Simple Resistor Question

[AndyID]

10 hours ago, melmerby said:

A LED has a voltage/current curve in the usable area similar to (but less well defined than) a zener diode.

For most purposes it can be assumed to be essentally flat so the curent flow (and hence brightness) is defined by the series resistor whilst the voltage across the LED only varies slightly.

 

 

Yes, probably a good way to achieve success is to include an external resistance from a voltage source that's quite a bit higher than the forward voltage of the LED, but for those who really care to understand how this stuff actually works (kindly move on now if you don't) it's incorrect to imply that the current/voltage relationship is "flat", by which I assume you mean the forward voltage is constant regardless of the current.

 

It ain't.

 

All LEDs have some internal, non-linear, resistance. A typical value is around 10 ohms so it is possible to take advantage of that resistance to limit the LED current. I would not recommend doing that but it is possible.

 

The best way to drive a LED is from a constant current source but that's going even further off topic.

[melmerby]

 

16 hours ago, melmerby said:

A LED has a voltage/current curve in the usable area similar to (but less well defined than) a zener diode.

For most purposes it can be assumed to be essentally flat so the curent flow (and hence brightness) is defined by the series resistor whilst the voltage across the LED only varies slightly.

 

 

 

4 hours ago, AndyID said:

 it's incorrect to imply that the current/voltage relationship is "flat", by which I assume you mean the forward voltage is constant regardless of the current.

 

It ain't.

 

All LEDs have some internal, non-linear, resistance. A typical value is around 10 ohms so it is possible to take advantage of that resistance to limit the LED current. I would not recommend doing that but it is possible.

 

 

Read my post again

What I said is true.

See this typical characteristic below.

The voltage across a LED in it's usable area does not vary greatly with current flow.

e.g. from 5mA to 50mA (a 10:1 ratio) the voltage varies from about 1.75V to 1.95 V (a 1:10 ratio)

Edited June 3, 2019 by melmerby

[Crosland]

On 02/06/2019 at 10:11, DavidCBroad said:

Your reply is also unhelpful.    If the Vf of the LED is less than the supply voltage and the resulting brightness its not too bright for your taste  you don't need resistors.  Don't perpetuate the myth that you must use resistors with LEDs per se.  The world has moved on.

Only if the internal resistance of the supply is so great that it substitutes for the resistor, or you use an appropriate constant current supply.

 

Please stop perpetuating this rubbish.

 

A lot of modelers will not have a clue about LED Vf, cell internal resistance or constant current sources so the best advice is to use a resistor. ALWAYS. FULL STOP.

[DCB]

Should have said if the MAX Vf of the LED is less than the supply voltage.   If it keeps you happy fit a resistor,  If you have a 12/24/9/6/5/4/240/110 volt supply fit a resistor.

[Gordon H]

Should have said nothing at all.

[sharris]

5 hours ago, DavidCBroad said:

Should have said if the MAX Vf of the LED is less than the supply voltage.   If it keeps you happy fit a resistor,  If you have a 12/24/9/6/5/4/240/110 volt supply fit a resistor.

 

Here's a challenge: 

 

For a typical LED (doesn't matter the colour) picked up from, say, the AllComponents stand at an exhibition (other purveyors of components to modellers are also available), tell me what the forward voltage is, and what voltage supply you'd be happy to use without a resistor. 

 

Here's a link to one manufacturer's list of colours and forward voltages (other LED manufacturers also available):

https://www.lumex.com/led-color-guide.html

 

Notice that depending on the chemistry you can have quite a variety of Vfs for a particular colour:

red: types with max Vf from 2.2-2.8V    (typical Vfs are, of course a bit lower)

yellow: types with max Vf from 2.5-2.8V

green: types with max Vf from 2.5-4V 

 

Even if you think your power supply voltage matches your LED Vf, or you think can get away with it, make the wrong assumption about which type of green, red or yellow LED you picked up from the stand and 'I can connect my LED directly to my power supply'  could easily blow your LEDs - or worse, tell someone else they can connect their green LEDs directly to a power supply and they happen to have a different variety of LED -  they could blow their LEDs without knowing why. 

 

 

[AndyID]

21 hours ago, melmerby said:

 

 

Read my post again

What I said is true.

See this typical characteristic below.

The voltage across a LED in it's usable area does not vary greatly with current flow.

e.g. from 5mA to 50mA (a 10:1 ratio) the voltage varies from about 1.75V to 1.95 V (a 1:10 ratio)

 

Read my post again.
 

This graph has a slope. It may approximate to a straight line but it is not "flat". In this case the the LED has a resistance of around five ohms.

 

Still, as Crosland strongly suggests, USE A BLEEPING RESISTOR! Resistor kits with nearly all the values you will likely ever need are available for minimal outlay online. This is one way you can store them so you can find the one you want.

 

 

The test-tubes are not expensive either, but you may need to find a cabinet maker if you want such a finely crafted holder.

 

Oh, and by the way, your clever flashing Ikon tends to trigger my ocular migraine. It might also induce epileptic seizures in some people. Just a suggestion, but you might consider an alternative.

 

 

 

 

Edited June 4, 2019 by AndyID
yup

[melmerby]

5 hours ago, AndyID said:

 

Read my post again.
 

This graph has a slope. It may approximate to a straight line but it is not "flat". In this case the the LED has a resistance of around five ohms.

Oh, and by the way, your clever flashing Ikon tends to trigger my ocular migraine. It might also induce epileptic seizures in some people. Just a suggestion, but you might consider an alternative.

 

Yes, I have read your post again , you accuse me of posting something I didn't do. At no time did I say it was flat. 

What I actually said "For most purposes it can be assumed to be essentally flat" which is true and not what you are accusing me of saying.

 

That internal 5 ohms you point out is a fraction of the total effective resistance of around 60 ohms* and insignificant when you are working out the value of series resistance to use.

It will also not be enough to stop destruction of the device if you don't limit the current by other means.

 

As regard the Icon, if it is a problem I will change it, I have used it for some years without previous comment.

 

*EDIT this LED has a restance of 180 ohms with 10mA flowing and 63 ohms with 30 mA flowing. i.e the effective resistance drops as the current increases. (as do all LEDs)

 

 

Edited June 4, 2019 by melmerby

[Junctionmad]

jeepers , many posts arguing about not using a 2 cents resistor

 

Leds need a current limit , thats easiest to implement as a resistor . End of story (  minute edge cases not withstanding ) 

[AndyID]

On 04/06/2019 at 01:55, melmerby said:

Yes, I have read your post again , you accuse me of posting something I didn't do. At no time did I say it was flat. 

What I actually said "For most purposes it can be assumed to be essentally flat" which is true and not what you are accusing me of saying.

 

That internal 5 ohms you point out is a fraction of the total effective resistance of around 60 ohms* and insignificant when you are working out the value of series resistance to use.

It will also not be enough to stop destruction of the device if you don't limit the current by other means.

 

As regard the Icon, if it is a problem I will change it, I have used it for some years without previous comment.

 

*EDIT this LED has a restance of 180 ohms with 10mA flowing and 63 ohms with 30 mA flowing. i.e the effective resistance drops as the current increases. (as do all LEDs)

 

 

 

Thank you very much for the new ikon!

 

I've struggled with unpleasant ocular migraines for sixty years. (For anyone who hasn't had the pleasure it's a bit like looking at the World through a kaleidoscope or an ever-changing stained glass window.)

 

OK, so not exactly flat.

 

The resistance thing does not work quite like that. An idealized model of a LED might be a constant voltage device (a battery perhaps) in series with a resistor. So there is already a "resistor" in the circuit, and it has an almost constant value. You can determine its value from the straight-line bit of the graph.

 

In the example you provided while the voltage increases 100mV (1.8 to 1.9 volts) the current increases 20mA (10 to 30 mA). From Ohm's Law, 100/20 = 5

 

It's not a large value and this might be a lot more information than some might like to hear, but it's not wickedly complicated either.

 

Andy.

[Grovenor]

And in his third post, a long way back in this topic the OP did get around to mentioning that the LEDs in question are to be drven by a DCC decoder, which for most decoders means an approximately 12V supply and makes most odf the above discussion a misleading waste of effort. If we pay attention to the questions we will be a lot more helpful.

Rgds

[AndyID]

1 minute ago, Grovenor said:

And in his third post, a long way back in this topic the OP did get around to mentioning that the LEDs in question are to be drven by a DCC decoder, which for most decoders means an approximately 12V supply and makes most odf the above discussion a misleading waste of effort. If we pay attention to the questions we will be a lot more helpful.

Rgds

 

True, but how much fun would that be?

[melmerby]

2 minutes ago, Grovenor said:

And in his third post, a long way back in this topic the OP did get around to mentioning that the LEDs in question are to be drven by a DCC decoder, which for most decoders means an approximately 12V supply and makes most odf the above discussion a misleading waste of effort. If we pay attention to the questions we will be a lot more helpful.

Rgds

I did in fact mention that until the drift set in

I asked whether Sir TH had a 12-15v power supply with which to try his LEDs with.

[BWsTrains]

LEDs (all of the same type) 1+ in number can be installed in series with one resistor matched to the unloaded PS voltage as shown in this spreadsheet.

 

Step 1 Input the Vf for the LED(s) and the supply voltage

Step 2 Derive the Resistance for the desired Voltage drop and LED current

 

The calc is shown for 1 and 2 orange LEDs that are quite low Vf. The resistor must have the balance of the voltage drop across it.

Using V=IR allows you to change the R variable (Ohms, green cells) to get a suitable mA for the circuit.

 

Here two different solutions for a single LED are shown, I tend to favour a resistor to run the LED somewhat below it's "IF Typ" rated value

 

 

 

Copy of LED Installation notes.xlsx

Edited June 5, 2019 by BWsTrains
typos and update image

[DCB]

The key phrase is "Check the supply voltage with meter"    ( see above)

 

[sharris]

26 minutes ago, DavidCBroad said:

The key phrase is "Check the supply voltage with meter"    ( see above)

 

 

I think the key phrases can be found on page one of the thread where SirTH indicates he has the answer - after which we've all been confusing the issue with science and bad science. 

 

[kevinlms]

46 minutes ago, sharris said:

 

I think the key phrases can be found on page one of the thread where SirTH indicates he has the answer - after which we've all been confusing the issue with science and bad science. 

 

Yes, yet again the thread got diverted by the 'You don't need resistors' poster.

[WIMorrison]

10 minutes ago, kevinlms said:

Yes, yet again the thread got diverted by the 'You don't need resistors' poster.

 

In fairness you don't need a resistor if you want to use the LEDs as single use 'flash bulbs' like we used to but for the top of cameras, however if you want to re-use them then it would be advisable to stick some resistors into the circuit  (and it saves you needing an unlimited supply of LEDs!)

[sharris]

Not just short flash-bulbs, but electronic mortar launchers - once, many years ago as a child before I understood how to wire up LEDs properly, I had a line of them in parallel and a single series resistor (don't do this!) - a short across the resistor led to the full supply voltage being applied to the LEDs shortly followed by very brief flashes, popping sounds and the tops of the LED packages being launched some distance across the room. I believe the smoke released from them during this 'experiment' was actually quite toxic.

 

 

[DCB]

1 hour ago, sharris said:

Not just short flash-bulbs, but electronic mortar launchers - once, many years ago as a child before I understood how to wire up LEDs properly, I had a line of them in parallel and a single series resistor (don't do this!) - a short across the resistor led to the full supply voltage being applied to the LEDs shortly followed by very brief flashes, popping sounds and the tops of the LED packages being launched some distance across the room. I believe the smoke released from them during this 'experiment' was actually quite toxic.

 

 

Exactly the problem you get when you use a high supply voltage and rely on the resistor to stop the LEDs failing.  Its very easy to short a resistor and blow an expensive LED signal or yard light when using 12 or 24 or even 6 volts.  I don't  have that problem,  I check every LED across the power unit and make sure they are happy, not too bright, or too dim, don't go bang or smoke,  That's sometimes 2 X AA batteries, sometimes one of two wallwart transformers.  I never had a LED fail after testing, sometimes a resistor is needed to trim the colour but its down in the low hundred ohmss rather than the 1 K range.  The last test off the power unit had an orange LED barely glow, that is the actual junction inside the LED barely glow, on a 1 K and was still too dim to be any use on a 560,  a 56 would be a good choice if I did feel the need to use as I don't think 0 ohms resistors are available.  The 30 odd LEDs on this low voltage bus  provide a good level of illumination for the buildings in a darkened room and are very reliable. Think of it this way, the wall wart takes the place of the resistor between the mains and the LED.  I'm Just off to try to test / blow up some  LEDs by putting them across 3 X AAA cells without a resistor. I'm not expecting any  to fail.  

Edited June 6, 2019 by DavidCBroad
Boredom

[WIMorrison]

Yawn ...

[sharris]

2 hours ago, DavidCBroad said:

Exactly the problem you get when you use a high supply voltage and rely on the resistor to stop the LEDs failing.  Its very easy to short a resistor and blow an expensive LED signal or yard light when using 12 or 24 or even 6 volts.   

 

1. I think we ascertained by post 10, about half way down the first page that we were dealing with DCC here, so we don't have the luxury of choosing a voltage source to exactly match the LED Vf - we're stuck with something around 14V.  Incidentally, doesn't it get rather expensive sourcing a power supply to exactly match each LED circuit, rather than spending a penny or two on a resistor for each LED and running them all from the same power source?

 

2. It's easy to short a resistor if you leave it openly waving around on a wire without sleeving - however it's equally easy to secure and insulate it properly to ensure that stray leads can't short across it.  Generally they're quite robust devices so if you don't abuse them they're not going to short out on their own. 

 

I would refer you to the phrase in my original text "before I understood how to wire up LEDs properly" as being the reason for my explosive LEDs. 

 

Quote

I don't think 0 ohms resistors are available.

You would think wrongly then:  https://en.wikipedia.org/wiki/Zero-ohm_link  - indeed, if you look in the resistors section of Farnell and limit your search to 0ohm in the 'resistance' box, it will give you 370 choices of them. 

 

Edited June 6, 2019 by sharris

[DCB]

This is actually a Non DCC section.      

[sharris]

2 minutes ago, DavidCBroad said:

This is actually a Non DCC section.      

Which still isn't an excuse for giving bad advice. 

[WIMorrison]

deleted

 

... duplicate post from 3 hours ago?

Edited June 6, 2019 by WIMorrison