Circuit Idea/Simplest Transistor Current Source

<<< contents - page stage 100% developed  as of Feb 23,2009 >>>

Building the Simplest Transistor Current Source

Circuit idea: Using a Bipolar Junction Transistor (BJT) to create a "bottleneck" for the current flow in a circuit branch.

What a current source is edit

 
Fig. 1: How does a constant current source behave?

We begin our story with general questions: "What is a constant current source? What do we want it to do? How does a current source behave? How do we figure out that some device is a current source? How do we make a current source? Do we need a "definition" for such a self-speaking term?" Obviously not; we can take for granted the notion of constant current source. Nevertheless let's mention that it's a thing (electronic device) producing electrical current. For such a simple device, we would expect it to have a simple behaviour. We would very much like it to produce a constant current, regardless of the load we apply.

What have you known about the current source from the basic course of electricity? Here is the well-known symbol of a current source (Fig. 1). How to make a good guess that it is a current source? If someone gives us a (black) box with two leads and says that it is a current source, how do we convince ourselves that it is really a current source? Let's conduct some experiments. For example, what will happen, if we connect the two leads (a short circuit - case 1 on the picture)? Is it dangerous for the current source (comparing with a voltage one)? Indeed, there is nothing dangerous for a current source, if we short its two leads. As the current source is designed for a given current, it keeps up (limits) the current thus protecting itself.

Then, maybe the opposite condition - an open circuit (case 3 on the picture) - will be dangerous as the current source will try to pass the same current as before. In this situation, the current source is misled: it doesn't see the open circuit; it "thinks" there is something connected there. It just "sees" that no current flows and raises its internal voltage up with the "hope" to pass a current through the harmful "thing". Of course, the current source does not manage to pass any current through the open circuit and reaches the final (compliance) voltage; a saturation and even a breakdown occurs. Thus, the compliance voltage represents the maximum voltage that the current source can reach when it strives to produce the desired current.

A passive current source edit

 
Fig. 2 : Making the simplest current source

The general idea edit

How do we make the simplest current source? As we all know, thanks to a certain wise man called Ohm, the only way to get yourself some current is to apply some voltage over some resistance. The easiest way to achieve this is to connect a voltage source in series with a resistor.

 

For example, take a standard 9 V battery and a 9 kΩ resistor, connect them in series - and there you have it, a 1 mA current source (you can measure it with an ammeter).

I = 9 V / 9 kΩ = 1 mA

OK, we all know Ohm's law as a mathematical equation... but what's the logic behind it? We know the fact that voltage over resistance produces current, but why? Voltage is in fact a genuine Force; another possible explanation would be to think of voltage as some kind of potential - like a smart man's potential to learn many things. In order for this potential to become something real, to move the smart man up the line of personal improvement, (this movement would be a kind of current) it has to be put in action over some obstacles - exams, tasks, etc. (this would be the equivalent of resistance). Another analogy would be the potential of a car's engine to propel the vehicle forward - there must be something to act upon - the road surface, for example.

Keeping a constant current by using an ideal current load
Passive voltage-to-current converter (voltage causes current)

Imperfections edit

 
Fig. 3: The load decreases the current.

But enough philosophy... now that we have a current source, let's use it to power up something - let's connect a load (it could be a light bulb, a resistor, anything consuming current). But now a problem appears... the ammeter shows less current, than we expected. We may look at the problem from two viewpoints.

...in terms of resistance... Why - because the load introduces some new resistance to the circuit, thus rising its equivalent resistance and lowering the current (note the voltage supply stays constant). If we put a higher load, we would see that the current decreases even more: IOUT = VIN/(R + RL).

...in terms of voltage. The real load introduces some voltage drop VL, which affects the excitation voltage VIN. Now, not all the input voltage is applied across the current-setting resistor R but only the voltage difference VIN - VL. In other words, here the voltage difference VIN - VL determines the current IOUT instead the voltage VIN. As a result, the current decreases: IOUT = (VIN - VL)/R.

It appears that our simple current source can't cope with various loads and is unable to keep the current at a constant rate. That's why these kind of circuits are called passive - they don't react to changes and disturbances. So, the simple current source that we've made ourselves is a passive one (it's like a reluctant worker, who doesn't alter his efforts regardless of how much work there is to be done). Well, this passive circuit is imperfect; it can't stand against the load "intervention", especially if the load varies. It is a static, fixed, non-adaptable circuit...; it needs some improvement... What do we do then?

Keeping a constant current by depreciating the load
Passive voltage-to-current converter (voltage-controlled current source)
Op-amp circuit builder (go to Stage 2 of this interactive flash movie)

Two ways to make an active current source edit

The basic "dynamizing" idea edit

Obviously, we have to reveal the truth by ourselves; let's then begin reasoning. The problem pops up when we connect a load and, what is worse, if we begin varying its resistance. Note the load may be not only a steady or varying resistor; it might be a charging capacitor, a diode (ordinary, zener, LED, base-emitter junction, etc.) or even a voltage source (e.g., a rechargeable battery)... It is not so important what exactly the load is; it is only important that a voltage drop appears across the load and this voltage drop confuses things.

If we look around, we can see that most things in this world are changeable, adaptable, dynamic... According to this observation, we have to make our current source react somehow to the load "intervention", to resist to its attempt to lower the current. That means something inside the current source has to change, in order to compensate the disturbing voltage drop across the load. What can change in this simple electric circuit?

In our recipe for current we have two ingredients - voltage and resistance. Hence, if we want to make our current source react to changes in load, we have to alter one of the two ingredients dynamically, in relevance with the change in load resistance.

Making the internal voltage dynamic edit

 
Fig. 4: Keeping up a constant current by a dynamic voltage source.

One possible solution would be to vary the internal voltage - when the load increases, the internal voltage rises too. We can call this one dynamic current source with following internal voltage.

It may seem like a smart idea, but unfortunately - it is not. It's actually a kind of bootstrapping - this term comes from the story about Baron Munchhausen, who once got stuck in quicksand, and pulled himself up by his own bootstraps.

This solution has quite a limited ability to compensate changes in load, because the varying supply voltage will quickly reach its maximum.

Making the internal resistance dynamic edit

 
Keeping up a constant current by a dynamic resistor.

The other possible solution is to vary the internal resistance. This time, when the load resistance increases, the internal resistance R will have to decrease, keeping the sum RL+Rint unchanged, so the current flow stays at a constant rate. Hence the name of the resistor R is current-stable resistor.

Of course, this method also has a limited ability to compensate changes in load resistance, because the decreasing internal resistance will reach zero at a certain time. But it's much more simple to implement (it's easier to play with resistance, then messing around with voltage).

We will continue our quest to destination Simplest transistor current source by using the second methodology. What element will we use to materialize the dynamic internal resistance solution? As the name of this topic implies, it's going to be the transistor. Let's talk about its behavior to find out why...

Active current source with a BJT edit

What is the transistor and why did we choose it edit

 
Transistors

Enter the transistor - considered one of the greatest inventions of the 20th century. A tiny device, that literally changed the world and now dominates all areas of life - it can be found in every electronic device around us, as it is the fundamental building block of modern electronic circuits.

OK, the transistor is pretty cool... but what is it actually? Its name comes from transfer resistor - transistor... which doesn't clarify its structure, neither its purpose. To keep it short, the transistor is a semiconductor device comprising two PN junctions and has three terminals. Let's talk a bit more about its behaviour, as it is much more interesting ;)

It can do a fascinating trick - amplifying a signal. At least that's how they call it, even though it's not amplification at all. Actually, the transistor allows us to control a larger current flow (the output) with a smaller current flow (the input). It's more like creating a scaled copy of the input signal using additional power, than amplifying it. Think of it as a current valve - valves are devices used to control the flow of a fluid, in our case the current would be the equivalent to fluid.

Transistor are also used for electronically controlled switches - just like ON/OFF buttons, operated by electronic means. This powerful idea is the heart of digital circuitry today.

 
An NPN Bipolar Junction Transistor

For a wonderful and detailed explanation of this transistor behaviour by means of hydraulic analogy - check the References/See also section.

Transistors come in a variety of types, with the most popular being MOSFETs and BJTs, but the list continues with lots of others (basically derivatives of the two most popular ones). We will use a BJT, which stands for Bipolar Junction Transistor. "Bipolar" indicates that both the electrons and holes take part in the transistor's operation.

Much more can be said about this wonderful invention, but that's a long story outside the scope of this article.

Hydraulic analogy - the BJT as a current valve edit

 
Valve - a variable water bottleneck

As we saw in some of the other articles of Circuit Idea, hydraulic analogy is a powerful and intuitive concept for explaining electrical circuits.

Current flow (hence the word "flow") can be represented by the flow of a fluid. Usually, when we want some fluid flowing at a constant rate, (e.g. when watering the garden with a hose), we create a constraint on the fluid path, creating a bottleneck.

 
Potentiometer - a mechanically controlled current bottleneck

A bottleneck (consider the name) is the point that has the lowest throughput, thus determining the maximum flow capacity. Bottlenecks can be found not only in hydraulic stuff, but in virtually any system imaginable. In hydraulic terms, the most simple variable element for creating a bottleneck is the valve.

In the electrical world, the closest analog to the valve is the variable resistor - also called a potentiometer.

Well, the transistor can behave as a variable resistor too, but unlike the mechanically controlled potentiometer, it's an electrically controlled one - the output collector-emitter(C-E) current can be controlled by the input base current or the base voltage, whichever suits us more. So here comes the simple idea - if we apply a constant base voltage, we will set a constant C-E current. When the C-E voltage increases (e.g. due to decrease in voltage on the load resistor Rl) the C-E resistance will increase too (remember we set the C-E current to constant using a constant base voltage). So what do we get - a dynamic varying resistor which keeps a constant current flow. In short - a current-stable resistor. Which is exactly what we need in order to make a constant current source.

Sprinkle that with theory... edit

 
IV curves (both RL and RT vary).
A nice and clear scheme to be added
 
Transistor - an electronically controlled current bottleneck

This behaviour is shown on the IV diagram. They name it output characteristic, because it represents the processes occurring at the output part of the transistor depending on changes in current or voltage. We can explain it by some ugly complex formulas found in thick textbooks on semiconductors, but we would rather think of it in a more user-friendly manner.

Imagine it as a mechanism with the I line acting as a "fixed rail" (representing the current which appears to be a constant), on which the crossing of the two "pivots" - Rl and Rt lines - point A slides. As we decrease the load resistance (Rl that is) point A slides down the Rl line, up the Rt line, and to the right on the I line. This means that the transistor resistance Rt increases, so the whole resistance of the circuit remains the same, and therefore the current flow remains the same. As we do the opposite - guess what - exactly the opposite things happen ;)

The solution - based on team work edit

 
Building the simplest BJT current source on the whiteboard.
A nice and clear scheme to be added

As the saying goes, two heads are better then one. More than two are even more ;) The point here is, that great things in life, no matter how simple, are more often produced by a team effort, rather than a personal one.

So is the case with our solution. Let's present our team...

In the middle, surrounded by a blue line is the "boss" of the circuit - the varying load resistor Rl. It's he who leads the concert, and the whole circuit is built for him - to satisfy his needs and desires for constant current. The nameless resistor just beneath the boss is his "bodyguard" - just in case someone curious moves the slider of the "boss resistor" to its very end and try to short-circuit the power source through the transistor.

Everything outside the blue line is the BJT current source itself - a whole team of elements working for the "boss". There's the "hero" of the circuit - the workhorse - the mighty bipolar junction transistor (BJT), who does all the dirty job (donkey work :) of keeping the constant current. This is our bottleneck point, which determines how much current can pass through. Connected to its base is his "manager" (modern heroes need to have managers), the varying resistor (potentiometer) P, who is basically a varying voltage divider, and by the means of base voltage Vb tells the transistor what to do. In this case - using a constant base voltage the manager tells him to keep the current flow at a constant rate. The "manager" also has a "bodyguard" - Rb keeping the power source safe from short-circuit's caused by someone playing with the potentiometer P.

Then there are our watchful eyes - the voltmeter V2, used to keep an eye on the voltage drop on the load resistor (the boss), another voltmeter V1, watching over the voltage drop on the transistor's collector-emitter junction, and an ammeter, which gives us a sneak-peak on the current flow.

Last, but not least - the power source VCC, who brings the whole circuit to life :)

Will the circuit pass its exams? edit

Now, let's put our solution on the examination table and make some observations.

What happens when we vary the load? edit

.. we observe a change in V1 and V2, but no change in the ammeter. We're happy: the current flow is constant :) How did that happen?

When we increase the load resistance Rl so does its voltage drop, which can be seen on voltmeter V2 (logically simple - more resistance generates more pressure). As V2 shows an increment, V1 respectively shows a decrement (their sum represents the voltage of the power source Vcc). Since we made the transistor act as a current-stable resistor, it "senses" the decrement of the C-E voltage drop and obediently reduces its resistance, so the sum of resistances in the circuit branch remains the same. Therefore the current flow remains the same too.

When we decrease the load resistance, exactly the opposite happens (V2 shows a decrement, V1 shows an increment, the transistor rises its resistance, the sum of resistance remains the same, and so does the current flow).

What happens when we vary the supply voltage? edit

Well, that was quite easy, so let's try a more difficult test - we will vary the supply voltage. Our expectations are that the current will remain constant, due to the fact that our "manager" (the potentiometer) keeps the bottleneck of the circuit at the same throughput level by applying constant voltage. Unfortunately, as we conduct the experiment we notice that the current changes... see you on the next exam session, constant current source, you've failed.

But what is the problem? Hence the power supply of the potentiometer - it's the one we changed in order to facilitate the examination, therefore we have also changed the bottleneck throughput. Our simplest current source fails this test, because we use the same power source for driving the transistor as well as for powering the whole circuit.

The next step: adding a diode edit

Our solution failed the last test, but is failure a bad thing? It can be considered good, because it gives us the opportunity and motivation to improve our ideas and to move up the ladder.

Obviously, we need to drive our transistor with an independent power supply. One way to do this is to use another voltage supply, dedicated entirely for this purpose. This is somewhat expensive solution, because it uses twice as much resources (two power supplies instead of one).

The more clever way of doing the trick would be to replace the voltage-divider circuit with a voltage-stabilizing one. This can be done by replacing the potentiometer with a diode (which is a voltage stabilizing element, as we all know) - an idea with more potential than you would first think of, because it's the way to go if you want to build the famous current mirror circuit... which is a vast topic itself and it will be considered in another article of Circuit Idea.