Hi Cyril, your classification sounds logical (all four controlled sources are active).

However, what about all the sensors that convert mechanical or optical into electrical energy?

Lutz

Dear Sir, i think that passive means dead i.e those elements which do not contain any internal energy initially (at the time of formation) and those have some internal energy as we called active e,g diode,Transistor,SCR family and all types of sources.

As engineers, we use models to simplify the behaviors of real-world devices in order to predict how they will behave under specific design conditions. It is important to understand that the model is always and will forever be an approximation to the actual devices. Furthermore, the best models do not attempt to "do it all". Instead they only describe just those behaviors we are interested in for a particular task at hand. A good example of that are the "small-signal" and "large-signal" models of a transistor.

I would expand Lutz's definition to include any transducer that converts energy to/from electrical current. These may be considered controlled sources, though they are generally more than just ideal sources. That is how we model them as engineers, e.g. a photovoltaic cell is modeled as a PN junction (passive) in parallel with a photon flux-controlled current source with a shunt resistance (Norton equivalent) in series with an "internal" series resistance. Here is the simple (first-order) model of the PV cell.

http://alturl.com/giy6e

In general, a passive element is any circuit element that cannot provide signal gain. A source is a circuit element that supplies energy to a circuit. Note that reactive elements are capable of being at once both passive and sources. A gain element is any element that can boost the power of an input, or maintain it at a constant level (unity gain) after some processing is performed, such as impedance transformation, frequency domain filtering, etc. The memristor element, as well as certain diodes, are unique in that they are non-linear devices as well as being passive. The non-linearity can, under specific conditions, allow these passive devices to oscillate and deliver signal energy to a load due to negative resistance. So under these operating conditions a non-linear passive can be considered active under the traditional definition I gave above.

Thus the traditional definition of passive -vs- active, in my thinking should really be based on the whether a device is capable of non-linear behavior. resistors, capacitors, inductors are strictly first-order linear devices. Semiconductors in general, are not. Sure we use them to "approximate" linear transfer functions. But it is the nonlinear behavior that drives the motivation for the small-signal model of the transistor. If they were truly linear we would have no need for it.

Dear Ashok,

I like your figurative "passive means dead"; we even intended to use this word in the question above, only in quotes...

Well, passive elements are "dead"; but can we and how do we "revive" them? How can we inject energy into passive circuits to make them active? This is the main question in electronics...

Regards, Cyril

Gregg, thank you for the comprehensive answer. It arouses more interesting questions about the nature of the amplification, nonlinear elements, storing devices...

I wonder if we can accept without reservation your explicit formulation - "a passive element is any circuit element that cannot provide signal gain"...

You made me ask yourself why transistors have to be nonlinear to amplify and why, "if they were truly linear we would have no need for the nonlinear behavior". IMO only the 2-terminal elements (diodes) should be nonlinear (to have a negative differential resistance) to amplify...

Regards, Cyril

Quote Gregg Pulley: "Thus the traditional definition of passive -vs- active, in my thinking should really be based on the whether a device is capable of non-linear behavior. resistors, capacitors, inductors are strictly first-order linear devices."

Are they really? STRICTLY? I am not yet convinced that this definition is a good one.

Perhaps I am somewhat too "pedantic", but - for my opinion - the difference in non-linearity between a resistor and a diode is not "qualitative" but only "quantitative".

That means: Of course is the diode much more non-linear than a resistor, but is this a good criterion to discriminate linear from non-linear? What about a resistor with a relatively large temperature coefficient? And what about a thermistor? Where do you want to set the threshold between linear and non-linear?

Lutz vW

Ther are such elements like immitance inverters. They are two-ports consisting of two voltage/current controlled voltage/current sources. E.g. voltage at the input of the inverter controls current of the output current source and vice versa. When the transconductance of VCCSs is imaginary then the inverter is passive. Such passive immitance inverters are used to model e.g. transformers.

I would like also point out that the nonlinearity is not related to passivity.

Great responses all! Now we are getting somewhere with this thought experiment.

Firstly, I never meant to claim that any element that exhibits nonlinearity is capable of amplification. I only stated my belief that non-linearity "appears" to be a necessary condition for amplification. Not a proof, only a useful observation. Again, I point out that the models we use as engineers are not-verbatim reality. They only serve to solve specific design problems. Reality is vastly more interesting than what can be described by systems of linear equations!

Adam: I didn't bring up immittance inverters, gyrators, SC circuits, etc because we do not know how to construct one as a basic circuit element like a diode, transistor, vacuum tube (remember those?) , R/L/C, etc. They are strictly models. You most certainly can, given an array of basic source models and other ideal elements construct a circuit that is quite linear within a design specification using nonlinear ideal elements. I would also point out that actual real-world transformers are also either passive or can be used to amplify by utilizing nonlinear core hysteresis. The saturable reactor is such a device, as it depends upon a high value of permeability in the unsaturated region and a low value of permeability in the saturated region of the magnetization curve. In general, the greater the departure of the magnetization from linearity, the more effective is the saturable reactor in its ability to amplify.

Lutz: No not strictly. The reason I believe is that true linearity in the mathematical sense is just a convenient good approximation in the real world, but only that. There will always be a set of conditions you can subject a practical device to that will break a model. Either that or the model is so unwieldy as to be useless. For example, the resistivity of a metallic conductor under "normal" conditions is also a linear phenomenon, however at extreme low temps below the Tc for superconductivity, the resistivity becomes exponential in nature. Is your circuit ever going to need to operate below, say, 5K? If not its a fairly good bet you don't care about the nonlinearity of copper then. thermistors, which are decidedly nonlinear, are either semiconductors or ferroelectric materials that utilize deferent mechanisms than basic conduction to provide NTC or PTC behaviors. Some even switch abruptly. But these devices are not simple resistors are they? The question "where do you set the threshold?" is a good one, because it strikes at the heart of the matter. If you're looking for a pedantic definition, then this is a really tough question that no one has a good answer for. From a practical perspective though, I think it's perfectly fine to say that all devices we know how to build that can amplify signals are fundamentally non-linear devices, because it's true. Most of them, thankfully can be coaxed into linear behavior if we restrict the operating range, but on the whole they are still nonlinear devices under operating conditions encountered in 90%+ of all design problems an engineer is likely to encounter during his/her career.

Cyril - When I said "But it is the nonlinear behavior that drives the motivation for the small-signal model of the transistor. If they were truly linear we would have no need for it." I meant the small-signal model would be unnecessary. The primary justification we have for such a model is that it simplifies our calculations for things like transconductance by linearizing the large-scale behavior.

Quote Gregg A.: "From a practical perspective though, I think it's perfectly fine to say that all devices we know how to build that can amplify signals are fundamentally non-linear devices, because it's true"

Yes - I agree. However, this must be considered as a necessary condition only.

That means: All devices capable to amplify are necessarily non-linear. But to serve as a basis for a definition (active/passive) , I think it must be also a sufficient condition like: All non-linear devices are capable to amplify. Is this the case? I don`t think so.

Lutz

Dear Cyril,

Dear colleagues,

In order to answer this questions i revert to the the basic definitions in the electrical and electronic circuits.

The elements of electrical and electronic circuits are classified into sources, elements and devices. All circuit elements and devices are fed by the electrical energy from the sources.All elements and devices are considered passive except those capable to amplify signals on the cost of qualified energy sources e. g. DC power supplies. The transistors are good examples of active elements. They can amplify input signals when properly biased and operated into their active mode.The activity is a capability to amplify signals rather than to act as a source of electrical power. The device is a converter of power from a form to the other form e. g. from DC to the signal form to be amplified.Assume that we have an amplifier circuit and we remove the transistor from it, will the amplifier still capable of amplifying signals.

Concerning the change of the nature of the element and or the device to be less passive as Cyril thought, it is possible to do that by combining with them active devices. In this case we have a circuit rather than an element. Then we have to speak of circuits rather than circuit elements.

Thank you Cyril for your questions.

OK - in short:

*Active: Capability to amplify signals (thus, necessarily two-ports)

*Passive: Not capable to amplify.

Right? I agree to this classification.

Lutz vW

Lutz - Correct. I agree for now that it's a sufficient criterion only. To define further would require that we dig deeper into what specifically we mean when we say a phenomenon is "nonlinear." The order of the system, it's Lyapunov dimension, etc. would all have to be examined. We all have also been taught that in linear systems a second order system is necessary, but not sufficient, for oscillation. There are similar nonlinear criterion for chaos. I also suspect that if one cared to look into it, we would find an analogous criterion for gain given the diverse and plentiful examples we see before us.

Lutz- in response to "Ok - in short" - yes, that is what we are all taught, and it's useful too I believe.

Thank you Lutz,and thank you Gregg.

Dear colleagues (friends-:),

You have made this discussion an exciting adventure-:) Thank you!

As we continuosly come across another fundamental category - nonlinearity, I have asked a new question about its nature and the relation to the passivity:

https://www.researchgate.net/post/What_is_linear_and_what-nonlinear_Is_a_linear_resistor_static_and_a_nonlinear-dynamic_Is_a_nonlinear_resistor_a_dynamic_linear_resistor

If you allow me, I would be glad to dedicate this fundamental question to the tomorrow's (24 May ) great Bulgarian (and my-:) celebration - The Day Of Slavonic Alphabet, invented by Saints Cyril and Methodius:

http://en.wikipedia.org/wiki/Saints_Cyril_and_Methodius

Best regards, Cyril

Lutz, I remember that "amplifying" active elements can be 1-port (2-terminal).

Dear Cyril I refer to your statement

The common notion, reigning in electricity, is that passive devices (loads) dissipate energy while the active devices (sources) produce energy

As far as active devices are concerned I do not agree that active means produce energy.

An active device such as a transistor is one which allows the control of power / energy through itself. It does not and cannot produce energy but may only control / regulate the flow of power through itself between source and load.

As you have stated elsewhere the appearance that an active device such as a transistor is able to AMPLIFY is an illusion when in fact all it is doing is using a lower power control signal to control the flow of a larger amount of power provided by a separate power source through the device.

All we are simply doing is exploiting the material properties which have been masterfully designed through advanced manufacturing techniques to allow us to do this.

So much so that today we think the device itself is the power amplifier rather than being the a device through which power may be controlled to allow amplification to take place.

On a lighter note many years ago I had second year students doing a practical involving OP AMPS in which they had to first simulate then implement a small circuit. The relevant components were placed in box After a while the students approached me and said they were unable build the circuit as some components were missing from the box. I went over and saw nothing was missing and asked them what it was that they wanted. They then said that there were no triangular shaped OPAMPS like in the simulation they had done and could not find them in the box only these rectangular shaped looking devices!

Maybe it's good to finally summarize what is "active" and what is "passive"?

Gregory,

" there were no triangular shaped OPAMPS "   

I am still laughing at this surprise response. 

And, to the other point, with reference to OA and transistors ,

Indeed, the Active Device is an Active Controller of Power. 

Glen