The constant current source is a very handy little circuit that gets used quite a lot in electronics.  The constant current source does exactly what it’s name implies – it will provide a maximum constant current regardless of voltage.

Normally the way in which we achieve current limiting is via a resistor.

Following ohm’s law which states that V = I x R we can easily calculate the values we need.

For example:  If we have a supply of 25V, and we need a maximum current of 100mA to flow then we need a resistor which is 25V / 0.1A = 250Ω.

This works very well and is used nearly everywhere.  It does have one problem however:  The current limiting is heavily dependent on the voltage.  If for example the supply voltage goes up to 30V, then the maximum current for the same sized resistor becomes 30V / 250Ω = 120mA.  Likewise if the voltage goes down the maximum current will also be reduced.  The effects of this is shown below:

 

CCS_Fig1

 

Imagine you have a power supply which is being heavily loaded and this causes 2V P-P worth of ripple.  A circuit which is dependent on a maximum constant current flowing will be directly impacted by this fluctuation in the voltage rail.  One such circuit that comes to mind immediately is the long-tail pair input stage used in nearly all discrete pre-amplifiers and power amplifiers.  This isn’t great because it automatically introduces rail voltage related ripple and other problems into the very first stage of your amplifier circuit.  Obviously a much better solution is required!

Now – there are many ways to come up with current limiting circuits.  I have drawn 5 different types so we can do some testing on how well they work.  For our test case each of the 5 circuits have been set up to draw 100mA on for a 25V power supply:

Note:  This is a theoretical exercise to show how different implementations of a current source work.  Please do not build any of these circuits without taking into account the maximum collector currents, collector-emitter voltages and power dissipation through the devices.  For a more detailed explanation of how to calculate these, please refer to this article.

CCS_Fig2_Types

There may be names for these different types of current sources (limiters) but I am not sure what they are, hence we will simply refer to them as Type A through to Type E.

So now, let’s increase the rail voltage by 10% to 27.5V and see how well each of our current sources are able to cope:

CCS_Fig3_27V5

TYPE A:  As expected, a 10% increase in voltage has resulted in a 10% increase in current.
TYPE B: 1% increase in current.
TYPE C: 1% increase in current.
TYPE D: less than 1% increase in current.
TYPE E: 1% increase in current.

Next, let’s look at what happens when we push the voltage up by 40% to 35V and see how they cope:

CCS_Fig4_35V

TYPE A:  As expected, a 40% increase in voltage has resulted in a 40% increase in current.
TYPE B: 3% increase in current.
TYPE C: 6% increase in current.
TYPE D: 1% increase in current.
TYPE E: 2% increase in current.

Next, let’s look at what happens when we push the voltage up by 80% to 45V and see how they cope:

CCS_Fig5_45V

TYPE A:  As expected, an 80% increase in voltage has resulted in an 80% increase in current.
TYPE B: 5% increase in current.
TYPE C: 11% increase in current.
TYPE D: 3% increase in current.
TYPE E: 4% increase in current.

So, we can see that TYPE D is in the lead with TYPE E slightly behind followed by TYPE B.  The other types have now fallen out of the race in terms of performance for increasing voltages.  So the obvious next thing to do is to see how they perform for lower input voltages.

We started off with a 25V supply, so let’s drop that by 10% down to 22.5V:

CCS_Fig6_22V5

TYPE A:  As expected, a 10% decrease in voltage has resulted in a 10% decrease in current.
TYPE B: less than 1% decrease in current.
TYPE C: 2% decrease in current.
TYPE D: 1% decrease in current.
TYPE E: less than 1% decrease in current.

Next let’s look at what happens when we apply a 40% voltage drop down to 15V:

CCS_Fig7_15V

TYPE A:  As expected, a 40% decrease in voltage has resulted in a 40% decrease in current.
TYPE B: 4% decrease in current.
TYPE C: 8% decrease in current.
TYPE D: 3% decrease in current.
TYPE E: 3% decrease in current.

Next let’s look at what happens when we apply an 80% voltage drop down to only 5V:

CCS_Fig8_5V

TYPE A:  As expected, an 80% decrease in voltage has resulted in an 80% decrease in current.
TYPE B: 15% decrease in current.
TYPE C: 24% decrease in current.
TYPE D: 10% decrease in current.
TYPE E: 20% decrease in current.


Conclusions

By looking at these tests you will see that all of the current sources are significantly better at maintaining a steady current than the standard resistor method.  Now that does not mean we need to change all resistors to current sources at all.  Current sources are normally only implemented where we need the extra stability and is highly recommended for use in the input stages of discrete pre-amplifiers and power amplifiers.

So out of all these tests it is clear the the overall performance champion is the TYPE D current source with the TYPE E a close second.  Therefore, let’s not concern ourselves further with the rest of these topologies and look deeper into the TYPE D design which will be the subject of it’s own article.