2023 Hackaday Prize: A Reinvented Solar Tracker

It probably goes without saying that solar panels need to be pointed at the sun for optimal performance. The tricky bit is that the sun has a funny habit of moving on you. For those with a solar panel on their balcony or garden, mysoltrk tracks the sun to get the most out of a small solar panel.

[Fulvio] built the tracker to be solid, low cost, and sturdy enough to survive outdoors, which is quite a tall order. Low cost meant WiFi and GPS were out. The first challenge was low-cost linear actuators that were 3D printed with a mechanism to lock the shaft. An N20 6 volt 30 RPM geared motor formed the heart of the actuator. Four photo-resistors inside a printed viewfinder detect where the sun is, allowing the system to steer the array to get equal values on all the sensors. An Arduino Nano was chosen as it was low power, low cost, and easy to modify. A L298N h-bridge drives the motors, and a shunt is used instead of limit switches to reduce costs further.

There are a few other clever tricks. A voltage divider reads the power coming off the panel so the circuit doesn’t brown out trying to move the actuators. The load can also be switched off via an IRL540n. As of the time of writing, only the earlier versions of the code are up on GitHub, as [Fulvio] is still working on refining the tracking algorithm. But the actuators work wonderfully. We love the ingenuity and focus on low cost, which probably explains why mysoltrk was selected as a finalist in the 2023 Hackaday Prize Green Hacks challenge.

35 thoughts on “2023 Hackaday Prize: A Reinvented Solar Tracker

    1. I think solar trackers are great as a supplement. I have a fixed system, 10kW if solar for now. I have plans to make it bigger but I’d also like a ln array of around 1000-2000w on a dual axis tracker just to get my system started earlier and ending later. Though I might just use some fixed vertical bifacials which seem to be almost as efficient as dual axis tracking.

  1. Two questions:
    1. If it’s powered solely by the sun, how does it go back to face east in the morning?
    2. What fraction of the day’s energy ends up powering the microcontroller, motors, sensors, and the quiescent draw of that buck converter?

    1. If you’re not totally space-restricted another panel or two is always cheaper than some form of tracking. Tracking always means mechanical parts which can and will fail and need maintenance.
      I did something similar as a university-project, too. So I like the idea of tracking the sun myself, but it’s just not with the hassle.

    2. I can hazard a guess…the Photo resistors would be facing West at sunset, and remain in this direction till sunrise which will cause the panels to start moving to the East. This move should be timed so that when start up voltage is reached, the panels are facing the sun. Since the motion is quite slow, the power does not need to be very high, and the net result would be more power generated from the same panels if immovable.

  2. Add a steel cable from base to main panel structure.
    So when it breaks, the wind doesn’t send it flying into some kid’s skull.

    Very basic financial/power breakdown.
    1 m^2 panel.
    20% efficient panels
    Perfect tracking, no obstructions, average 12 hours/day of power.
    Done someplace frozen, dark and evil as Mordor…MooseJaw. Where solar normally makes no sense at all. 2 hours average solar equivalent for fixed panels.
    No heat required to keep unfrozen.
    $0.15/kwh power.

    These are very generous assumptions…power price for tundra might be way too low.

    Net 10 hours average gain from tracking.
    2 kwh/day net gain.
    Roughly the same as just putting up 6 similar panels set for latitude.

    $0.30/day. $110/year rough value of power.
    That doesn’t cover a tech with grease gun to lube and inspect, much less repairs.

    Conventional wisdom remains: ‘nothing beat cubic inches except cubic dollars’
    Sorry wrong one. remains: ‘buy more panels and nail them down’.

    1. Trouble is, the fact that you make the most money by hogging all the panels, pointing them due south, and raking in money at noon while letting some other guy or a battery pick up the load at all other times means people think there’s something inherently wrong with the panels that they can’t work at any other time of day. When actually they work at whatever time they’re lined up with the sun and not shaded.

  3. I remember back in the 70’s seeing a completely mechanical tracker – I think they used air pressure to do it – probably 2 air tanks with one face painted black, angled slightly from the solar panel, and whichever got hotter produced higher air pressure, thus moving the ram until they equalized.

    But I still respect you kids and your electric doodads :-)

    1. I remember that one. It was two cross-connected propane tanks, painted black and half-hidden under shade on the east and west edges of the array. Heat from the sun boiled a bit of propane on the hot tank, pushing liquid to the cold side, making it heavier and tilting the array. No electronics. Genius.

    2. Interesting. Clever. Looks to be a single axis tracker, which is better than fixed but not as good as possible.
      I wonder how it behaves in a strong wind?
      Freon might be a poor design choice now.
      Given the cost of solar panels today, I suspect a fixed pair facing westwards and eastwards would still be a cheaper and more reliable solution.

    3. It’s amazing how much heat you can get from something painted black in the sun… Or even something sitting in the sun..

      I shifted my hot water system tank so that it was on the east side of the house instead of at the back (in the shade) – mainly to get it off the back porch and close to the main water usage.

      I forgot to turn the power back on after moving it (it’s on it’s own circuit). We didn’t find out for over a month, as the sun was enough to heat the tank each day (in the tropics)…

  4. True. You can dispense with the second axis if you can persuade the sun to travel along the equator instead of the crazy ecliptic doing that radical 46 degree swinging up and down between summer and winter.

    Though it might be simpler to just re-orient the earth so the north pole points to the Cat’s Eye nebula instead of Polaris. Bonus: It’s much cooler-looking.

    1. comment glitch. this was in reply to Darjac’s comment above: “A single axis of rotation is necessary to compensate the rotation of the earth. And at constant angular speed. cf equatorial astronomical mounts.”

      (HaD, when are you going to fix all these commenting system issues?)

    1. Neither BTU nor kWh is a measure of heat rate.

      What is the question you have been asking for years?

      If you’re asking what is the thermal power available from solar radiation, a good approximation at moderate latitudes and low altitude is 1 kilowatt per square meter in direct sunlight.

    2. If you’re using sunlight for heating and choose appropriate materials and design,
      you can obtain high efficiency heating of water or air. For instance, skylights can be good during part of the year in some climates. Maybe you’ll get 3000+ BTU/hr per square meter.

      If you’re trying to heat something by sunlight and then turn that heat (differential) into electricity, you’re fighting an uphill battle. Seebeck devices have miserable efficiencies. Mechanical devices will be complex and inefficient. Carnot is not your friend.

      Best quality photoelectric cells under optimum conditions can turn the sun’s radiant energy into electricity at efficiencies of 23% (silicon) to 44% (concentrated exotic multijunction). Your results won’t be that good.

      Let me say in good faith: you should be doing your own learning on this subject, not asking the same ill-defined question repeatedly.

  5. I have designed a simple, inexpensive, mechanical device which can track the sun from morning till evening. If anyone is interested to check the device please contact me on my email kvkishore1952(at)gmail(dot)com.

  6. Thank you for sharing this very interesting and novel design.

    In order for solar tracking to work, there must be a financial benefit. i.e. One expects the tracking solution to produce a better financial solution than a stationary one. Cost of build/operation vs. profitability.

    The financial benefit of moving the panel(s) must exceed the cost of this mechanism. If the panels are to be rotated 180° (e.g. East to West) during the course of the day. A 1m square panel will need 1.4142m area to rotate within (3ft 3″ vs 4ft 7″ ). This is the corner-to-opposite-corner distance to prevent them from colliding. Therefore, increasing the land you need to house the panel(s) by 41 % or reducing the number of panels you can house on the same piece of land by the same amount.

    This cost is incurred even before the more costly construction takes place.

    When one concedes, that the above solution may be inadequate for moving large solar panels or solar panel arrays, then one may divide the problem, as I did, into three parts;

    – The first, being locating the sun at a given place and point in time (mathematical).
    – The second, being the actual movement of the solar panel in response to the earth’s rotation (mechanical).
    – The third, being the execution of the previous two (electrical).

    When I began looking at the theme of solar tracking, I came to somewhat different solution.

    Given $4; a $1 compass, a $2 GPS module and a $1 microcontroller, I can virtually locate the sun with ease using a (free) software algorithm.

    Now I can calculate the position of the sun and convert this into X-Y-Z co-ordinates for the motor movement. The solar panels are no longer “solar-tracking clever”, but dumb.

    The movement of a bearing-mounted solar panel or solar panel array could be initiated using the same microcontroller.

    Optimization: Further bearing-mounted panels could potentially be attached mechanically to the first (a linkage arm) in order to save cost, given that the motor rotating the first was substantial enough to handle the load.

    Which questions the need for “solar trackers” in the first place.

    Sorry, if this sounds negative. It’s not my intention. I merely wish to explore alternative solutions. Which is probably why you’re here too.

  7. You kids with your microcontrollers!

    I did this in pure analog with two CDs resistive photosensors and a h bridge driving a single motor (single axis, clearly). Won second place iirc in the Santa Clara County science fair in… 1980 ish? I was still in grade school. (and had help from my hardware engineer dad)

    It’s not a difficult concept to design and build. Making a device that can work for years reliability is.

  8. What this individual may have failed to realize is the sun travels a regular path in the sky and you would really only need one axis of rotation. Also it travels at 180° per 12 hours, so all you technically need is a gearbox capable of traveling this speed and reversing course overnight (and in the morning).

    As far as “tracking” goes a PID loop with a photocell in a tube is more than enough ( if it is even necessary). In practice you can just water cool the cells to get better efficiency from them and also add 10-20% more panels.

    If this is for a mobile application with constraints on how many panels and their location then so be it.

    Note: you can pretty much copy plans for a telescope mount that has “automatic tracking”. They will explain the angle to put your axis on, and the motor control necessary to follow a fixed point while the earth rotates.

    I am glad to see people hacking along and trying things, but also sad that it is hard to get people on board with thinking logically about what they are accomplishing, or checking that people have been building these trackers for more than 50 years (specifically for solar panels, much longer overall) and didn’t look up any instructions or plans, just took off in a vacuum and invented Johnny 5’s head.

  9. Trouble is, the fact that you make the most money by hogging all the panels, pointing them due south, and raking in money at noon while letting some other guy or a battery pick up the load at all other times means people think there’s something inherently wrong with the panels that they can’t work at any other time of day. When actually they work at whatever time they’re lined up with the sun and not shaded.

  10. The lowest tech solution to tracking with a solar panel was by a friend at AfrikaBurn. Panel was on a hinged frame but off balance. Counter balanced by a 2 l bottle of water. Small hole in bottle lets the water out slowly so it gets lighter. Panel moves correspondingly.
    Reset and refill bottle in the morning.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.