Hi All;

Wilf and I have been developing another solar tracker
that is based on a 74AC240 Dual Quad Tristate Buffer.
There have been a number of variations. This is
the results. See:
http://www.redrok.com/images/beamstepper7e.gif

The 74AC240 stepper driver works by enabling each half
of the buffer. Only one half can be enabled at a time.

Let's assume that the top half of the driver is enabled.
U1A & U1B along with R8, C1, & the input protection
resister R7 form a square wave oscillator. The outputs
of U1A & U1B directly drive one coil of a bipolar stepper
motor.

U1C & U1D along with R9, C2, & the input protection
resister R10 form a 90 degree phase shift. The outputs
of U1C & U1D directly drive the other coil of the bipolar
stepper motor. The motor turns in one direction.

If the second bottom half of the driver is enabled the
oscillator using U1E & U1F work as before. U1H & U1G
along with R12, C3, & the input protection
resister R11 form a 90 degree phase shift. Except it's
connected the other way around from before so it's
actually 270 degrees. The outputs of U1H & U1G directly
drive the other coil of the bipolar stepper motor. The
motor turns in the other direction. Neat, Huh!

An earlier version of the circuit didn't work well
because the the sensors presented an analog enable
signal. This was sometimes at the threshold voltage
which caused the buffer to have high idle current and
sometimes cross coupling which was a bad thing. %^(

What was needed was a sensor that had a Schmitt trigger
input. This could be done using a Schmitt trigger gate
which works well. I suggest a 40106 or 74AHCT14. However,
this needs a second IC.

A better solution is to make the sensor have Schmitt
action. The first version was:
http://www.redrok.com/images/beamstepper7a.gif
The problem was that it worked over a limited voltage
range.

http://www.redrok.com/images/beamstepper7e.gif
works better. Q1 & Q3 and Q2 & Q4 each form a bistable
latch similar in operation to an SCR.

Let's start with the left side without the LEDs.
Initially no current flows. The series resisters
R5 & R2 cause a small bias current to flow in the base
of Q1. Which pass current through R1 causing Q3 to
conduct. Since Q3 shorts out R5 the current through
R2 doubles. The output at the collector of Q1 snaps
high disabling the connected buffer.

(Note, R5 & R6 aren't actually required. It turns
out that leakage currents in the transistors is enough
to get started. I tried many transistors and never found
one that didn't work as expected. Prudent circuit design
demands that R5 & R6 be included because one might find
a transistor that is so perfect it won't work. Bummer. )

The now connected and lit LED1 has the ability to
absorb the current through R2 starving Q1 which
switches off resulting in the output snapping low.
Q3 also switches off reducing the bias current
in R2 to 1/2. This condition persists until the
LED goes dark.

You might ask where the current for the other side of
the LED comes from. It is from base of Q2 on the right
side. Actually, when the left side is turned off the
right side is turned on doubly as the current from
both R2 and R3 go through the base.

The right side works the same way. Since the LEDs
are connected anti parallel only one latch can
be off at a time. This is safe for the buffers.

When both of the quad buffers are supposed to be off
it is essential that all inputs not be near the
threshold to have the lowest idle current. R13 & R14
ensure that all inputs be near ground. All inputs
are connected to R13 or R14 either directly, through
input resisters, or through the stepper motor. I
added R15 & R16 for testing when the stepper motor
is disconnected. If the motor is permanently
connected R15 & R16 aren't needed. R13 & R14 can also
be connected to VCC. They don't even need to be to
the same voltage, although it operates quicker if
they are the same.

I have tested this circuit with about 25 different
74AC240s. They all worked as expected.

I ran the circuit from about 2.4V to 8.5V.
OK, one shouldn't go past 7V to be within the specs
of the 74AC240.

The sensor section was tested to 40V. It still works
well, the sensitivity is less because the bias current
is proportional to voltage which requires brighter
illumination to work.

The step patterns are not perfectly symmetrical because
this is essentially an analog circuit. Some resister
adjustment can be done.

To change the speed of the motor adjust the capacitor
values. Note, all three need to be the same value.

I have chosen the time constants of R9-C2 & R12-C3
to be about 3/4ths of R8-C1. Try to keep these ratios.
( BTW, I'm not sure this is the exact ratio but it
seams about right. )

The 10M resisters in the sensor are the largest
commonly available resisters in 1/8W size. I tried
22M in 1/4W and that worked well with added
sensitivity. I suppose if you could find 100M they
would work even better.

I have a variation which is even more sensitive to
low light levels. Ask me if you want this variation.

I have to thank Wilf for his invaluable help in the
circuit design. Thanks Wilf.

Have fun, Duane

--
Home of the $35 LED solar tracker.
http://www.redrok.com/electron.htm#led3
CUL8ER \ \ \ \ \ \\ \ \ Receiver
Powered by\ \ \ \ \ \\ \ \ [*]
Thermonuclear \ \Solar\Energy\from the Sun \ /////|
Energy(the Sun) \ \ \ \ \\ \ / / /\/ / /|
\ \ \ \ \ /\ / \/ / / / |
WA0VBE \ \ \ \ / /\ \/ / / \/ /|
Ziggy \ \ \/ / / \ \/ \/ /\ |
\ / \ \/ / /\ \\ / \ / / |
"Red Rock Energy" === ===\ / \ / \ === \ / ===
Duane C. Johnson, Designer=== === \ \ === / |
1825 Florence St Mirrors,Heliostats,Controls & Mounts|
White Bear Lake, Minnesota \ \ / |
USA 55110-3364 \ \ |
(651)635-5O65 work \ \ / |
(651)426-4766 home use Courier New Font \ \ |
(413)556-659O Fax copyright \ / |
(651)583-2O62 Red Rock Energy Site (C)980907 ===\ |
redrok@... (my primary email: address) \ |
redrok2@... (Hotmail address) \ |
duane.johnson@... (Unisys address) \ |
http://www.redrok.com/index.htm (My New Web site) \|
These are my opinions, and not that of Unisys Corp. ===