Maker Pro
Maker Pro

Op amp darlington current source oscillations

J

Justin Dobbs

Hello!

I have a simple linear current source circuit pictured below. The
test
loads are an 8 ohm wire-wound resistor and a 7-8 ohm 450 foot roll of
hookup
wire (tested separately).

The purpose of the circuit is to take a PWM signal from a
microcontroller, whose average DC value is 0.2 volts per amp desired,
and feed the load with the desired current. I am testing the circuit
on the bench using a separate DC power supply (with the grounds tied
together in one spot). The transistor is actually a 2n6284 darlington
pair, and the op amp is a dual rail-to-rail LMC6062AIM "micro power"
op amp (in a very frustrating SOIC package). My interpretation of the
spec sheet is that it will source up to 25 milliamps, which should be
adequate for the darlington pair. The goal is to regulate between 0-5
amps continuously with a 10mS rise time.

I thought this circuit was working a few days ago. Now it seems to
oscillate/switch on and off at 1-10 kHz depending on the lowpass
filter I put after the first op-amp stage.

This circuit initially had just the first op-amp stage driving the
darlington pair though the 10-ohm resistor, with no lowpass filter,
and the behavior was the same.

With the large inductance (coil of wire), the current climbs to the
desired value, then the transistor is appears to be switched off
quickly until the current has fallen by 50%, and then the current
begins to rise again. The current drops so quickly that the transorb
conducts.

With the wire-wound 8 ohm resistance, the current spikes up to the
setpoint, then drops immediately to zero, and after a 1 millisecond
delay, spikes again. In other words, a delta-train with no DC.
Almost like you would expect from a fixed-frequency switchmode current
chopper driving a resistive load.

I want the circuit to provide smooth DC to the load. The op-amp
appears to source enough current to turn the darlington sufficiently
"on" to reach the current I want, but the regulation seems
non-existent. I tried adding a 2n2222 to drive the base of the
darlington, to make a transistor triplet, and this made the problem
worse. The lowpass filter between the two op-amps only changes the
frequency of this switching action. Eliminating the lowpass filter
changes the ripple to a higher frequency (50 kHz and above). The
behavior is the same with or without the 10 ohm resistor feeding the
gate.

For amusement I put a 18,000uF capacitor across the output of my power
supply, just in case of poor regulation, but that doesn't seem to
help. The 12 volts is straight from the (bench-top) regulated supply,
and the 5 volts is derived from a linear regulator with a 470uF
capacitor on its output. Putting a 220uF electrolytic across the op
amp's .1uF tantalum bypass cap does not seem to cure the problem.

I'm wondering whether my choice of op-amp is the problem due to its
low output current. But the circuit seems to me like it should be at
least stable. Although the transistor may be oversized a tad, the
topology is simple and classic enough that it ought to work! What am
I missing?

Thanks,

Justin


Rload=8ohm Lload
12V |--------------/\/\/\------88888----------------------.
|
5V |
--- .1uF |
|---| |---. |
| /// |------.
| LMC6062AIM | |
10k |\| __|\ | |
---/\/\/\----------|+\ 10k / |+\ 10ohm |C |
| | >---/\/\/\---' | >--+--/\/\/\-----| _|_/
| . --|-/ | .-|-/ | |E //\
75V trans
-- | |/| | | |/ | | /__\
zener
.1uF -- | | | |_______| | |
| | /// | | |
| | --- | |
/// | --- 10uF | |
| | | |
| /// | |
|___________________________________________+______|
|
|
\
/ 0.2 ohm
\ 5W
/
|
|
///
 
B

Ban

Justin said:
I want the circuit to provide smooth DC to the load. The op-amp
appears to source enough current to turn the darlington sufficiently
"on" to reach the current I want, but the regulation seems
non-existent. I tried adding a 2n2222 to drive the base of the
darlington, to make a transistor triplet, and this made the problem
worse. The lowpass filter between the two op-amps only changes the
frequency of this switching action. Eliminating the lowpass filter
changes the ripple to a higher frequency (50 kHz and above). The
behavior is the same with or without the 10 ohm resistor feeding the
gate.

For amusement I put a 18,000uF capacitor across the output of my power
supply, just in case of poor regulation, but that doesn't seem to
help. The 12 volts is straight from the (bench-top) regulated supply,
and the 5 volts is derived from a linear regulator with a 470uF
capacitor on its output. Putting a 220uF electrolytic across the op
amp's .1uF tantalum bypass cap does not seem to cure the problem.

I'm wondering whether my choice of op-amp is the problem due to its
low output current. But the circuit seems to me like it should be at
least stable. Although the transistor may be oversized a tad, the
topology is simple and classic enough that it ought to work! What am
I missing?

Thanks,

Justin


Rload=8ohm Lload
12V |--------------/\/\/\------88888----------------------.
|
5V |
--- .1uF |
|---| |---. |
| /// |------.
| LMC6062AIM | |
10k |\| __|\ | |
---/\/\/\----------|+\ 10k / |+\ 10ohm |C |
| | >---/\/\/\---' | >--+--/\/\/\-----| _|_/
| . --|-/ | .-|-/ | |E //\
75V trans
-- | |/| | | |/ | | /__\
zener
.1uF -- | | | |_______| | |
| | /// | | |
| | --- | |
/// | --- 10uF | |
| | | |
| /// | |
|___________________________________________+______|
|
|
\
/ 0.2 ohm
\ 5W
/
|
|
///


Take the first stage out completly and connect the feedback to the -input of
the remaining amp. The 10k/10u will give a good stable DC voltage from the
PWM control. Learn about opamps and try to understand how they work apart
from cpying bits and putting them together. This is *not* software.
 
W

Winfield Hill

Justin Dobbs wrote...
I have a simple linear current source circuit pictured below.
The test loads are an 8 ohm wire-wound resistor and a 7-8 ohm
450 foot roll of hookup wire (tested separately).

The purpose of the circuit is to take a PWM signal from a
microcontroller, whose average DC value is 0.2 volts per amp
desired, and feed the load with the desired current. ...
The transistor is actually a 2n6284 darlington pair, and the
op amp is a dual rail-to-rail LMC6062AIM "micro power" op amp

Go back to a single opamp section, but with a compensation path.

.. Rload=8ohm Lload
.. 12V |---------/\/\/\------88888-------.
.. |
.. 5V |
.. --- .1uF |
.. |---| |---. |
.. | | +------.
.. | /// | |
.. 10k |\| | |
..---/\/\/\--+------|+\ 220 C |
.. | | >--+--/\/\/\--- B _|_/
.. | .--|-/ | E /_\
.. --- | |/| --- | |
.. .1uF --- | | --- 0.01uF +------'
.. | | /// | |
.. | | | 10k |
.. | '--------+---/\/\/\-----+
.. | |
.. | \
.. | / 0.2 ohm
.. | \ 5W
.. | /
.. | |
.. '---------------------------+
.. |
.. ///
 
F

Fred Bloggs

You do know that with a 12V power supply, 8.2 ohms load resistance, and
allowing 2V for the Darlington, that the maximum possible output current
will be (12-2)/8.2= 1.2 Amps? or 1.2 x 0.2V/A= 0.24 Volts at the input,
which you are most likely exceeding. For a 5 Amp limit, you would need
5*8.2+2=41.2V of power supply in place of the 12V. Are you even close?
 
J

Justin Dobbs

You conveniently cut out a key part of the original post:

To clarify, this is what was oscillating:
 
J

Jim Thompson

You conveniently cut out a key part of the original post:


To clarify, this is what was oscillating:

Connect anode of zener to ground and see if the oscillation stops.

...Jim Thompson
 
J

Justin Dobbs

Whoops, I wasn't clear about that at all. The supply will be a 12
volt battery which might range from 11 to 15 volts depending on state
of charge and external charging.

Ultimately I wish to swing 0-5 amps. The 8 ohms is a large roll of
wire that I am using as a test inductance; I do not want to test with
the real solenoid (a proportional fuel control valve), since the coil
will quickly overheat without cooling measures in place. This means I
will probably fine tune the circuit once I have the real load set up,
but I am trying to achieve stability with some load (any load)
initially.

I have tried other test load combinations (using wire-wound resistors)
to yield anywhere from 2 to 8 ohms. The real load will be a 2.6 ohm
solenoid coil that I wish to drive proportionately. This yields a
fully on current between 4 and 5 ohms.

Sorry about the confusion!

Justin
 
Top