# How to build a constant current source?

Discussion in 'Electronic Basics' started by [email protected], Jun 17, 2005.

1. ### Guest

I have been frustrated for months working through an introductory
circuits book, as I want to build many of the circuits, but many of
them include constant current sources. Why is it so easy to find
voltage sources at your local store (i.e., batteries!), but searching
for current sources on the web leads to a complicated bunch of circuit
diagrams?

Here is my naive question: using thevenin-norton equivalent circuits,
couldn't I transform a battery (i.e., voltage source) into the desired
equivalent current source using Vth=InorReq. That is, can I put a
voltage source in series with a resistor (as opposed to its equivalent,
a current source in parallel with the same resistor)? What is the
problem with doing that?

Does anyone know where I could buy a cheap but reliable current source?
Why is this so hard?

2. ### Larry BrasfieldGuest

Your desire to build many circuits but not any of the
simpler current sources is a puzzling combination.
Maybe nothing. How much of a current source do you want?
How constant does it have to be versus terminal voltage?
Your subject indicates you want to build a current source.
If that is out of the question, one of these parts,
http://www.rohm.com/products/shortform/26rstr/rstr_index.html
together with a battery could be called a "current source".

You can use a JFET, with its gate connected to its
source, as a current source. Put a variable resistor in
series with the source to adjust its value downward.
Somebody used to sell a two-terminal current source
which was just a JFET hooked up that way.
It's not.

3. ### Rich GriseGuest

If a current sink will do, then you can do it with an NPN transistor,
a couple of diodes, and a couple of resistors.

"output"
|
|= x mA
|
C
+V ---[R1]--+--B
| E
V |
--- |
| [R2]
V |
--- |
| |
GND GND

Where R2 is sized to give .6V at your desired current, R1 is maybe
1K or so, and you should recognize the two diodes. I think a 2N4401
can go up to almost a half an amp.

Turning that into a source, with a PNP transistor, is, of course,
left as an exercise for the reader. ;-)

Have Fun!
Rich

4. ### John PopelishGuest

Voltage sources produce a fairly constant output voltage for all loads
from infinite resistance (zero current) to some minimum resistance
(more than zero current), but they are only approximate voltage
sources, because their output voltage does sag and current passes
through them. This is because they have some internal resistance.

Practical current sources hold a very nearly constant current for all
loads from zero ohms (zero voltage drop) to some maximum value of
resistance (some maximum voltage drop called the compliance voltage).

Ideal current sources would produce a constant current, regardless of
the load resistance, including infinite resistance, by producing a
compliance voltage up to infinity volts.

It is a lucky accident that we have discovered chemical cells that
produce something close to a voltage source over useful ranges of load
current. There is no similar simple power sources that provide a
constant current output.
The problem is that it is not a current source, but a current source
in parallel with a fairly low value of resistance. Mathematically
transforming a voltage source that has a little series resistance to
the equivalent current source with parallel resistance does not
actually alter the characteristics of the source, at all.

Almost all practical current sources are actually active variable
resistances that waste all the extra voltage from a voltage supply
that is not needed to force the specified current through the load.
This active circuit has to be designed to pass the required current,
while measuring it, in some way, and using some sort of feedback, vary
an active resistance as needed to hold the measured value of current,
constant, till the load uses up all the available voltage. These
active current regulators are specified with their effective parallel
resistance (equivalent to a perfect current source in parallel with a
resistor), their voltage compliance, their frequency response, and
sometimes how long it takes for them to recover current regulation
after being exposed to an excessively high load resistance and then
the load falls to within the normal operating range.

Name your specifications, and people, here will help you design
current sources (regulators) that meet your needs.

5. ### Bob MonsenGuest

You can buy a constant current source. I have a cheap one myself. It is
basically a power supply with a voltage source and current limit. Turn
the current to 0, short the output, turn the voltage all the way up, and
then adjust the current to your liking. You can get them here:

http://www.mpja.com/

You can also build a constant current source, given a voltage source. An
LM317 voltage regulator makes a good one. It tries to keep the voltage
between it's output and adj terminal at 1.24V. Thus, if you put a
resistor of value R on the output, and connect the adj terminal after
it, before the load, the LM317 will try to adjust the voltage so that
there is a constant 1.24/R A across the load. If you use an R = 100, for
example, then it will try to keep 12.4mA across the output by varying
the output voltage. Note that it will only be able to adjust the voltage
within the limits of the original voltage source, minus some for
whatever the LM317 + resistor needs, probably 3 or 4 volts.

6. ### Guest

Wow, you have all been very helpful. I will print these suggestions out
and try to figure out which would be best. As far as specifics, I would
be happy with something that could generate a single DC current of
~100mA. In the best of possible worlds, I would love something that
would let me, through a turnpot, vary the DC current between 0 and 1
Amp. I don't want anything fancy, just something so that I can play
with my breadboard while working through an electronics book.

I guess we are lucky to have batteries. I wonder why no companies sell
a cheap current source that takes AC in or even uses batteries to power
it. They could probably make a KILLING with all the EE departments
around the country! Maybe I'll make one and provide the supply!

7. ### John GGuest

Nobody as asked:-
Why do you want a constant current supply?

Most of the electronics experiments you will encounter will require
constant Voltage and these supplies are every where because everyone
uses them for most everyting.

8. ### John PopelishGuest

Take a look at page 16 of the data sheet for the very common voltage
regulator LM317:
http://www.national.com/ds/LM/LM117.pdf

All you need for a variable current regulator is an isolated voltage
source (battery or wall wart) and a variable, low value resistor, or a
selector switch that selects various low value resistors.

The current setpoint is 1.2 volts/resistance. The LM317 may well need
a heat sink for currents above 100 mA. At 1 amp, the resistor will
also produce some significant heat ((1.2 volts^2)/resistance).

9. ### John PopelishGuest

Have you ever gone through basic electronics text book? They are full
of hypothetical circuits that contain voltage sources and current
sources, to demonstrate the analytical principles. If you are an
experimentalist, you need both sources to verify that they are not
lying to you.

10. ### Guest

Have you ever gone through basic electronics text book? They are full
Exactly, my main motivation. Secondarily, I am a neuroscientist. We
inject current into neurons all the time, and I want to stop treating
the stimulator as a black box!

11. ### kellGuest

Whose neurons?
Hey has anybody seen my foil hat!!?

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Continue to site