Sure you could. However, the HFE will drop drastically, I tried this myself in discrete circuits for my own curiosity.

Hi Masoud,

I highly appreciate your curiosity; it is an exceptional phenomenon nowadays-:)

Well, we can do it... but how is it possible? I thought the emitter (like the collector) current is the output quantity and the base-emitter voltage is the input quantity of this transistor stage... and, imagine, we want to pass the input current through the output and to take the input base-emitter voltage as an output... What an idiocy-:)! Why do we do it?

Regards, Cyril.

Dear Cyril,

Again with an interesting a provocative question. In fact nothing prevent us from exchanging collector with emitter, and explore what is known as reverse active mode. A mode that was particularly useful for TTL logic gates, Unlike MOSFETs which are perfectly symmetrical device, BJT are not designed to be symmetrical. In MOSFET the notion of source or drain terminal, only makes sense when you imagine the device properly biased, until then, it is not possible to distinguish one terminal from the other. In BJTs, the collector has larger area than the emitter, this makes the device unsymmetrical. When you use the reverse active mode, BC is direct biased, while BE is reverse biased. For this strange situation, what happens is that the transistor will operate normally, but the current gain is reduced. This is a common error with students doing their first steps into the fascinating world of transistors.

However, as pointed out in your remark to Masoud, this is not the answer to your original question. Again the interest is to reverse the input and output quantities, which is a very different matter. I see several circuit employing this concept, linear voltage to current conversion, where the emitter current is set by series-series feedback involving the the BE and the emitter terminal. Differential amplifiers which rely on the relation between the emitter current of a emitter coupled pair and the BE voltages of the pair. The same concept can also be found in translinear circuits. So the concept is not only interesting but also very useful in circuit design.

Regards

Luis Nero Alves

The attached image shows the basic topology of a practical circuit that will drive a current into the emitter of a transistor. With R1 = 1Kohm, the operational amplifier will ensure that the emitter current is 1mA per volt of applied input voltage. It drives the base of the transistor to the corresponding Vbe, which in this case will appear negative with respect to ground. (One practical issue: if you apply a negative voltage to the input, the operational amplifier will drive the base-emitter junction into reverse breakdown - to ensure that this cannot happen, you could run the operational amplifier from limited supply voltages such as +/-5 volts.)

Why do it? I'm not sure! There are some mathematical functions that can be implemented in analogue circuits based on the logarithmic relationship of collector current to Vbe. One example is analogue circuits used to generate a RMS value for an alternating (but not necessarily sinusoidal) input voltage.

Yes, Luis, yes... I do not mean the BJT reverse active mode since there is no some eccentric circuit idea there-:) As you have guessed (as I can see, you have already a very good notion about my way of thinking and can predict my intentions), I have ment a clever phenomenon that can be seen everywhere around us in our routine. Particularly, as you have noted in the second part of your comment, where I have seen a few ingenious speculations, we can see this phenomenon in elextronics as well...

Thank you for the so interesting insertion! Regards, Cyril.

"Mamma mia" (I would say if I was an Italian-:), how wonderful the circuitry is! I like these clever circuits behaving like people... I love them! I admire the circuit geniuses invented them... I bow down before these minds...

Daivd, thank you for the excellent example. Obviously, this is a more sophisticated op-amp version of the classic emitter-biased transistor stage or the cascode circuit (see the attachment below). Let's compare all these famous circuit solutions to see what the transistor and the op-amp do.

The classic "only transistor" version exploits the lucky chance that the emitter is a "common-born" point for the input and the output voltages (emitter follower). So a series-series negative feedback (Luis) naturally exists and we do not need to make any additional connection to introduce it. The situation is very interesting and deserves a special separate question. Shortly, the input current source (Q2) tries to pass the desired current through the emitter-collector junction and the transistor (Q1) "helps" it by adjusting its base-emitter voltage so that to conduct this current. This simple configuration does not ensure perfect load conditions (short connection) for the input current source but it is still good (in the most cases, Vbe can be ignored). The output emitter voltage is below the Vb (if the base is grounded, it is negative)...

Now, about the Daivd's op-amp version (for convenience, I have attached it again)...

Here the input voltage source "blows" a current Vin/R1 into the Q1 emitter trying to "pull up" the emitter voltage. As above, the transistor begins increasing its base-emitter voltage so that to pass the desired current. But now, in addition, the op-amp "moves down" the base voltage below the ground (zero) potential until the emitter voltage becomes zero (virtual ground). As a result, the input current source (Vin-R1) operates at perfect load conditions (short connection). An additional advantage is that the input current does not enter the op-amp input where the only base current flows. BTW this allows to use the collector as an output of a perfect and buffered current source (I have seen this application somewhere). Here the output base voltage is negative...

And finally, let's consider the classic "transdiode". It uses the same negative feedback idea but here the input current is "pushed" in the collector; see the dual question:

https://www.researchgate.net/post/Can_we_reverse_a_BJT_by_injecting_the_input_current_into_the_collector_and_taking_the_base-emitter_voltage_as_an_output

But there is a big difference between the operation of this circuit and the other two above. Here the transistor cannot "feel" the input variatons since the input quantity (the current) is applied at (through) the collector (a BJT can react only to the voltage and current applied to/through the base-emitter junction). So, here the op-amp "feels" the input quantity by "observing" and keeping up the virtual ground point (the collector voltage) by "moving down" the emitter. Here the emitter voltage is negative and the input current enters the op-amp input...

Maybe, for completeness, it would be useful to remember the simple transistor version of the idea above (injecting a current into the collector). Here, to "sense" the input quantity (the current, accordingly the collector voltage), the Q1 base is connected to its collector. This allows the transistor to adjust its base-emitter voltage so that to sink the desired current. There is a negligible error caused by the Vce voltage acting as a positive output voltage...

Regards, Cyril.

Hi Cyril...I am from an older generation!....the comment about the non-symmetric structure is true. However, connection-wise it is still like two diodes connected at anodes(cathodes). The asymmetry allows for high current gain which means if you replace collector and emitter it may still work but at a point of disadvantage!

Please do not get confused with use of PNP in negative logic; in all the images within this discussion, the biases are all correct and proper for both P- and N- types.

Nice talking to you.

Masoud

Thank you, Masoid! Note that the load is not shown in this picture but, of course, it exists. Imagine it is simply a resistor. Then can we say that we have as many as four cascaded converters?

Regards, Cyril.

Cyril, thanks for this interesting discussion, comparing a variety of circuits that translate a current to a corresponding Vbe in a transistor. I believe that the best-known circuit designers (here, I am thinking of the designers who pioneered analogue integrated circuits in the earlier 'silicon valley' companies) have always cultivated an intuitive appreciation of the operation of various circuits, together with a good grasp of formal methods of analysis.

It's interesting to contrast the two op-amp circuits (1) which I described, and (2) the classic 'transdiode' circuit. (1) generates the Vbe that corresponds with the EMITTER current whereas (2) generates the Vbe that corresponds with the COLLECTOR current. It may seem at first glance that there is no difference, but there is a subtle difference that is important: the relationship between Vbe and the collector current approximates the ideal logarithmic relationship much more closely than does the relationship between Vbe and the emitter current. The reasons for this are a matter of transistor physics, but I'd appreciate comments from anyone better informed than I am about this.

Hence, (2) has been used quite widely to provide a logarithmic function in analogue computation application; but I have never seen (1) used for analogue computation (it may have been used as a current source, yes).

There are practical difficulties with using the Vbe potential in circuits, because Vbe depends on viables other than the collector current, such as: i) temperature, ii) collector emitter voltage (Early effect), and iii) saturation current (a manufacturing-process-dependant parameter). These difficulties are usually addressed by the use of transistors fabricated on the same piece of silicon, largely eliminating temperature and saturation current differences. Integrated circuits, in which inherently well-matched transistors are available, made translinear circuits a practical reality.

Hi, Daivd! It is so beautiful there are people as you trying to penetrate deeply into circuit phenomena!

Yes, I think exactly as you about the way of thinking (creating) of the pioneer circuit designers in 60's. There many legends about them, some of them quite funny... The problem is that these circuit geniuses, because of various reasons, have not revealed what we, educators and our students, need nowadays - the way of thinking when creating these famous circuit solutions. For example, try to understand something from the Widlar's articles in NI application notes and you will see what I mean. And we have only to guess at their intentions. I will illustrate this procedure by demystifying completely your op-amp circuit (2).

You made me think deeply about it and derive at unexpected conclusions. After one sleepless night-:), I realized it is just a buffered (by Q1) inverting amplifier without R2 (maybe, the circuit designer has omitted to draw it-:) where the op-amp compensates the "undesired" base-emitter voltage drop Vbe with the purpose to keep the virtual ground. It does this magic by adding the same voltage in series to Vbe so that the result is zero (virtual ground). We can think of this circuit as of an op-amp version of the transistor stage with parallel-parallel negative feedback thoroughly considered in the question below:

https://www.researchgate.net/post/Can_we_reverse_a_BJT_by_injecting_the_input_current_into_the_collector_and_taking_the_base-emitter_voltage_as_an_output

The clever trick here is that we use this "mirror" compensating voltage Vout as an output quantity instead the original Vbe voltage (all the op-amp inverting circuits exploit this idea). I have thoroughly considered this technique in the Wikibooks article below:

http://en.wikibooks.org/wiki/Circuit_Idea/Voltage_Compensation

http://en.wikibooks.org/wiki/Talk:Circuit_Idea/Voltage_Compensation

If we use it as a current source, maybe it is better to ground the input and to apply the input voltage to the non-inverting input. Then this circuit will be just a buffered non-inverting amplifier. In this case, we can ground the negative supply terminal and the circuit becomes a single-supplied. We can connect the load into the emitter (the current will be exactly Uin/R but the load will be ungrounded) or into the collector (it will differ with Ib but will be grounded)... the choice is yours-:)

In 2002, I was fascinated by some Tom Hayes's (Student Manual for the Art of electronics) speculations about some unique negative feedback properties. Then I created and dedicated to him the funny interactive Flash movie "Strange things can be put in feedback loop". You can enjoy it by clicking on the transistor from the library on the left; the transistor will "fly" over the circuit and "land" on the op-amp output-:) Then you can take the output after (hide the dog) or before (show the dog) the transistor "disturbance":

http://www.circuit-fantasia.com/collections/tools/strange-things.html

Best regards, Cyril.

Hi Cyril, when I sketched the op-amp circuit (2), I did not anticipate that it would deprive you of sleep :) I was just thinking 'how can I make an emitter current to Vbe converter that has ground-referred inputs and outputs?' and played around with the op-amp and transistor.

As you may be aware, there are a number of famous circuit configurations that have been arrived at (in a more or less serendipitous way) by designers who were just 'playing' with circuit elements and asking themselves 'what will happen if I join this little circuit to that little circuit?' Examples I can think of where serendipity played a part are the Gilbert multiplier cell, the Wilson mirror, and Hans Camenzind's 555 timer chip. Of course, these examples are very useful circuits - I do not think that the emitter current to Vbe converter will be so useful!

I agree with your insight that the op-amp circuit (2) is fundamentally an inverting amplifier. Here, the feedback resistor normally connected between the op-amp's output and inverting input is replaced by the transistor's base-emitter junction.

Regarding the circuit geniuses of the 60s and 70s - actually, a few of them have revealed their way of thinking when creating circuits. The late Jim Williams (of Linear Technology) edited two books, 'Analog Circuit Design - Art, Science, and Personalities', and 'The Art and Science of Analog Circuit Design', in which some of the world's foremost circuit designers have described their thinking, their approach, their methods, their philosophy in creating circuit designs. You may already know these books, of course - if not, I think you will find much illuminating, inspiring, and entertaining reading in them.

Best wishes with all your work in teaching and encouraging tomorrow's electronic engineers . . .

Daivd, thank you for your words pregnant with meaning... I continue thinking about all these problems (including the simple current mirror) but I am too booked this week with preparing some materials for my students.

You have evoked my interest in the long-tailed pair (the differential amplifier) and particularly, in Gilbert's multiplying ideas. I have the remarkable Jim Williams book, "Analog Circuit Design" and I have examined some of the papers, particularly with the Hornak's exciting story (True Analog Circuit Design) and the extremely interesting for me Sheingold's story about the op-amps. The Gilbert's "Where Do Little Circuits Come From?" is too sophisticated (including the style) and I will scrutinize closely it to realize the truth about his cell. In addition, he has said extremely interesting "philosophical" and prophetic thoughts as, "...the conception of fundamental new topologies will always be a highly heuristic process, which relies far more on lateral thinking (de Bono) than on a mechanized approach..."

Best regards, Cyril.