some strippers like these that hold the wire in place while it strips/pulls the insulation off.

I Googled wire strippers and found some that were well over $200.

For #30 wire, this is the tool I use, works perfectly.

Miller NN014W No Nik Wire Stripper 86194

JGL, I'm 99.9% sure that no additional hardware is necessary in my situation. That comment in the documentation (second link above) was most likely intended for those who are using outputs from a "non-Chubb" source, such as Arduino. The point was to make sure that the inputs on the SSD card were being driven by the proper signal. My outputs are from Chubb output cards, they should be compliant. Unfortunately, there's no way to run tests at this time. I'm still a number of months away from powering the system up. The bulk of the system was preserved from enterTRAINment, but the serial line, linking the nodes, needs to be restored. I have my C/MRI manuals right by my side. I'll grab a pad of post-its and start marking the important "needles" in those two "haystacks". Trust me, there's a LOT of "hay".

Bob, part of me wishes that this was solid wire. It would be a little easier to deal with. I might have some wire strippers that could work, out in the garage. I'll go look. $200 is a little over the top for this, I have better places to put that money.

Thanks John, $40 is more in the ballpark.

The No-Nik stripper I use strictly for the #30 solid wire-wrap wire I use to wire many LED lights in engines and rolling stock.

I also have one of these from MicroMark, and I use it extensively for various wire sizes, it works very well.

Wire Stripper, AWG #20 - 30

I like that one John. It would also work well for the Cat5 conductors. The whole signaling system will use Cat5 to cover the distance between the C/MRI boards and the signals in the field. There will be a lot of wires to strip.

Jameco has it...

http://www.jameco.com/webapp/w...ybmcwCFQyHaQodyXEAXg

Cool Chuck, I will be placing a Jameco order when my cards come from Chubb anyway. I'll add it to my parts list.

Good news - bad news. My package from Jameco came today, and the wire strippers were in it. Unfortunately, when I went to try them on the wire in question, I discovered that it's smaller than 30 gauge. I can still put the strippers to good use, as they are far superior to the ones that I've been using for Cat5 conductors.

Back to the drawing board for stripping the tiny wire. We now know we are looking for 32 or smaller.