Do you think an LED is a resistor?

[Sredni]

Ok, so you too think that there are no nonlinear resistors. They are just one-ports. In my view, using set theory

One port < Resistor U Capacitor U Inductor U Memristor U Any circuits composed of the connection of the previous four

And then I specialize each sunset into linear and nonlinear. So to me (and to Chua, Desoer and Kuh) nonlinear resistors are a specialization of one ports, like linear resistors from which they are disjoint.

But let's switch for a second to inductors: what do you call a coil with a ferrite core? A one port (but not an inductor)? Because I guess inductor only means linear inductor, right?

I call it a non-linear coil.  Under reasonable limited conditions, it exhibits inductance.  In Spice, it can be modeled as a one-port that includes hysteresis.
Of course there are one-ports, encased in plastic, that can be characterized as non-linear resistors (with non-linear V vs. I equations), but extending that nomenclature to semiconductor devices is useless sophistry.

Ok, therefore the big toroidal core coils used in switched power supplies are not inductors, but something different called nonlinear coil. Are we clear on that?
Can we also say that all inductors in a SEPIC are
 linear? What does the I stand for in the acronym?

[TimFox]

Yes, the “big toroidal core coils” (your phrase) have inductance.
They also have resistance.
Therefore you (but not I) can call them “resistors” and confuse them with RN65D metal-film thingies.

[Sredni]

As I have already explained at least three times, I am neglecting secondary effects. The ESR of a coil is a side effect. (While the static resistance exhibited by a diode is the main effect).

A 300uH coil to me is an inductor, irregardless of the fact that it be wound around a ferrite core or not. If it is nonlinear and shows saturation and hysterisis it is a nonlinear inductor and as such, it still belongs to the broader category of Inductors, side by side with the linear inductors.

While for people who do not believe in the existence on nonlinear inductors, it is a nonlinear coil (NLC, in brief). Like those used in SEPNLCC circuits.

[TimFox]

Enough nonsense—back to the original question.
I have two similar-size cylinders with axial leads.
One is an RN65D metal-film resistor.
Testing it, I see that it obeys Ohm’s law quite well, but has small parasitic inductance and capacitance, along with a slight self-heating and temperature co-efficient.
Ignoring the minor deviations from the Platonic ideal, I confirm the common opinion that it be called “resistor”.
The other is labeled 1N4007.
Initial measurements show it is not close to Ohmic, especially with respect to polarity reversal.
Testing it, I see that it conforms well to Shockley’s equation for a semiconductor diode (including temperature effects) with small parasitic series resistance (in the package), along with non-linear junction capacitance and charge storage.
Again, following your advice to ignore secondary effects, I confirm its proper name as “diode”.
Calling it a “resistor” would only be confusing.

[Kim Christensen]

What are the physical units of this "rectification"?
I use resistance and it is expressed in ohms. I can tell you exactly how many ohms a silicon diode 1N4148 exhibits at a given voltage or current.

[Sredni]

Enough nonsense—back to the original question.
I have two similar-size cylinders with axial leads.
One is an RN65D metal-film resistor.
Testing it, I see that it obeys Ohm’s law quite well, but has small parasitic inductance and capacitance, along with a slight self-heating and temperature co-efficient.
Ignoring the minor deviations from the Platonic ideal, I confirm the common opinion that it be called “resistor”.
The other is labeled 1N4007.
Initial measurements show it is not close to Ohmic, especially with respect to polarity reversal.
Testing it, I see that it conforms well to Shockley’s equation for a semiconductor diode (including temperature effects) with small parasitic series resistance (in the package), along with non-linear junction capacitance and charge storage.
Again, following your advice to ignore secondary effects, I confirm its proper name as “diode”.
Calling it a “resistor” would only be confusing.

Ok, this is the VI characteristic of a 1N4148 diode in LTSpice (default temperature) - sorry I went on simulating and picked 1n4148 instead of 1n4007. But it's immaterial.



and this is the resistance offered by the diode as a function of the voltage across it



So, we are now seeing the diode as a voltage dependent resistor. Let's see... what is the resistance 400mV? Let's zoom in:



I'd say it's about 23.2 kohm.
Let's see what is the resistance at, I don't know, 660 mV (about 5mA of diode current). We can compute it by hand of course, but on the graph we see it is 132 ohm.



Now, let's see if we can make something with these values...
Let's say I have a 5V supply and a diode that has exactly that characteristic (you can find yours experimentally at your lab temp).
If the diode at 5mA shows a resistance of 132 ohms, we can create a circuit that drives 5mA by using a resistor such as

5mA (Rx + Rdiode) = 5V

or Rx = 5V / 5mA - 132 = 1000 - 132 = 868 ohm

Oh, look: 660 mV voltage, 5 mA current. What a lucky break.



How about the other one?

The diode shows 23.2 kohm at 0.4V.

Let's change equation, since I did not record the current. Let's use a voltage divider equation

Rdiode/ (Rx + Rdiode) 5V = 0.4V

23.2k / (Rx + 23.2k) = 0.4 / 5

Rx = 23.2k 5 / 0.4 - 23.2k = 266.8 kohm

and, oh look:



0.4V.
Now, let's see what we can do with resistors:



Ok, exact same results, if we neglect a bit of rounding error in reading and setting the values.
Now, take your black boxes out of the fridge. Put the diodes D1 and D2, and the resistors R1 and R2 inside a black box each. Shuffle them around. And tell me: without looking inside the black boxes and without resorting to second order effects (like temperature dependence, or changing the other circuital parameters to change the operating points) can you tell me which are the diodes and which are the resistors, by simply measuring voltages, currents and powers?

[schmitt trigger]

I sense an opportunity to bring this daft thread to a close.

I believe the thread is rapidly approaching the 200 post milestone.

[Sredni]

What exactly is your equivalent circuit for a diode (such as the 1N4148, you mentioned)?

Let's put it in a black box, and let an independent observer, see if they can tell your equivalent circuit, from another black box, containing a real diode.

There is no need for an equivalent circuit: you specify the VI characteristic. Then if you want you can use other elements to approximate it, but it's not necessary. There is a symbol for nonlinear resistors. And nonlinear inductors. And nonlinear capacitors...

Inspiration for this post, came from here:
https://www.sciencedirect.com/topics/engineering/circuit-theory

Ah, ScienceDirect. There are several articles linked at that page. Which one are you referring to? Could it be this one?
https://www.sciencedirect.com/topics/engineering/nonlinear-resistor

[Circlotron]

There the cats whiskers when compared to many other illuminating devices

I wonder if anyone has succeeded in using a suitably biased LED as an RF detector?

[Kim Christensen]

Now, take your black boxes out of the fridge. Put the diodes D1 and D2, and the resistors R1 and R2 inside a black box each. Shuffle them around. And tell me: without looking inside the black boxes and without resorting to second order effects (like temperature dependence, or changing the other circuital parameters to change the operating points) can you tell me which are the diodes and which are the resistors, by simply measuring voltages, currents and powers?

All we have to do is change the supply voltages.

Or stick an ammeter across each "resistor" in turn. 

[Sredni]

You still don't get the meaning of nonlinear? It means it is NOT linear. The resistance is NOT constant. That is the whole point.
But if you wish I think you could waste some time creating a voltage dependent resistor in LTSpice by copying the V-R characteristic I show above, and then you can change supply voltages all you want.

Nonlinear resistors exist and the term is used in engineering.

Go read the science direct link I posted above.

[Kim Christensen]

I don't need to. I understand the concept 100%. I just don't see the point. 

[Someone]

non linear system has linear approximation around a small operating point....

STOP THE PRESSES!

... oh wait, that's the underlying principle of how spice AC analysis works.

[electrodacus]

Ok, this is the VI characteristic of a 1N4148 diode in LTSpice (default temperature) - sorry I went on simulating and picked 1n4148 instead of 1n4007. But it's immaterial.

You can do the same thing for an inductor. Do you consider that all components are just resistors ?
A diode is the equivalent of a one-way valve.
While there is a minimum pressure needed to have the one-way valve open same as a minimum voltage is require for the diode to allow electrons to pass trough.
You will not call a one-way valve a restriction same as you will not call a diode a resistor.
Also one way valve will do what the name implies and allow the fluid to flow only in one direction but same as a diode if the pressure / voltage is high enough it will get damaged and then when they are damaged they are no longer a one-way valve or a diode but your favorite component a resistor.
So I'm OK to call a damaged diode a resistor but a working diode is not a resistor.

A real diode unlike an ideal diode has non ideal characteristics that impact the performance. It has resistance, inductance and capacitance and this non ideal proprieties are refereed to as parasitic elements. This including resistance are not desirable properties.

[Sredni]

non linear system has linear approximation around a small operating point....

STOP THE PRESSES!

... oh wait, that's the underlying principle of how spice AC analysis works.

Small signal analysis has nothing to do with what we are discussing here.
The resistance I talk about is the static resistance, not the dynamic or incremental resistance of small signal analysis.

Try again.

[Sredni]

Ok, this is the VI characteristic of a 1N4148 diode in LTSpice (default temperature) - sorry I went on simulating and picked 1n4148 instead of 1n4007. But it's immaterial.

You can do the same thing for an inductor.

No, you can't.
Different state variables.

[BU508A]

A LED is not a resistor.

This is a resistor which can emit light:

[SiliconWizard]

That's, once again, just about modeling. You can model all your heart's content, if you find that fun or see any practical use. Ultimately, it's just MODELS.

You're taking an I-V characteristic, basically, and claim that R = U/I, except that R here is a function of U (or I). Which makes it a dubious use of the "resistor" term, as quite a few have already said.
It doesn't serve much purpose either, other than playing with I-V characteristics, which is something that is pretty basic electronics, and obviously useful in itself.

And yes, inductors and capacitors are a slightly different beast, as they involve derivatives and their model is thus a differential equation. But you can still express them in terms of a relationship between U and I; no pun intended.
If you further twist your definition of a resistor with differential terms, then inductors and capacitors also become "resistors" in your definition.

Obviously that's still playing with models, as real-world parts all have resistive, capactive and inductive components as parasitics, and while useful, even the Shockley approximation is ultimately just for the birds, as someone would say.

[Berni]

Resistor < Linear resistor U Nonlinear resistor
Nonlinear resistor < incandescent lamps U diodes U ...

So, you can still call it a diode, recognize that it is a nonlinear resistor and, as such, that it belongs to the more general set of resistors.

Does this make any sense to you?

It does not make sense.

All resistors exhibit electrical resistance.
All of those semiconductor devices indeed also exhibit electrical resistance.
However not all devices that exhibit electrical resistance are resistors as they are not components purposely designed to implement a well defined resistance.

Does not matter what you understand as "resistor", the vast majority of forum members you are talking to on here understand it as a device specifically designed to implement a well defined amount of resistance. You don't have the authority to redefine established industry words used by others.

What you are trying to say is that a diode exhibits the effect of "electrical resistance" or "resistivity".

This effect is not particularly special and just describes that the device can consume electrical power and turn it into something else,

The point I make is that this is ALL a diode does.
Huge resistance when reverse biased, small resistance when forward biased. This is not a side effect. It is what it does (if we neglect secondary effects due to parasitics in real devices).
It does not store energy in the electric field.
It does not store energy in the magnetic field.
It does not do whatever sorcery a memristor does.
It just oppose a resistance that takes power out of the circuit.

Yes we all agree on here that at some fixed DC operating point a diode acts like resistance.

The point is that a diode is nothing special at doing this. The laws of physics force all power consuming components to look like resistors in steady state DC. This means that at a fixed DC operating point even a memristor is actually just electrical resistance, much like a diode acts as purely electrical resistance at that DC state.

Like what else do you expect an component to do at DC? It can either act as a power source (like a voltage or current source) or it can resist the flow of current hence resistance. There is nothing else for a component to do at DC.

[Someone]

non linear system has linear approximation around a small operating point....

STOP THE PRESSES!

... oh wait, that's the underlying principle of how spice AC analysis works.

Small signal analysis has nothing to do with what we are discussing here.
The resistance I talk about is the static resistance, not the dynamic or incremental resistance of small signal analysis.

Try again.

That was your long winded and dithering proof:

So, we are now seeing the diode as a voltage dependent resistor. Let's see... what is the resistance 400mV? Let's zoom in:

[MASSIVE IMAGE]

I'd say it's about 23.2 kohm.
Let's see what is the resistance at, I don't know, 660 mV (about 5mA of diode current). We can compute it by hand of course, but on the graph we see it is 132 ohm.

[MASSIVE IMAGE]

Now, let's see if we can make something with these values...
[snipping conversational fluff]
Ok, exact same results, if we neglect a bit of rounding error in reading and setting the values.
Now, take your black boxes out of the fridge. Put the diodes D1 and D2, and the resistors R1 and R2 inside a black box each. Shuffle them around. And tell me: without looking inside the black boxes and without resorting to second order effects (like temperature dependence, or changing the other circuital parameters to change the operating points) can you tell me which are the diodes and which are the resistors, by simply measuring voltages, currents and powers?

So to try and claim you're not relying on the well known small signal AC parameters is plainly incorrect.

If you dont like relying on small signal characteristics, perhaps "try again" with your explanation/justification.

[Siwastaja]

So,

A resistor is a component, which is purposely made resistor-like.

Most people would agree that to be resistor-like, there are two conditions;

1) Absence (or minimizing) of capacitive and inductive behavior, in other words, lack of time-dependent state (voltage and current are in relationship at any point in time, but not in relationship based on their history)

2) Linearity of the relationship between voltage and current, such that a simple scalar constant defines the relationship.

Sredni accepts (1) but doesn't accept (2).

Second condition is sometimes relaxed - e.g. variable resistors, thermistors, light-dependent resistors - but they are never called just "resistors", but with resistor only as part of the full name.

I don't suggest renaming something which already has another name, though, such as "light-emitting diode", as "non-linear light emitting resistor", even if you technically could. Maybe if it was originally named that way, we would be fine with the name. Then we would likely see Sredni creating a thread about how this non-linear resistor should be actually called a "diode".

[magic]

Resistors are bipolar (non-rectifying), non-light-emitting blue LEDs whose forward voltage just happens to depend on forward current linearly rather than logarithmically.

Prove me wrong

[Gyro]

A LED is not a resistor.

This is a resistor which can emit light:

Its operation has been studied and characterized too (attached, those happy days!).

[m k]

If only tool is a wheatstone bridge...

[Retep]

The poll should have had the option: "No. it is not a resistor, so it has a different name."