How to plot the I-V curve of a tunnel diode?

My answer to Codidact question How to plot the I-V curve of a tunnel diode?


I wrote this post yesterday. Initially, I did not intend to write an answer but rather wanted to give some guidance to OP through comments. But it turned out this was not desirable there and was treated as a "discussion"... and our comments have been deleted.

My answer is short. My goal is to use it to explain, in the simplest and most convincing way possible, the behavior of the tunnel diode in the negative resistance region... what it does there...

Answer

Setup. We can best understand the behavior of the tunnel diode in the region with negative resistance if we imagine it as a self-variable (dynamic) resistor R driven by a variable voltage source V - Fig. 1. If OP has a subtle sense of humor, I suggest we conduct this experiment in the form of a fun (but useful) game where he is the voltage source and I am the "tunnel diode":)

Fig. 1. A setup for measuring the tunnel diode IV curve in the negative resistance region
Fig. 1. A setup for measuring the tunnel diode IV curve in the negative resistance region (Wikibooks)

Graphical representation. The voltage VA across and the current IA through the two elements are the same. So their IV curves can be superimposed on the same coordinate system - Fig. 2. The IV curve of the variable resistor is a straight line (in orange) beginning from the coordinate origin and having a slope depending on the instant (static) resistance R. The IV curve of the input voltage source is a vertical line (in red) shifted to the right from the Y-axis. The intersection point A (aka operating point) represents the instant magnitudes of the current IA and the voltage VA.

Fig. 2. The graphical representation of the circuit operation as two superimposed IV curves
Fig. 2. The graphical representation of the circuit operation as two superimposed IV curves (Wikibooks)

Operation. Let's now see what this "tunnel diode" (I) does when the voltage source (OP) varies its voltage:

1. Low positive resistance region. When OP gradually increases the input voltage from zero to the beginning of the negative resistance region, I keep a low constant resistance R... and OP (i.e., the voltage source) imagines it as a steep line (IV curve).

2. Negative resistance region. When reaching the negative resistance region, I decide to play a trick with OP and begin increasing R vigorously while OP gradually increases the input voltage. As a result, in Ohm's law I = V/R, both V and R change in opposite directions and with different rates... so the current decreases and OP sees a negative resistance. In Fig. 2, my (R, tunnel diode) IV curve begins rotating clockwise; the operating point A moves down and pictures the negative resistance part of the tunnel diode N-shaped IV curve.

3. High positive resistance region. After the negative resistance region, OP continues increasing the input voltage. My resistor possesses high positive resistance and OP imagines it as a sloping IV curve.

More considerations

Let's see what is the most important to measure the tunnel diode IV curve. IMO this is not the scope. You have to see that when increasing the voltage across the diode, the current through it decreases. For this purpose, first at all, you have to remove (destroy, neutralize) any resistance in series. Here is my philosophy about how we can do it...

The problem. What is this resistance? First, this is the internal resistance of the voltage source and next, this is the ammeter resistance (as you provably now, VOM measure the current by measuring the voltage drop across a small resistance). How do we do this magic? The usual answer is, "It is very simple, just apply a negative feedback". Yes, but you probably are not satisfied with ready-made formal explanations and you want to understand the idea behind all this. Here is what it is ...

Solution. To zero a resistance actually means to zero the voltage drop across it (V = I.R = 0). The natural way is to replace the existing resistance with a "piece of wire"; then really V = I.0 = 0. Only, we cannot do it in this way. However, since we are inventive enough, we decide to do it in an artificial way - by adding a voltage equal to the voltage drop in a series manner, according to KVL. As a result, the total voltage across this network will be zero... as though the resistance is zero. Wonderful, isn't it?

We can add the compensating voltage in two ways:

1. Increasing VIN with VR. First, we can make the input voltage source to raise its voltage with the value of the voltage drop. We can do it by applying a series negative feedback to an op-amp and putting (hiding) the ammeter resistance inside the feedback loop. This idea is implemented in Olin's and coquelicot‭'s answers. A disadvantage of this solution is that the ammeter (ADC) is floating. Also, you can want to test another device by current; then the device will be floating.

2. Adding VR to VIN. The second idea is just amazing - instead to increase the input voltage with VR, we decide to add the compensating voltage to the input voltage. This means to connect an additional voltage source in series to the diode so that its voltage adds to the input voltage according to KVL. For this purpose, its voltage must be negative in regards to ground (you can see for yourself if you travel along the loop). The sum of the compensating voltage and the voltage drop across R is zero (VR - VR = 0) and the so-called _virtual ground_ appears.

We can explain this technique in terms of resistances. The op-amp output can be considered as a "negative resistor" with equivalent but negative resistance -R that is added to the positive resistance R. The result is zero resistance.

Implementation. Thus you have two possibilities to artificially zero the undesired resistance - by a non-inverting configuration or by an inverting one. I prefer to use the second; this old setup from 90s is implemented this way (here is a movie showing how a rectifying diode can be investigated by voltage). It is actually an inverting amplifier with buffered input and output. For your purposes, you have to connect the tunnel diode in the place of R1 and the ammeter (VOM or movement) in the place of R2. If you want to measure the current by a grounded ADC or a microcontroller port, then use the op-amp output voltage as a measure of the current. The only problem is that it is negative (I = -Vout/R).

See also:

Demystifying the Negative Differential Resistance Phenomenon (Circuit Idea wikibook)

My answer to the SE EE question, "Art of Electronics - Zener Diode Example" (the same explanation about the dynamic resistance but in the case of a Zener diode)

Voltage compensation (a Wikibooks story about the philosophy behind op-amp inverting circuits)

Comments

  1. I understand what you mean by "The IV curve of the variable resistor is a straight line (in orange) beginning from the coordinate origin and having a slope depending on the instant (static) resistance R". For newbies, I would say the following: "For a particular linear (static) resistor obeying Ohm's Law, the relationship of current I, a dependent variable, depends on the independent variable V can be shown as a straight line originating from 0, 0 of the graph. For many resistors with different resistance (R = V/I) values, we see a bunch of such straight line starting from the V, I origin 0. 0.

    ReplyDelete
  2. This comment has been removed by the author.

    ReplyDelete
    Replies
    1. Ha, adding the moving potentiometer wiper bit makes thing much more clear. Perhaps you should add a picture of the pot, with a hand moving it, .. :)

      Delete
  3. This is very confusing: "The IV curve of the input voltage source is a vertical line (in red) shifted to the right from the Y-axis." for the following reason: If I-V curve denotes the I depending on V relationship, then how come the vertical lines denote many voltage sources. Or perhaps I was misled in the very first beginning. Or perhaps I should interpret your description as "A variable resistor can be represented by a bunch of orange lines, different resistances with different slopes." But then do you mean the red vertical line represent different voltages of the same voltage source? This new one-variable-resistor, one-variable-voltage-source seems to make more sense. But I still find it confusing. Perhaps another presentation using moving pictures by animated GIF or Flash/HTML5, showing a time sequence. Of course an oscilloscope display in slow motion is even better. Actually this is what we can do using your 50Hz sine signal to the tunnel diode and display by a scope. The only thing I wish to do is to show the events happening in the NDR region, ...

    ReplyDelete
  4. Let's try to say it even more simply to the newbie: "You should already know Ohm's law, right? Then you know that the IV curve of a resistor is a straight line starting from the origin of the coordinate system and its slope depends on the resistance R. So, in our case, when you move the potentiometer wiper, its IV curve will rotate around the origin (clockwise when you increase the resistance and v.v.)

    ReplyDelete
  5. Ah, need to go now, see you later. Nice chatting with you. Cheers.

    ReplyDelete
  6. OK, I will explain it... IV curves of elements behaving as voltage stabilizers (voltage sources, diodes, circuits of voltage stabilizers, etc.) are measured by varying the current and measuring the voltage. So, in the case of a variable voltage source, its IV curve will move horizontally (translates) when its voltage varies. I have Flash movies where this is animated (e.g., https://www.circuit-fantasia.com/my-students/ske2004/classes/class1/v-to-i-old-page6-3.html). But how do I show it on a sheet of paper? Of course, as a family of IV curves... a bunch of red vertical lines... Look how I had drawn it in 2004 by the help of the Corel Draw...

    ReplyDelete
  7. Here is an interesting laboratory experiment from the early 90s, where the diode IV curve is measured and plotted on the screen by a computer-based curve tracer:

    https://photos.app.goo.gl/qobqm7Fwp3fTEA4o7

    Since it is created not by a "humble engineer":) but, in addition, a teacher, inventor, thinker, dreamer, fantasist:)... it shows not only the final result - the diode IV curve, but also the way how it is obtained. The computer and its AD periphery serves as a variable voltage source whose IV curve (vertical line) is drawn on the screen. When I vary the voltage (by the arrows on the keybord), this line moves left or right and the operating point draws the diode IV curve on the screen. Isn't it great? Moreover, it was made 30 years ago... How much -1s would I get in SE EE if I showed it there? I think at least a dozen:)

    ReplyDelete
    Replies
    1. Well, I would give 8/10 marks to this computer aided diode curve tracer experiment report if published 30 years ago in a BBS (Bulletin Board System). Actually it was in early 1990''s when I was working in a university research project on CIS (Clinical Info System), and I setup up in an IBM286PC (sorry, funny characters system bug), a BBS, using a BBS system program called Remote BBS. At that time I had several phone lines and use 9k6 baud modem for communication and sharing software (linux 0.9 need 12+ 3,5" 1.44MB floppies. In those years a 170MB hard disk was very expensive (HK$1,700) . And in the 1980's 40MB hard disk costed HK$4,000.

      But this afternoon I went to the local computer centre and bought a SamSung T7, 1TB SSD for only HK$1,250.

      So it is in this historical background that I think your curve tracer should got 8/10 marks.

      Delete
    2. The funny characters seem disappeared. So let me continue. So let us fast forward from 1990 to 2020 now. We still do diode curve tracing, but the hardware is completely different. Let me go though my suggested versions.

      (1) Version 1 - Voltage source is a manual button adjustable in mV steps of a digital PSU, form 0V to 1V. Measuring meters are cheapy multi-meters.

      (2)Version 2 sorry, funny chars again.Need to take a break.Cheers. 

      Delete
    3. Version 2 - Using programmable triangular wave signal generator as the voltage source to power the (tunnel) diode, ...

      Version 3 - Using AD5933 Impedance Converter - This is what I am exploring now. AD5933 can frequency sweep an impedance and return two integers, real and imaginary part of the impedance. In other words, we can measure the resistance and inductance of a (tunnel) diode.

      I want to measure inductance at any particular frequency, becuse I want to know

      (1) which frequency the tunnel diode would oscillator,

      (2) which (external) inductance would force the tunnel diode to crawl down the slope after the peak, instead to jumping over. (According to the GE TD 1962 User Guide, these parameters need careful tests to find.)

      Delete
  8. This comment has been removed by the author.

    ReplyDelete
  9. Well, once we started this topic, let's move on to the other way of investigating elements - by varying the current and measuring the voltage. As I said above, it is more suitable for testing diodes. Here is a video of such an experiment from the distant past:

    https://photos.app.goo.gl/Rng6drGK63s1BjhN8

    Now the computer (Apple II manufactured in 80's in Bulgaria under the name "Pravetz") and its AD periphery serves as a variable current source whose IV curve (horizontal line) moves up and down when I vary the current by the arrows on the keybord.

    ReplyDelete
    Replies
    1. Ah, this diode voltage Vd vs diode current Id of testing diode characteristics is interesting. I guess you can do this curve plotting in two ways: software and hardware.

      (1) To do it in software we know the diode equation Id = f(Vd) , and so Vd = g(Id) where function g( ) is the inverse of f( ),

      (2) To do it in hardware, we can use a variable current source, controlled by hitting one key, say "i" to increase current magnitude, and another key "d" to decrease current magnitude.

      I agree this can be done using Apple II and associated hardware, current source, voltage and current meter.

      Delete
  10. TL Fong, I wonder why there are so few normal people like you and me and there are so many idiots on the web ... I think that Q&A sites are one of the reasons for the idiocy of people because the normal form of communication between people is not the monologue (asking and answering questions) but rather the dialogue (discussion)... ie, what we both do here... However, let us not forget the fact that they are very popular and what we write there becomes available to many more people than here. So I suggest you exchange some of these valuable thoughts also there under my answer to your question. Olin should not react negatively to this because it is directly related to the topic and is not written under someone else's answer.

    ReplyDelete
    Replies
    1. Well, I already told you that ALL the answers to my question in CD are not answering my heavily edited question, which I pointed out my question is why I cannot plot the curve in the NDR region. One answer using 25 words and a picture showing that it can be done using simulation, but that simulation is for a negative resistance of a transistor circuit, NOT even a tunnel diode. Another answer very briefly describes how to use LM317 adj volt reg to provide the variable voltage source for the diode, but no details, and that what is described is Gunn diode, which is not tunnel diode. Another answer in very detail suggest how to use op amp for precise measurement, which is off the point, and not what I asked. Another answer describes how to plot the graph, but again does NOT answer my question of he missing NDR. If my long answer has not been heavily edited, perhaps everybody can see what I am asking, is not at all the general plotting method, but the missing Twilight Zone.

      If my question can be further expanded, I would suggested that the question have already been answered in GE,sorry funny char, will come back later,

      Delete
    2. Continued from above incomplete reply with funny characters. I was saying that GE TD 62 Manual already answer my question about the NDR Twilight Zone, but I think a complete answer should include how to plot the NDR region, with specific parameter of external R and L components to force It to crawl down the neg negative slope, and also who to design the load line for oscillation (I already measured the oscillation It is approx 1.8mA, for 2SB4b).

      I don't think it a good idea to comment on your answer in CD. As you see, all my references to your image, wiki books, in my long answer before being heavily edited by some elite have already been "knocked out". So I don't wish to waste my time any more, because the whole question might be closed, sooner or later.

      But I am happy to comment on you TD curve plotting here, to clarify the meanings of the slant current and vertical voltage lines, and perhaps update the picture with more elaboration with numbers. That is way I am exploring ICL8038, AD5933, and PCA8951 to update your 30 years old invention using Apple II (and Turbo Pascal?) and hand knob adj pot, ...

      Delete
    3. The link about GE TD 62 Guide's answer to my NDR plotting question is here: https://tunneldiode.blogspot.com/2020/12/tunnel-diode-twilight-zone-problem.html.

      I will continue reporting my progress on experimenting with ICL8038, AD5933, and PCA8951 there. Stay tuned.

      Delete
  11. From what you have written above, I understand that you have many years of experience in complex computer systems and that it is very important for you that the implementation is virtuoso. You pay special attention to the form and the means ... you want them to be modern so that you get a product with a sales value. I, on the contrary, am interested in the basic idea in its most general form, which does not depend on the concrete implementation and therefore does not die ... it is eternal and immortal. I guess this is a popular view of the world in the Eastern teachings and philosophies you mentioned to me ...

    So I want, when discussing, to distinguish the idea (1) from its implementation (2) and the measuring means for its investigation (3)...

    ReplyDelete
  12. ... not to oppose them to each other. Also, I want to distinguish between the functional notion about the element (NDR) and the notion about the specific element (tunnel diode).The idea is the most important to me and least important to most people. The vast majority of people (including the "elite" of Q&A sites) think specifically ... "they see the trees but do not see the forest" ...

    So, as you can see in my answer to your question in CD, actually I do not explain the behavior of the tunnel diode; I explain the behavior of the N-shaped negative differential resistor. That is why, I can replace it by a humble potentiometer (rheostat) and show how it operates. This is possible sinve there is a common idea between them...

    ReplyDelete

Post a Comment