The BEAM Circuits Collection is a
Simple photopoppers, based on one
Back in the last century, Wilf drew up a nice design for a
functions plus reverse -- all on a single chip!) here:
In 2002, this design was resurfaced / reinvented on the
Yahoo BEAM list. Here are Wilf's resulting sketches of 3
Note that in the first diagram, the inverters labeled "3"
indicate that for each, 3 inverters
are wired in parallel
together. Wilf later added this explanation:
is for positive feedback for fast switching between the
two motors and to slow down and avoid high frequency
oscillations. With average light and 100pf, the frequency
is 1KHz so there is a kind of vibration as each motor
rapidly turns on and off. The overall effect is
continuous rotation of both motors while moving straight
towards the balanced light condition. Larger values will
slow the oscillations so that each motor stays on longer
(1 second on / 1 second off) and the motion is a "waddle"
or even a twirling motion. With so much "overshoot" the
motion will be somewhat random but the average direction
should be towards balanced light. Try 1000pf or 0.01 uF
for more purposeful motion.
In May of 2003, Wilf revisited this circuit again:
Wilf explains the circuit
is as follows:
This is a revised reversing photovore
schematic showing R2 and R3. The values of R2 and R3 are
experimental and are related to the "resistance"
of the photosensitive devices used in the photo bridge.
R2 should be roughly equal to the minimum resistance
of the the LDR,
etc in bright light. If necesary these sensors should be
shrouded or masked so that the current
through them is not too high in bright light to avoid
power losses. R2 together with C2 limits the maximum
frequency of the monocore
and motor drivers
when the light is bright and the sensors are equally lit.
That further avoids power losses as switching losses
increase with frequency and motors don't respond very
well to high switching frequencies (quivering).
R3 should be equal to the maximum resistance
of the sensors in low light. R3 together with C2 sets the
minimum frequency of the waggle even in the complete dark
which is more interesting than twirling endlessly in a
As I mentioned the values are experimental and should
be adjusted for most interesting range of behaviour.
I also show alternative light sensors including 3 lead
phototransistors used as photodiodes
(AFAIK it is the collector
junction that acts as a reverse biased photodiode.
This adapted photodiode
is not as sensitive as large area types so C2 may need to
be reduced to 0.01uF while the value of R2 and R3 can be
increased by a factor of 10. Some LEDs
can be used as photodiodes
although these are less sensitive still.
Two leaded phototransistors can also be used but may
require extra shielding to reduce light current
in the bridge to acceptable levels. Note that here,
sensitivity is less of a factor than dynamic range.
In March of 2003, Wilf turned
John-Isaac Mumford's Mazibug
into a 74HC240-based circuit he posted:
Try this 74*240 variation which uses an
detection and so is much more efficient. Of course this
design asumes that there is enough light to charge and
run the MAZIBUG at the same time (i.e. full Sun
light). It is only when the cap
is fully charged to 5V that it looks for shadows. When
is down to half voltage
(about 2.5V), MAZIBUG searches out the light again. The
were used to add a reversing circuit.
Wilf followed up some more on the
discussions of Maxibug derivatives, and came up with the "CD
MaxiBug" in a March, 2003 post:
CD MaxiBug is based on the MaxiBug by Math Vos.
If I had only taken some time to read his excellent
tutorial in which he describes feeding stations etc,
I would not have wasted my time reinventing the wheel.
Having said that, maxibug is not perfect: it churns its
wheels while feeding and does not back out of the feeding
station when full. CD MaxiBug v5 uses just a few more
parts but has powerful and efficient motor drivers,
its motors are off while feeding, and it backs up when
The CD Maxibug v5 uses just one 74AC240 chip. The
are of the Current
Dump variety which do not waste any base
and can handle a wide range of motors. The voltage
comparator uses two inverters
with positive feedback to make a Schmitt trigger circuit.
The 1K feedback potentiometer has a small value to
control the LED
when changing state. A steady LED
indicates low supply voltage
and the phototropic mode. In addition, the comparator
uses an FLED to provide efficient voltage
sampling to minimize the 74AC supply current
associated with a slow changing analog input voltage.
Note that the Schmitt trigger configuration does not
exclude the analog level from the input but does insure
positive switching when triggered and reset.
MaxiBug is designed to work with a 5V feeding station
that charges Maxibug's super cap
used for energy storage. The cap
must be rated 5.5V and is usually constructed with two
2.5V super caps
connected in series to provide the voltage
rating. The charging current
is limited by a 22 ohm resistor
in series with the feeding electrodes. A bright light
beacon just above the feeding station electrodes is
designed to attract and to line up MaxiBot's charging
electrodes during the incoming trajectory. The physical
design of the feeding station is up to the reader and
will require considerable ingenuity to perfect. One
consideration for "jacking" into the charger is to avoid
closing the reverse switch when in the docking bay since
that may cause it to back out before the charging
contacts mate. Once contact is made MaxiBug v5 charging
applies +V to pin 19 and automatically turns off the B
section motor drivers
for a low power standby mode. The Maxibot charger
also charges the reverse delay circuit
through a diode
so that when the motors are turned on the MaxiBug will
back up for the duration of the reverse delay.
The super cap
charges up fairly rapidly while the storage cap
is monitored by the voltage
comparator. The Maxibug v5 comparator is designed so that
when the output is low, the feedback pot forces several
milliamps of current
through the reference LED
to set the forward voltage
to about 1.8V. The pot should be adjusted to set the
trigger level close to 5.5V . Use a 6V supply (or
for the charging station. Turn the wiper of the feedback
pot all the way towards the LEDs.
When the cap
is fully charged (~5.5V) slowly adjust the pot until the
triggers. Once triggered, the comparator applies a low
to pin 19 to enable the B section motor drivers
and MaxiBug v5 disconnects from the charger and backs out
of the feeding station.
When fully charged MaxiBug is photophobic and avoids
bright light. The FLED
will both flash as the comparator samples the lower
with each FLED
pulse. When the voltage
drops to about 4V the comparator switches low again and
turns on solid. MaxiBug now seeks out the beacon.
In August of 2003, Wilf posted
an improved version of the "Bare Bones Photovore":
The BBPV is a little vulnerable to damage if a
motor with too low resistance is connected to
the first driver stage. That can cause the
logic level on the second driver paralleled
inputs (connected to first driver outputs) to be
near the switching threshold. That can cause up to
200ma to flow through the second stage output
stage and the circuit can self destruct from
internal thermal overload.
A slightly different design uses two inverters
in series to process the photo input stage and uses
the first and second inverter outputs to drive
the respective parallel inputs of 2 groups of 3
inverters used as motordrivers. A small layout
drawing is attached.
The addition of the cap changes behaviour in
different light conditions like the BeamANT.
For more information...
The 2002 list discussion of 74*240-based
Wilf's later explanation came in this
posting. The May, 2003 version of the circuit
was presented in 2 list posts -- here,
Meanwhile, a more historic 74*240-based
photopopper design, the BEAMant, is documented on
its own page here.