Constant Current Source Design

[confused soul]

I need to design a constant current source of 0.1 mA. The load varies
between 1Mohm and 1 Kohm. Using a constant voltage source, is it
possible to design a constant current source?
Any help in this is greatly appreciated.

--Confused Soul

 

[Jim Thompson]

I need to design a constant current source of 0.1 mA. The load varies
between 1Mohm and 1 Kohm. Using a constant voltage source, is it
possible to design a constant current source?
Any help in this is greatly appreciated.

--Confused Soul

0.1mA thru 1Meg is 100V. What supplies do you have available?

...Jim Thompson

 

[John Popelish]

I need to design a constant current source of 0.1 mA. The load varies
between 1Mohm and 1 Kohm. Using a constant voltage source, is it
possible to design a constant current source?
Any help in this is greatly appreciated.

--Confused Soul

Good enough approximations may be possible.
Is the load resistance connected to a positive supply, a negative
supply, ground, or none of the above?

What supply voltages are available?

What kind of frequency response must this current source have?

0.1 mA through 1 Mohm implies 100 volts drop across that resistance,
so a voltage supply greater than 100 volts will be necessary to feed
the current regulator.

 

[John Woodgate]

I read in sci.electronics.design that confused soul
z.googlegroups.com>) about 'Constant Current Source Design', on Mon, 24
Jan 2005:

I need to design a constant current source of 0.1 mA. The load varies
between 1Mohm and 1 Kohm. Using a constant voltage source, is it
possible to design a constant current source?
Any help in this is greatly appreciated.

Yes. It sounds like homework to me. We don't do homework but we may give
hints to polite enquirers.

Have you given all the data? How much variation in the 0.1 mA is
allowed? Do you have to use just a voltage source and resistors, or are
transistors allowed?

0.1 mA through 1 Mohm means that there is 100 V across the 1 Mohm. If
you had a 1 kV supply with a 9 Mohm resistor in series, how would that
work out? Do you think that's a practicable solution?

 

[Winfield Hill]

John Woodgate wrote...

confused soul asked...

Yes. It sounds like homework to me. We don't do homework but we
may give hints to polite enquirers. ...
0.1 mA through 1 Mohm means that there is 100 V across the 1 Mohm.

Hints: check out common bipolar-transistor current-source circuits,
with opamp feedback. Think about using two MPSA42 (npn) or MPSA92
(pnp) transistors, wired in the Darlington configuration.

 

[confused soul]

Looking at the previous replies, I kinda feel dumb that I didnt even
think about the possible voltage required. 100 V seems impractical in
my case. It is possible for me to change the constant current value to
0.01 mA or even 1 uA instead of 0.1 mA. In that case, what kind of
circuits can be used for building a constant current source.

 

[John Woodgate]

I read in sci.electronics.design that confused soul
z.googlegroups.com>) about 'Constant Current Source Design', on Mon, 24
Jan 2005:

Looking at the previous replies, I kinda feel dumb that I didnt even
think about the possible voltage required. 100 V seems impractical in
my case. It is possible for me to change the constant current value to
0.01 mA or even 1 uA instead of 0.1 mA. In that case, what kind of
circuits can be used for building a constant current source.

A few weeks ago we had a thread here in which a FET-based current source
could give output source impedances of gigohms. That should be OK for
you! If you can't search the group archive, maybe someone will give you
a message ID to look up or a copy of one of the relevant posts.

 

[Nicolai]

Have a look at the Fixed Current Source schematic on page 10 of
http://cache.national.com/ds/LM/LM124.pdf
Exchange the PNP transistor with a P-channel MOSFET with low Vgs threshold
and recalculate the resistor values for your own needs and it should do the
trick. If you need better accuracy just use a better OP-AMP and resistors.

The High Compliance Current Sink on page 12 will do the same trick if you
need to sink instead of source the current. Again just substitute the NPN
transistor with an N-channel MOSFET and recalculate for your needs. The
current sink circuit has the advantage that it can be made to work with any
voltage as long as you use the right MOSFET.

You can keep the transistors if you don't like MOSFETs but then you will
have a small error contribution from the base current. Old datasheets are
abundant with similar circuits.

HTH

 

[Paul Burridge]

Looking at the previous replies, I kinda feel dumb that I didnt even
think about the possible voltage required. 100 V seems impractical in
my case. It is possible for me to change the constant current value to
0.01 mA or even 1 uA instead of 0.1 mA. In that case, what kind of
circuits can be used for building a constant current source.

A FET-based solution is well worth considering!

 

[Spehro Pefhany]

Have a look at the Fixed Current Source schematic on page 10 of
http://cache.national.com/ds/LM/LM124.pdf
Exchange the PNP transistor with a P-channel MOSFET with low Vgs threshold
and recalculate the resistor values for your own needs and it should do the
trick. If you need better accuracy just use a better OP-AMP and resistors.

The High Compliance Current Sink on page 12 will do the same trick if you
need to sink instead of source the current. Again just substitute the NPN
transistor with an N-channel MOSFET and recalculate for your needs. The
current sink circuit has the advantage that it can be made to work with any
voltage as long as you use the right MOSFET.

You can keep the transistors if you don't like MOSFETs but then you will
have a small error contribution from the base current. Old datasheets are
abundant with similar circuits.

HTH

With a bipolar op-amp such as the LM324, there's an error component
from the input bias current. At 1uA, it could be as much as 50% error
(typically 4%). It also varies with supply voltage and somewhat with
temperature. There are better op-amps.


Best regards,
Spehro Pefhany