This is only a preview of the January 1995 issue of Silicon Chip. You can view 30 of the 96 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Articles in this series:
Items relevant to "Build A Sun Tracker For Solar Panels":
Items relevant to "Simple Battery Saver For Torches":
Items relevant to "Dolby Pro-Logic Surround Sound Decoder; Pt.2":
Articles in this series:
Articles in this series:
Items relevant to "Build A Stereo Microphone Preamplifier":
Articles in this series:
Items relevant to "Amateur Radio":
Articles in this series:
|
Build a Sun Tracker
for Solar Panels
The addition of a solar tracker to control a rotating frame
can increase the daily energy output of a solar panel by
30% or more. This tracker frame is built around a 60W
panel supplied by Dick Smith Electronics.
14 Silicon Chip
This simple design will increase the
daily output of your solar panels by
around 30% or more. It was designed to
suit a 12V 60W panel but it can be used
without circuit modifications to control
any size panel.
By NENAD STOJADINOVIC
Imagine the scene: an alternative
energy expo with all sorts of nifty
gadgets to make one’s life away from
the power companies just a little
easier. As I wander around, I can’t
help noticing the large numbers of
solar panels bolted immovably to
their stands.
“Surely they would deliver more
power if they followed the Sun
around?”, I ask. “Yes they would,”
they answer, “but solar trackers are
expensive items”. I checked around
and let me tell you, they are not
kidding.
Everyone complains about the
weather but nobody does any
thing
about it. Thinking these sage thoughts,
I sat down at my desk and came up
with a circuit that eventually evolved
into the design you see before you.
Tracker fundamentals
When a solar panel is aimed directly
at the Sun its output is at a maximum
but for a fixed panel this only happens
for a short time each day; before and after that the output drops off markedly.
Ideally, the complete tracker would
follow the Sun in both altitude and
azimuth but that means two motor
drive circuits would be required.
It is more practical to just track from
east to west (ie, altitude) and have a
fixed azimuth which can be changed
manually from time to time to account
for the changing position of the Sun
from summer to winter. Most pub-
lished designs use this approach and
this one is no exception.
Not only should a tracker follow
the Sun from east to west but at the
end of each day the panel should be
swung back to the east so that it faces
the sunrise next day. Most published
designs that we have seen do not
do this. Apart from that, the tracker
should incorporate limit switches so
that the panel is not driven against the
stops if a fault occurs.
Finally, the tracker should only run
intermittently, swinging the panel by
just a few degrees from time to time
and then consume very little power
at other times.
The circuit
The design is based on an LM324
quad op amp where IC1a and IC1b are
configured as a “window” comparator. A window comparator works on
the principle that when the input to
the two comparators is at the desired
level, the output of both will be low
Fig.1: the circuit is essentially a
“window” comparator involving IC1a
& IC1b. Whenever a shadow falls
across LDR2, the output of IC1a goes
high & the motor drives the panel
westward until both LDRs are once
again fully sunlit.
+12V
R1
10k
0V
RE
4.7k
LDR1
EAST
ORP12
VR1
5k
IC1a
3
1
D3
1N914
R8
VR2
10k
9
LDR3
ORP12
10
6
4
IC1c
LM324
8
2
C1
4.7
11
3
LDR2
WEST
1 ORP12
IC1b
E
B
VIEWED FROM BELOW
C
R9
D4
1N914
100k
G
D
S
D
Q3
G
S
M
7
RM
2. 7
5W
C2
0.1
D1
1N914
6
R2
10k
C
R5
100k
D2
1N914
+4.8V 5
E
2xMTP3055
S
14
IC1d
13
RW
4.7k
B
D
12
8
IC2
555
Q2
G
R7
22k
A
7
R3
1k
B
100k
R10
1M
Q1
S1
BC548
HG
C
E WEST
+7.1V 2
S2
HG
EAST
R4
1k
R6
100k
Q4
D
G
S
2xMTP3055
D
Q5
G
S
SOLAR TRACKER
January 1995 15
Fig.2: the PC board
is straightforward to
assemble. Note that
mercury switches S1
& S2 must be angled to
set the limits on panel
rotation, as detailed in
setting-up procedure.
Take care with
component orientation.
SHIELD
LDR2 LDR2
S1
Q1
10k
4.7k
100k
1k
D3
22k
G
D
S
VR2
LDR3
D1 D2
Q3
12V
0.1
1M
1
D4
G
D
S
2. 7
5W
IC2
555
10k
100k
IC1
LM324
VR1
Q2
1k
4.7uF
G
D
S
Q4
100k
4.7k
100k
G
D
S
Q5
S2
MOTOR
The PC board has a blackened cardboard light shield placed between LDR1 &
LDR2. Note that the mercury switches are fairly fragile & will break if roughly
handled. Warning: do not handle mercury it is poisonous.
(off). In this case the input voltage
at point A (pins 3 & 6) is to remain
between the upper and lower trigger
voltages of say, 7.1V and 4.8V, re-
spectively. If the voltage at pin 3 rises
above 7.1V, the output of IC1a jumps
to around 10.7V (on). Similarly, if the
voltage on pin 6 falls to below 4.8V,
the output of IC1b jumps to 10.7V
(also on).
Why a window comparator? Why
not a simple feedback device? The
answer lies in the zone of non-operation while the input voltage is in
the window – the so called “dead
zone”. A propor
tional device will
attempt to follow the Sun exactly and
consequently the motor will always
be either running or on the verge of
running.
The input voltage signal for both
comparators is provided by two light
dependent resistors, LDR1(E) and
LDR2(W), connected in a resistive
divider configuration. It can be seen
that when equal sunlight falls on both
LDRs they will have (roughly) equal
resistances and the supply voltage will
be dropped in two equal increments,
leaving point A somewhere around
+6V. We will pretend that RE and RW
are not there for the moment.
As the Sun moves west, the cardboard divider panel between the two
LDRs throws a shadow onto LDR2(W),
causing its resistance to rise, and so the
RESISTOR COLOUR CODES
❏
❏
❏
❏
❏
❏
❏
❏
No.
1
4
1
2
2
2
1
16 Silicon Chip
Value
1MΩ
100kΩ
22kΩ
10kΩ
4.7kΩ
1kΩ
2.7Ω 5W
4-Band Code (1%)
brown black green brown
brown black yellow brown
red red orange brown
brown black orange brown
yellow violet red brown
brown black red brown
not applicable
5-Band Code (1%)
brown black black yellow brown
brown black black orange brown
red red black red brown
brown black black red brown
yellow violet black brown brown
brown black black brown brown
not applicable
voltage at point A rises correspondingly. As soon as this voltage rises above
+7.1V, IC1a switches on and drives the
panel motor westward.
LDR1E and IC1b together are there
to drive the panel eastward. Normally
this function is not used since the Sun
does not travel east, but is provided in
case of whoopsies such as the family
dog crashing into the panel.
If the panel is moved too far west, a
shadow falls on LDR1 and this causes point A to fall below +4.8V. This
causes IC1b to switch on and drive
the panel east until normal conditions
are restored.
The threshold voltages for IC1a
and IC1b are set by R1, R2 and VR1
acting as a three-way voltage divider.
Reducing the resistance of VR1 will
bring the threshold voltages closer
together and thus reduce the size of
the non-operation “window”.
SC13101951
Fig.3: this is the full size etching pattern for the PC board.
END OF SPINDLE
THREADED TO
ACCEPT MATCHING
NUT ON CIRCUITRY
CASE
V-PULLEY
H-pack output stage
The drive side consists of a set of
four Mosfets in an H-pack arrangement. A west signal from IC1a causes
Q2 and Q5 to switch on and an east
signal switches on Q3 and Q4. R5 and
R6 are included to make sure that the
Mosfets switch off.
At the end of the day, rewind to the
east is taken care of by IC1c which
works as a simple comparator. As
darkness falls, the resistance of LDR3
increases until the voltage on pin 9
reaches the voltage on pin 10. The
reference voltage at pin 10 can be
any arbitrary value above zero and
so is tied to pin 5 of IC1b; ie, 4.8V.
The actual darkness threshold is set
by VR2.
The output of IC1c is fed to a IC2, a
555 timer wired as a monostable with
an ‘ON’ time of about 10 seconds. As
the light faded it was found that shadows from trees and things would start
to trigger the dark sensor but the west
sensor would argue the point, resulting
in the panel hunting back and forth.
IC2 prevents this by fully rewinding
the panel to east the moment the darkness sensor triggers.
When IC2’s output goes high, it
does two things. First, via diode D2,
it switches on Q4 and Q3, driving
the whole show eastward. Second,
it causes IC1d’s output to go low and
thus removes the base drive from
Q1. This shuts off any possible drive
to transistors Q2 and Q5, ensuring
that the tracker will not follow the
(D) SUB-FRAME
(C) TELESCOPING
PROP SECTION
(B) PIVOTING
PROP
(E) SOLAR CELL
CARRIER FRAME
12.7 PILLOW
BLOCK
(A) BASE
(D) SUB-FRAME
Fig.4: this diagram shows the
tracker frame in perspective view.
The dimensions of its various parts
are given in Figs.5-8.
Moon, car headlights, or other such
light sources.
The mercury switches are there
to set the travel limits of the panel.
S1(W) disconnects the gate signal to
the west drivers as the panel tilts to its
westerly limit; similarly, S2(E) cuts off
gate signals to the east drivers as the
panel tilts to its easterly limit.
Some refinements
The tracker will function with
January 1995 17
12.7 DIAMETER SPINDLE
WELDED TO FRAME END
ON CENTRE LINE
12.7 DIAMETER SPINDLE
WELDED TO FRAME END
ON CENTRE LINE
140
232.5
85
555
232.5
25 x 2.5 FLAT IRON LUGS
WELDED TO FRAME TO
SUPPORT SOLAR CELL
MODULE. POSITIONS AND
DIMENSIONS MAY NEED
ADJUSTMENT TO SUIT
YOUR MODULE
305
305
1162
(E) SOLAR CELL CARRIER FRAME
MATERIAL: 25.4 SQUARE MS TUBE 1.6 WALL THICKNESS
ALL CORNERS WELDED
DIMENSIONS IN MILLIMETRES HOLES 7 DIAMETER
Fig.5: the solar cell carrier frame on the prototype was designed to suit a
standard 60 watt panel from Dick Smith Eectronics (Cat. MSX-64).
twitching back and forth but not quite
running. It took some time to figure it
out but this is the result of the com
parators being switched on by a very
slowly changing input.
Comparators have extremely high
gains. As the voltage on the input pin
approaches that of the comparison pin,
peaks of noise on the input will be
amplified tremendously and cause the
comparator to rapidly switch on and
off. Unfortunately, the driver transistors will do the same with consequent
wastage of power. The solution is to
30
just these components but is not as
efficient as it could be. I found a few
small modifications reduced current
consumption dramatically – never a
bad thing with alternate energy.
The first mod came about when I
bought some LDRs that had a very low
resistance in bright sunlight. It seems
that not all ORP12 equivalents are
the same. The addition of RE and RW
reduced the current through the LDRs
and kept them a lot cooler to boot.
The second mod came about when
I noticed the drive motor rapidly
30
25
27
10
703
(B) PIVOTING PROP
MATERIAL: 30 SQUARE MS TUBE
2 WALL THICKNESS
Building the tracker
45
97
107
85
104
85
707
106
53
25.4
45ø x 6mm CHAMFER
12
(C) TELESCOPING PROP SECTION
MATERIAL: 25.4 SQUARE MS TUBE
1.6 WALL THICKNESS
DIMENSIONS IN MILLIMETRES ALL HOLES 7 DIAMETER
Fig.6: construction details for the pivoting prop & telescoping prop sections.
18 Silicon Chip
introduce some hysteresis. On IC1a,
the first instant the comparator switches on causes a high signal to travel
through D3 and R8 to the junction of
the LDRs. The nett effect is to cause the
input on pin 3 to suddenly rise about
0.4V, thus well and truly turning the
comparator on.
The same system is used on IC1b
via D5 and R9. The hysteresis circuit
also causes IC1a to turn off a trifle
later then it otherwise would, thus
causing the panel to turn a bit further
before the motor switches off. This
is a handy thing in preventing the
input voltage from sticking too near
the upper limit.
(Note: the current drain of the tracker is around 10mA when the panel is
stationary, rising to about 1.5A when
the motor is running. Most of the 10mA
quiescent current can be attributed to
the 555 timer and this could be reduced to under 2mA by using a CMOS
555 (ie, a 7555).
The circuit for the solar tracker
is wired on a PC board measuring
125 x 73mm and coded 13101951.
There is nothing at all tricky about
the board assembly. Make sure that
the diodes are all in the right way
around or strange things will happen.
Use sockets for the two ICs but don’t
install them just yet.
PARTS LIST
2.5
87
TWO PIECES OF 25 x 2.5 FLAT IRON,
87 LONG WELDED TO TOP OF
UPRIGHTS TO SUPPORT PILLOW
BLOCKS. HOLES DRILLED TO MATCH
PILLOW BLOCKS
25.4
25.4
25.4
HINGE TANG WELDED
TO UNDERSIDE
25.4
TWO PIECES OF
25 x 2.5 FLAT
IRON WELDED TO BOTTOM
OF HORIZONTAL FOR PIVOT
1220
(D) SUB FRAME
DIMENSIONS IN MILLIMETRES HOLE DIAMETER 7
Fig.7: this diagram shows the dimensions of the frame pivot support.
25.4
Miscellaneous
Hookup wire, solder, blackened
cardboard for LDR shield.
Testing
Before you power up the board,
remove the ICs from their sockets and
make sure that your power is supplied
via a 2A fuse. Alternately, use a current
limited supply if you have one. Hook
up the power and make sure that the
correct voltage is going to pins 4 &
11 of the IC1 socket and pins 1 & 8
25.4 SQUARE MS
TUBE
WALL THICKNESS
1.6
HINGE TANG WELDED TO
BOTTOM END OF SUB FRAME
505
Be careful with the Mosfets. These
little fellows are really rugged once
in the circuit but are easily damaged
when being handled beforehand. Do
not ever touch the pins with your fingers and if they come wrapped in foil,
just peel a bit back around the pins and
solder them in like that. The villain is
static electricity and standard precautions include grounding yourself and
your soldering iron, etc.
25.4
225 x 140 T-HINGE
CUT TO CLEAR SQUARE TUBE
AND WELDED TO ANGLE IRON
32
Resistors (0.25W, 1%)
1 1MΩ
2 4.7kΩ
4 100kΩ
2 1kΩ
1 22kΩ
1 2.7Ω 5W
2 10kΩ
15
Capacitors
1 4.7µF 16VW electrolytic
1 0.1µF monolithic
HORIZONTAL AND TWO UPRIGHTS
25.4 SQUARE MS TUBE WITH
1.6 WALL THICKNESS
UPRIGHTS WELDED TO ENDS
OF HORIZONTAL
8
Semiconductors
1 LM324 quad op amp (IC1)
1 NE555 timer (IC2)
5 1N4148 diodes (D1-D5)
4 MTP3055 Mosfets (Q2-Q5)
1 BC548 NPN transistor (Q1)
3 ORP12 light dependent
resistors (LDR1,2,3)
305
1 photovoltaic solar panel (see
text)
1 tracker frame to suit panel
1 3V barbecue spit motor
1 weatherproof box with
transparent lid
2 pillow blocks
2 V-belt pulleys
1 V-belt to match pulleys
1 4-way insulated terminal block
2 mercury switches (S1,S2)
1 8-pin IC socket
1 14-pin IC socket
1 5kΩ trimpot (VR1)
1 10kΩ trimpot (VR2)
2 OFF 25.4 x 25.4 x 3
ANGLE IRON WELDED
TO SQUARE TUBE
(A) BASE
DIMENSIONS IN MILLIMETRES
HOLE DIAMETER 7
25.4
1250
25.4
Fig.8: the hinge details for the frame pivot support.
of the IC2 sock
et. Measure around
the various other pins, especially the
driver transistors, for the same reason.
When finished, power down and put
the ICs in.
The circuit falls naturally into
west and east drives, and so that’s the
way testing proceeds. Power up and
measure the voltage on pin 3 or 6; it
should be about +6V. Place your finger
alternately over LDR2(W) and LDR1(E)
and the voltage should swing up and
down in unison.
It seems logical that putting one’s
January 1995 19
This view shows the tracker frame with the solar panel removed to reveal the
barbecue spit motor which is driven by the tracker circuitry.
finger over LDR2(W) should result in
the tracker going west. Try it; pin 7
of IC1b should immediately jump to
around +10.5V (high) while pin 1 stays
low and vice versa. Put your finger on
LDR3 and confirm that pin 14 jumps
to the same high level and then low
again when uncovered – you may need
to adjust VR2.
Note that once IC2 seizes control of
the system it doesn’t let go for about 10
seconds and that S2(E) will stay high
during that time. If you want to disable IC2 while you check other parts
of the circuit, short out LDR3 with an
alligator clip.
If all is well, hook up a motor to the
output pins. Covering LDR2(W) will
now drive the motor one way (mark it
on the motor) and LDR1(E) the other.
20 Silicon Chip
LDR3 will always drive the motor east,
regardless of any other LDR you have
covered.
The mechanicals
Being primarily a mechanical engineer, I decided to try doing a little
better than some of the jury rigged
trackers I had seen around. The tracker illustrated in the diagrams was
designed to suit a standard 60 watt
panel from Dick Smith Eectronics (Cat.
MSX-64; price $599.) This is encased
in a welded frame which, in turn, is
mounted on pillow blocks so that it can
be rotated. The stand is hinged at its
base, allowing the panel to be raised
and lowered to account for seasonal
variations in the Sun’s altitude.
The PC board was mounted in
The tracker circuitry is mounted in a weatherproof plastic
box with a transparent lid. The box is mounted on the
same shaft which drives the solar panel via a V-belt.
a weatherproof plastic case with a
transparent lid. A blackened piece
of cardboard is used as a light shield
between LDR1 and LDR2 so that as the
Sun moves across the sky, it repeatedly
throws a shadow across LDR2(W). The
weatherproof box then is mounted at
the end of a belt-driven shaft and in
the same plane as the solar panel, as
shown in the photos.
Motive power was a problem, with
the commonly available windscreen
wiper motors being found to have not
enough torque and too much power.
The best solution comes in the form
of a cute little spit motor from the local barbecue place. The speed is low,
torque is high and they’re cheap. The
only glitch is that the actual motor is
a 3V unit while the Mosfets put out
about 7.5V.
The answer is to use a current
limiting resistor but the value has to
be chosen carefully. Too much and
starting torque ends up too low, while
too little fries the motor. Around 2.7Ω
is about right.
Commissioning
If everything has checked out on
the bench, commissioning should be
fairly simple. Point the LDR’s directly
into the Sun and turn on the power.
The PC board is mounted in the weatherproof box so that
when the frame is rotating from east to west, LDR2 will
have a shadow cast upon it as the Sun moves further west.
Nudge the panel so that LDR2(W)
is about half shaded and the motor
should immediately scream to life. If it
doesn’t, carefully turn VR1 clockwise
until it does. Switch off the power and
wait until sunset.
Darkness level is set by VR2; start
with it fully anticlockwise. When it’s
reasonably dark, adjust VR2 until the
motor starts up. Watch that the panel
frame does not hit the stand at the end
of its travel.
At the fully rewound position,
check that S2(E) has switched off. If
not, gently bend the glass case down
until the mercury falls from the contacts. Do the same with S1(W) for the
westerly limit.
There are two things to be aware of
before you sit back and enjoy watching
your panel do its sunflower routine.
First, putting a plastic lid over your
LDRs can alter their light level settings
and you may need to make some fine
adjustments. Second
, overcast days
are bad news as there is no clear target
for the LDRs. Instead, the tracker will
orient on any patch of lighter sky it
sees, meaning it spends all day winding itself back and forth. In that case,
it is best to switch it off.
(Editor’s note: as an alternative to
turning the tracker off on overcast
days, the output of IC1b could be disabled to prevent the unit from tracking
east. This is easily accomplished by
placing a toggle switch in series with
SC
diode D1).
Is A Solar Tracker Worthwhile Having?
In presenting this article on a solar tracker we should comment on
whether it is worthwhile for all solar panels. In our opinion, it is probably
not worthwhile for panels rated at less than about 18 watts. This is because
the cost of the tracker itself, which could be $100 or more, depending on
how much of it you build yourself, has to be added to the overall cost of
the installation.
Clearly, if you add $100 to the cost of a 10 watt panel, you could buy an
18 watt panel (using Dick Smith Electronics catalog prices as a guide) and
thereby increase the output by 80%. However, for the bigger panels the use
of a tracking mechanism is clearly worthwhile.
January 1995 21
|