This is my home-made solar panel sun tracker. It is based on a 1960s vintage TV antenna rotator, driven by 21st century microcontroller technology.
It was pretty easy to build. This web site shows how I did it.
This is one of my more ambitious and complicated projects. Unfortunately, I have to include the usual disclaimer that I won't be able to give people
trying to build one of these much in the way of individual attention. My inbox gets flooded with questions and requests for help every day, and there
is just no way I can help everyone, or even most people. You'll need to have a decent grasp of mechanics, electronics and programming to duplicate
this project. If you don't have them, well I am not going to be able to teach it all to you. Basically, you are going to be pretty much on your own. I will be
continually updating this web site with answers to frequently asked questions as time goes on. So if you don't get a response to a question, check back
from time to time. If a lot of people are asking the same thing, I will address it here.
Why build a tracking platform for my solar panels? Solar panels produce a lot more power if they are pointed directly at the sun all the time than they do in a fixed position. I got tired of manually moving my panels to keep them pointed at the sun throughout the course of the day. If I wasn't around to move them every few hours, they wouldn't make enough power to keep my batteries fully charged. I decided to automate the process and free myself from having to manually move the panels.
Here is the box the antenna rotator came in. It still has the $15 yard sale price tag on it. The box is beat up and faded from being
in storage for so long, but the unit inside was still brand new and wrapped in the original plastic.
It is an older unit, based on 1960s technology. The person had purchased the unit new, but never used it. It had been sitting in a
box in their garage for decades until they finally decided to get rid of it at the yard sale.
The first step was to come up with a way mount the drive motor and solar panel(s). I did a little back of the envelope brainstorming and
designed a mount for the tracking system that was simple, inexpensive, and easily broken down for transport. It is made mostly from 2X4s
and standard pipe fittings, and is held together with carriage bolts.
Here is a photo of the north side support of the solar tracker. It measures 48 inches wide at the base and stands 43 1/2 inches tall.
Keep in mind that these dimensions are only correct for use at 34.6 degrees north latitude. If you are significantly further north
or south, you will need to modify the dimensions of this piece. More on that below. The support is made of 2X4s that are screwed and
glued together. Note that there are two little feet on the bottom. They aid in levelling the unit when setting it up. The gap between
the upright 2X4s is exactly the thickness of another 2X4, or about 1 1/2 inches.
Here is a photo of the south side support of the solar tracker. It measures 24 inches wide and stands 13 1/2 inches tall. It too
is made from 2X4s glued and screwed together. This piece also has little feet to aid in levelling the entire unit when setting it
up. This piece is probably more or less universal, and will work for a wide variety of latitudes. Again, the gap between
the upright 2X4s is exactly the thickness of another 2X4, or about 1 1/2 inches.
The horizontal 2X4 brace that goes from the bottom of the north support to the bottom of the south support is 48 inches long. It fits
between the uprights and gets bolted through them. This is another piece that will have to be sized for your particular latitude, since
the distance between the north and south supports will change as the angle of the driven pipe changes.
Here is the heart of the tracker unit. This is the drive motor and rotating assembly. The antenna rotator drive motor is at the left, with its
associated mounting structure. A 4 foot long, 1 inch steel pipe is driven by the rotator, and will carry the solar panels. A bearing and
mounting structure are at the right end. Details below.
Here is a close-up of the motor end. The antenna rotator is designed to be clamped onto a fixed mast, and rotate a shorter mast with
an antenna attached to it. So I created a pseudo fixed mast to clamp it to. The short piece of 1 inch pipe at the top (under the coil of wire)
serves as the mounting point for the rotator. The short piece of pipe screws into a floor flange, which in turn is bolted to a 3 1/2 X 3 1/2
square piece of wood that is glued and screwed to a 12 inch long piece of 2X4. The 2X4 slips between the uprights on the north support and
gets bolted in place.
Here is a close-up of the bearing end. The lower end of the 4 foot long pipe that carries the solar panels screws into a union that has been modified to serve
as a bearing (more on that below). A close nipple connects the other side of the union to another floor flange. The floor flange is bolted
to another wooden mounting structure identical to the one at the other end, but one corner was cut off to prevent it from interfering with the lower
brace holding the north and south supports together.
The first time I assembled the unit, I held all the pieces together with large c-clamps. Once I got the angle of the drive axis correct and
everything nice and squared up, the clamps were tightened down to hold it that way. Then I drilled holes for long carriage bolts to bolt
Here is another view of how the rotator head is mounted. I don't draw actual blueprints for the stuff I build. So don't bother writing and
asking for them. I tend to just visualize stuff in my head and then build it without bothering with the intermediate step of drawing up plans.
I know this makes it difficult for others wanting to reproduce my work, sorry. If you live at a different latitude than me, you are going to have to
modify the design anyway. So there isn't much point in my being too specific about dimensions anyway. But I'll take lots of pictures and post them here. People with
the knack for building things should be able to figure it out from the photos and a few dimensions. If you are having trouble figuring out something,
write and ask a question. I will get back to you with info, advice, dimensions, and/or more photos as time allows. If I find a lot of people asking
the same question, I will put the answer here on the web site.
This photo shows how the lower bearing end of the drive pipe fits into the south side support and gets bolted into place with
carriage bolts. The other end is similarly attached to the north side support. The lower end of the diagonal brace is also visible.
Here is a close-up of how the union has been converted into a bearing. This is an old amateur astronomer's trick for building telescope
mounts using pipe fittings. I'm an old amateur astronomer, so I used it here. It works great and is dirt cheap.
This photo shows one of the aluminum frames that hold the solar panels. It is made from aluminum angle. This particular frame holds a 100W panel,
and measures 47 1/8 by 21 1/2 inches inside dimensions. Basically, it is just slightly larger than the outside dimensions of the solar panel.
The panel drops right into the frame, and will be held in place with screws that go through the frame into the sides of the panel.
This photo shows how the aluminum angles are butted and screwed together at the corners. I didn't have time to get fancy. I only spent about
an hour tops building my frames, but they work great. If you have the time and skills, feel free to miter and MIG/TIG weld the corners.
Here is a close-up detail shot of how the hose clamps are used to mount the frames on the tracker drive pipe. Tightening down the hose clamps
really locks the frames onto the pipe quite tightly. I was somewhat surprised at how well it worked.
During initial indoor testing I only mounted one solar panel length-wise on the tracker, taking up the whole drive pipe. My intention all along though was to
eventually mount two panels. The motor seemed to have plenty of torque, and with counter-weighting, I was confident it could handle two panels. If you only
have or need one panel, this is a way to mount it.
This photo shows two aluminum frames for holding panels clamped onto the drive pipe. By turning the panels 90 degrees, two will now fit. The frames
are slightly different sizes because the panels they will be holding are slightly different.
This photo shows the two solar panels in place. The panels just drop right into the frames, which have been made slightly larger than the outside
dimensions of the panels. Screws hold the panels in place so the wind can't blow them out of the frames.
This photo shows the counterweight pipe. It is a piece of 1 inch steel pipe 30 inches long. It gets screwed into the elbow at the top end
of the motor unit. The pipe alone is a little more counterweight than is needed for one panel. For two panels I added a steel T fitting at
the end of the pipe. The antenna rotator was designed to move a balanced vertical mast. The counterweight is necessary to reduce the amount
of torque the motor has to exert to move the panels which are hung off the side of a nearly horizontal mast. Your panels will probably
have a different weight than mine and need a different counterweight arrangement. Experiment with different lengths of pipe, and/or extra
fittings to get the balance as close to perfect as possible and prevent burning out the motor or stripping the gears.
Here is the original schematic for the antenna rotator. It is totally electro-mechanical. Very old-school, almost primitive. On the other hand, it
still worked after decades of storage. One of the quirks of this old unit is that the motor in the rotator head runs on 24V AC. That made designing
a new control system for it challenging. I looked for ways to modify or automate the original control box, but couldn't figure out a way to make it work.
So I gave up on trying to use the old control box, stripped it of usable parts, and started designing something completely new.
Here is a schematic of the controller electronics I came up with after several iterations. Click the image for a larger view. the circuit
is based on the MBED rapid prototyping platform. The
MBED is basically a complete computer on a tiny module. It can be programmed in C using an online IDE. The MBED is quite powerful, has lots of IO capability and
just about every bell and whistle anyone could want. It is really overkill for this project, but I was familiar with MBEDs from using them in
projects at work. So it was my first choice for this project. You could easily substitute an Arduino, a Raspberry Pi, a PC, or a handful of
analog components to do the same thing. This is just the way I did it. Feel free to roll your own.
The code (software) for this project can be found at http://mbed.org/users/omegageek64/code/suntracker/. It is a fairly simple program. As I said above, the MBED is overkill for this project. However its untapped potential could allow for adding more features and functions in the future. A second motorized axis could easily be controlled. Battery management, charge control and temperature compensation could be added. Logging of power production and usage data could be added. The sky is the limit. Let your imagination and ingenuity run wild.
The electronics for driving the unit are housed in an old ammo box that I got at a yardsale for $5. It makes a perfect enclosure. It is rugged,
weather-proof, and plenty roomy. The box holds two 40 Amp automotive relays, a power inverter, a 120V to 24V step-down transformer, the
breadboard containing the actual drive logic, a fuse holder, and terminal blocks for wiring everything together, There are feed-through
terminal blocks mounted to the outside of the ammo box so that various signal and power connections can get into and out of the box.
Here I have begun mounting parts inside the ammo box. The relays, transformer, terminal strip and one of the feed-through terminal strips have
Here is a view inside the ammo box with all the electronics installed. The white breadboard with all the logic is on the upper right.
The long black rectangle is the power inverter. The breadboard and inverter are held in place with industrial strength velcro. A second
pass-through terminal strip has been mounted at the lower left. Everything has
been wired together creating a real rat's nest effect in the bottom of the box.
Here is a close-up of the breadboard with the "brains" of the system on it. The MBED computer module is on the right. Left of
the MBED are the two trimpots for adjusting the signals from the sensor head. Below them are the power transistors for driving the relays.
Further left are the manual override push buttons for moving the tracker manually. At the far left is the 9V voltage regulator.
The sensor head consists of two small thin-film Copper Indium di Selenide (CIS) solar cells of the same type I used in
my home-made folding 15 watt solar panel. I had several of these
cells left over, so it seemed like a no-brainer to use them as sun sensors in the tracker.
The two small solar cells are mounted at 90 degrees with respect to each other. The idea was that one cell or the other would get more sun, and the tracker
would move until they were getting equal sun.
Here is a view of the completed sensor head. It is mounted on a short piece of square aluminum tubing, which in turn will be mounted on the drive
pipe of the tracker. I drew on some dimensions for those who are always asking me to include them. The sensor head is held on with a hose clamp in
the same way the panel frames and limit switch bar are attached.
Here is a view of the sensor head attached to the tracker. The sensor head is mounted on the stub of pipe coming out of the top of the rotator unit.
This keeps it out of the way when mounting the panels, and makes it less likely to be shaded by anything.
The paddles of the switches engage with the heads of long screws protruding from the wooden support structure for the drive motor. The
limit switches stop the motion of the motor at both the east and west ends of travel. The switches are wired normally closed, and open
when the limit of travel is reached.
This photo was taken during a marathon testing and debugging session at my workshop the last weekend before leaving for Arizona. My netbook computer
is plugged into the MBED unit in the ammo box. A large, deep-cycle battery is powering the electronics and the tracker unit (not in the shot). I
needed to get the unit working, and then break it down and pack it for transport by Sunday night. It all came right down to the wire (doesn't it
always?). But it appeared to all be working well by Sunday afternoon. So I disassembled it all and packed it for shipping. Little did I know that there
was going to be a big problem when I got to Arizona.
Here is a photo from the above testing and debugging session.
The sensor head idea worked well indoors in my workshop. After tuning the pots on the controller, the tracker would follow a lamp quite well,
always staying pointed toward it, no matter how I moved it around. I was very pleased. I wanted to take the unit outside for some testing under
the actual sun, but there wasn't time to disassemble it, move it outside, reassemble it, test it, then disassemble it again. I had simply run
out of time. So I hoped for the best, packed it up, and trucked it to Arizona.
A solution to the problem was found by mounting an occulting bar in front of the solar cells, and applying black tape to cover part of the solar cells.
As the sun moves west the bar shades the east cell more and the west cell less, resulting in the necessary voltage difference needed to make the tracker
move and follow the sun accurately.
The first prototype occulting bar worked so well that a permanent occulting bar, made of 1/32 sheet aluminum, procured from the hardware
store in town, was fabricated the next day. This bar was made wider so so it would cast a wider shadow and I could do away with the tape
on the solar cells.
The occulting bar is mounted on two screws that allow it to pivot east and west. This allows for fine-tuning the pointing accuracy of the tracker.
with this bar in place, the tracker really began working well.
The occulting bar works great. Here it is late afternoon and the tracker is all the way over to its west limit after following the sun
all day. The unit is working very well. I couldn't be much more pleased with it.
Here the tracker is a little east of center late in the morning on a somewhat cloudy day. Even through thin clouds the tracker
works well. The tracker stops tracking when thick clouds roll in and the brightness of the sky tends to be fairly uniform. As soon
as they thin out though, the tracker latches back on the sun.
It gets really windy on my Arizona property. On any given day we can see gusts of 35 mph or even more in the afternoons. Something about the location, between
the mountains in the south and the desert in the north, drives a daily strong wind out of the south. It gets even worse if there is a storm. The
winds could easily blow over the solar tracker. So I take the precaution of staking it down. This photo shows wooden stakes at the four corners of
the tracker base to keep it in place. Once I decide where to permanently place the tracker, I will probably use steel stakes to hold it in place
because they won't rot.
UPDATE - I think I have found a cheap and easy way to weatherproof the sensor head. I sliced a 2 liter bottle in half and slipped it over
the sensor head. I had to cut a few slits in the bottom part of the bottle to make it slide around the square tubing at the base of the head.
Some clear tape holds it together. I can adjust the position of the occulting bar (if necessary) through the cap hole with a notched stick.
It will be interesting to see how well it stands up to the weather.
UPDATE - I have made some changes to the solar tracker. Firstly, as you can see in this photo, it has been painted to help protect the wood structure from the
weather. It is also now sitting on top of bricks to keep it up off the ground at least a little bit and prevent the wood from absorbing moisture.
The original wooden stakes used to hold the tracker down in high winds have been replaced with long steel stakes driven deep into the ground. Long screws
go through holes in the stakes and into the wood to securely anchor the tracker. A storm front came through with hurricane force winds shortly after I
made this modification. The tracker handled it with no problem.
A horizontal support bar has been added to help stabilize the panels and prevent them from flapping around in strong winds. Wire ties are threaded
through holes in the edges of the panel frames and loop around the bar. The wire ties had not yet been installed when I took this picture.
The horizontal support bar was attached by welding a 1/2 inch steel pipe coupling to the main 1 inch support pipe. Two 24 inch long pieces of 1/2 inch
pipe then thread into it to make the horizontal support bar.
UPDATE - I have built a new, weatherproof sensor head for the system. It is still based on two small, thin-film, Copper Indium di Selenide (CIS) solar cells set at an angle to each
other, but now it is mounted inside a clear plastic container. The container originally held a Bluetooth speaker I bought for use with my phone. I immediately saw a potential new use
for the container. I had also considered using peanut butter jars and Mason jars, but this particular container was just perfect.
The occulting bar is now located on the outside of the container for ease of fine tuning the tracking. It is held in place with a simple hose clamp. Once the new sensor head
is mounted on the tracking system, a bead of silicone caulk will be applied around the lip of the jar lid to seal it against moisture.
Here is a view of the sensor head with the jar removed. The original head had the two CIS solar cells mounted at 90 degrees to each other. Such an arrangement would not fit
in this jar, so I mounted the cells at a more acute 60 degree angle.
This photo shows the underside of the sensor head where a terminal block was attached so that wires could be terminated neatly. It also shows how the mounting foot is screwed
onto the jar lid. The mounting foot will be clamped to the main shaft of the tracker with a hose clamp.