Connecting LED with Long Wires

[Luchik]

I'm designing a circuit with a LED.
The LED is not located on the PCB. it is connected using long wires, which will probably cause inductive load since the LED is toggled at a high rate.
My question is: will the parasitic induction be any different if the LED resistor is located on the PCB or close to the LED? is there any benefit from locating the resistor close to the LED?

[Blastcap]

How long is "long"?

[Luchik]

about 20cm.

[Whales]

How high of a frequency?

[Blastcap]

From gut feel anything below a few meters shouldn't be any problem, and then your first problem would probably be the wires resistance.
From what you are saying I'm guessing you are dimming it thru PWM, and those frequencies shouldn't really a problem.

Someone more experienced than me can probably get into the nitty gritty details, with simulations and all that jazz.
For a wire of that length, why don't you just take a 2m wire and try it. If it works with a wire 100 times the length, it should works with a 20cm wire.
... or do you by any chance have a 100kW UHF sender in your backyard, then 20cm might work as an antenna.

[batteksystem]

I'm designing a circuit with a LED.
The LED is not located on the PCB. it is connected using long wires, which will probably cause inductive load since the LED is toggled at a high rate.
My question is: will the parasitic induction be any different if the LED resistor is located on the PCB or close to the LED? is there any benefit from locating the resistor close to the LED?

Well how fast are you driving it?

[Back2Volts]

From gut feel anything below a few meters shouldn't be any problem, and then your first problem would probably be the wires resistance.
From what you are saying I'm guessing you are dimming it thru PWM, and those frequencies shouldn't really a problem.

Someone more experienced than me can probably get into the nitty gritty details, with simulations and all that jazz.
For a wire of that length, why don't you just take a 2m wire and try it. If it works with a wire 100 times the length, it should works with a 20cm wire.
... or do you by any chance have a 100kW UHF sender in your backyard, then 20cm might work as an antenna.

...and I thought the people "challenged" with the metric system where on this side of the pond !         Sure you meant 10 times...

[StillTrying]

Option 2, every time.

[Ian.M]

*IF* you have two wires per LED, its always better to put the current limit resistor at the supply end between the +V rail and the LED so that the rail is protected if the wiring gets shorted to ground.

[stj]

worse case scenario - you create a short range radio transmitter running in the low KHz band - what exactly would you effect with it in your house / place of work - likely nothing unless your a spy or something.

[Blastcap]

...and I thought the people "challenged" with the metric system where on this side of the pond !         Sure you meant 10 times...

Got me there.
What is an order of magnitude in these days?
I think i got confused by the 20cm as I'm used to work in mm.

[IanB]

Electrically it makes no difference where you put the resistor. It is always exactly the same circuit.

But for practical reasons note what Ian.M said and go with option 3: put the resistor on the positive supply side at the inboard end so that the output is current limited in the event of any faults.

[Kjelt]

If you cause interference at MHz's which can happen with the modern led driver ic's then place a ferrite bead with the impedance at the correct troubling frequency in series of the power supply to the led.
You have to experiment a bit with this, a spectrum analyzer can help. Alternative use a shielded cable with the shield grounded.

[Seekonk]

Option 3:  Always put the resistor on the power side.  You never know when the  leads can short to common.

[StillTrying]

The OP says:

connected using long wires, which will probably cause inductive load since the LED is toggled at a high rate.
My question is: will the parasitic induction be any different if the LED resistor is located on the PCB or close to the LED?

So, despite being a lone voice, I still say Option 2 would be best to reduce parasitic induction (and capacitance) effects on the switching transistor - and radiated from the wiring.

Of course barely matters in this particular case. But would it make any difference if the circuit was higher power, say a 12v 5A 12 meter PWM controlled LED strip around a ceiling, and the TR a fast switching MOSFET.

Would it still be best to put the resistance in the 12V line in case the 12V gets shorted, and then try to transmit the raw PWM around the room, - I know which option I'd choose if it was in my living room!

If you cause interference at MHz's which can happen with the modern led driver ic's then place a ferrite bead with the impedance at the correct troubling frequency in series of the power supply to the led.

"power supply to the led" is very ambiguous, and I think the spectrum analyzer might be a bit OTT. 

[IanB]

Would it still be best to put the resistance in the 12V line in case the 12V gets shorted, and then try to transmit the raw PWM around the room, - I know which option I'd choose if it was in my living room!

So you are saying the position of the resistor makes a difference to what the drive transistor sees? Can you elaborate on why you believe this to be the case? It does not seem to me that there would be any difference in the behavior of the circuit, but perhaps I am missing something.

[Ian.M]

How have we gone from the O.P.'s "about 20cm" wires to wires long enough to go around the living room?

Also, if you are PWMing LED lighting, and running a 12V supply through your ceiling, anyone with any sense at all will protect that circuit with a fuse or circuit breaker, which renders my response to the "Where to put the resistor?" question moot, and will use screened cable so it doesn't wipe out AM radio reception in the area.

[StillTrying]

So you are saying the position of the resistor makes a difference to what the drive transistor sees?

Yes, when you're passing square waves down a cable. In my defence you'll notice others are mentioning how some of the digital switching would escape from the cable.
Almost none of this applies DC, and not in this OP's case where the effects are going to be too small to notice.

but perhaps I am missing something.

Probably me too!

[mikerj]

Would it still be best to put the resistance in the 12V line in case the 12V gets shorted, and then try to transmit the raw PWM around the room, - I know which option I'd choose if it was in my living room!

So you are saying the position of the resistor makes a difference to what the drive transistor sees? Can you elaborate on why you believe this to be the case? It does not seem to me that there would be any difference in the behavior of the circuit, but perhaps I am missing something.

It does if the parasitic capacitance of the wires is significant.

[Zero999]

At higher frequencies the wire acts like a transmission line so to stop reflections, one end needs to be terminated in its characteristic impedance. In this case,  it's not possible to do that, since the LED needs a 20R resistor. I'd probably keep the resistor at the LED side and at the source side, add a reverse parallel diode in series with a suitable termination resistor to absorb the negative reflected pulse.

[MosherIV]

Hi

Why even think about pwm down a wire, why not pwm into a capacitor and put the capacitor before the wire/led. Then you will not get high freq going down the wire (well not much).

Or am I missing something? 

[Ian.M]

That will cause a severely non-linear response driving a LED.  Better to PWM an inductor in series with the LED, with a diode across both to allow the current to continue to circulate.   While you are doing that, add a low side current sense resistor and ditch the LED's series current limiting resistor, control the current by the PWM duty cycle and gain efficiency.

However the O.P. may not be doing PWM for brightness control - the circuit and component values look typical for an IR remote sender, in which case, its essential that any filtering has negligible effect at the IR carrier frequency.

[IanB]

Hi

Why even think about pwm down a wire, why not pwm into a capacitor and put the capacitor before the wire/led. Then you will not get high freq going down the wire (well not much).

Or am I missing something? 

Feeding PWM into a capacitor will waste energy at best^* and will blow up the drive transistor at worst. A capacitor looks like a short circuit to AC, and PWM contains a lot of AC.

* Any time you charge a capacitor from a voltage source you lose 50% of the supplied energy

[SethGI]

To answer the first question, for an application this simple, it shouldn't matter at all. The only thing you may want to take into account is that if the wires are REALLY long, they will have their own, no-longer negligible resistance which you should subtract from the value of the resistor.

[macboy]

Use twisted pair, like a pair from a CAT-5 cable, or a parallel pair like two adjacent conductors from a ribbon cable. The inductive coupling between them will ensure nearly zero total effective inductance, since there are equal and opposite currents in each conductor.

[Brumby]

To answer the first question, for an application this simple, it shouldn't matter at all.

This is what I was thinking.  20cm is not that long at all - and, although the OP hasn't come back with a frequency, I wouldn't be at all concerned running it up to 10kHz or even more.  If you were really worried, twist the wires together.

I wouldn't be getting into some of the 'advanced' considerations mentioned here unless we were rocking along at RF frequencies or had some particularly sensitive circuitry that needed an electrically quiet environment.

The only thing you may want to take into account is that if the wires are REALLY long, they will have their own, no-longer negligible resistance which you should subtract from the value of the resistor.

If the wires were 20m long, then that would be worth checking - but even 2m isn't much of an ordeal for the LED itself ... unless there are strong EM fields from sources in the area that could induce problematic currents.

Yes, there are considerations for looong wires, but I wouldn't be too worried about 20cm unless you were getting up in the UHF band.

[Kjelt]

Long time ago for me but 20cm is enough for a perfect couple of hundreds MHz antenna.
Just make sure those frequencies are not in the signal, and if I remember correctly from college, the square wave with Fourier harmonics, you must take care that the flanks of your switching signal are not steep, the steeper the higher frequencies will be present, so use a cheap RC filter, ferrite bead or shield it.

[kmossman]

What is gauge of the wire?   That is the most important thing.  There are resistance per metre/yard charts available on the Internet. Just look up and determine the total resistance of the wire. Subtract that from the calculated resistor value. If it is negative then use a heavier gauge.  Simple.