Well, not only was it uphill to school and back for me, I was on occasion chased by dinosaurs.
So don't believe that nonsense about evolution, either.
There are those who argue the earth is flat and those that argue it is otherwise and it's unlikely that both parties will ever agree.
In that context, asking "is the world round or flat?" is akin to asking "what is the resistance of an LED?"
In the general sense, the world is round and an LED (or other diode junction) does not really have the property of linear resistance.
On the other hand, in the context of building a house, I treat the earth as if it were flat, because over the very short distance of the size of a house, it can be considered flat.
By the same token, if I wish to know the slope of the curve of a diode at a given current over only a -very- small change in current, for instance as part of an attenuator for very small signals or in order to calculate the small signal ac gain in a bypassed common emitter BJT amplifier, I can treat the diode/LED junction as if it were linear. But only over that very short distance.
68tjs:
fungus:
At any instant in time though, Ohm's law applies.+1
UnoDueTre:
Or the fungus extrapolation of ohms law.In either case for a DIODE you must consider only the current, voltage is imposed by the current,
I disagree, voltage determines the current...
However
Since the current will rise when heat is applied, ohms law is used to prevent the led from consuming more current.
(If the LED is heatsinked I've gotton away with powering them via voltage alone without issues for several 10 watt led)
If voltage is controlled, then current is the variable. If current is controlled, then voltage is the variable.
polymorph:
If voltage is controlled, then current is the variable. If current is controlled, then voltage is the variable.
Supply 1 volt to an LED, see how far you get controlling the current.
uphill both ways.
How did that work?
JimboZA:
uphill both ways.
How did that work?
Simple
/^
/
home_ / _school
hill
retrolefty:
JimboZA:
uphill both ways.
How did that work?
Simple
/^
/
home_ / _school
hill
Ah ok.... so net change in height = 0 so no work done. Therefore actually the same as this:
home --------- school
and therefore the world is flat. 8)
I bet that back in the day Lefty, you were an ace at doing those line printer calendar pix of Snoopy and so on.
LED's do not follow "Ohm's Law". Their current vs voltage characteristic is not linear at all. It is not even locally linear.
An LED has a threshold voltage, which depends on the type of semiconductor junction in them, and tends to be a characteristic value for different colours. The threshold voltage is typically between 1.9 and 3.5 volts.
If you apply a voltage less than the threshold voltage to the led, very little current will flow and it won't light up. The apparent resistance is very high.
Once the threshold voltage is reached, a lot of current will flow. If you try to increase the voltage some more, too much current will flow and the LED will blow up.
You can safely use a 20vdc power supply to an led without a resistor...
Providing .....
- there's no stored capacitance.
- The 20v power supply is rated 20ma 8)
In that case, it becomes a 20mA power supply and no longer puts out 20V.
OK, thanks guys - hope you had fun at the expense of a newbie thread... 
No problem really, i like going off-topic too !
UnoDueTre:
The resistance of the LED is not linear and will vary according to the applied voltage.
See this for more info:
Is a current-limiting resistor required for LEDs if the forward voltage and supply voltage are equal? - Electrical Engineering Stack ExchangeBTW, like the avatar.
Thanks 
And thanks for the reference page, will try to undersand it properly.
polymorph:
...So we can play semantics 'til the cows come home, or realize that calculating a resistance number for an LED based solely on the voltage drop divided by the current is a meaningless number, because as soon as you change anything, you'll come up with a different answer.
OK, i'll just have to accept that what i'm trying to do will not work.
Will just have to measure the AnalogRead() in the actual setup and that it cannot be theoretically pre-calculated.
that it cannot be theoretically pre-calculated
Yes it can, it is just that it is difficult and involves non linear equations.
It would involve knowing the voltage / current curve for the LED and using that relationship in the calculations in place of resistance.
Anyway no one in there right mind would attemp to make a circuit like you are suggesting. What are you trying to do? What ever it is you are going about it wrong.
Grumpy_Mike:
that it cannot be theoretically pre-calculated
Yes it can, it is just that it is difficult and involves non linear equations.
It would involve knowing the voltage / current curve for the LED and using that relationship in the calculations in place of resistance.
Anyway no one in there right mind would attemp to make a circuit like you are suggesting. What are you trying to do? What ever it is you are going about it wrong.
I'm fiddling about with this thread;
http://forum.arduino.cc/index.php/topic,8558.0.html
I added the LEDs to act as visual confirmation for this n00b's peace of mind.
What i'm trying to do is set up a spreadsheet to play around with various resistor values to get the ideal AnalogRead() for all the buttons which won't overlap.
I get that if one were mathematically inclined, one could calculate it totally without having to plug in numbers by trial and error - but having the graph plots (in the spreadsheet) helps my understanding as i see the visual changes with each different resistor used.
I have everything working as per theory with just the resistors (which should really be sufficient) but was just curious how one would go about this if an LED was "necessary" with that circuit.
Basically, i'm satisfied with the "not (practically) possible" answer but wouldn't turn away any purely academic response.
Hi, if Georg Ohm could see this thread, wow.
[stir] Now can I mention Kirchoff and Norton [/stir]
Tom..... 8) (Gunna be 27 DegC tomorrow, 12/1/2014, looking at 41 to 43 DegC later in the week.) 
TomGeorge:
Hi (Gunna be 27 DegC tomorrow, 12/1/2014, looking at 41 to 43 DegC later in the week.)
High 20s here as we speak