Two-Axis Solar Tracker

Solar panels are an amazing piece of engineering, but without exactly the right conditions they can be pretty fickle. One of the most important conditions is that the panel be pointed at the sun, and precise aiming of the panel can be done with a solar tracker. Solar trackers can improve the energy harvesting ability of a solar panel by a substantial margin, and now [Jay] has a two-axis tracker that is also portable.

The core of the project is a Raspberry Pi, chosen after [Jay] found that an Arduino didn’t have enough memory for all of the functionality that he wanted. The Pi and the motor control electronics were stuffed into a Pelican case for weatherproofing. The actual solar tracking is done entirely in software, only requiring a latitude and longitude in order to know where the sun is. This is much easier (and cheaper) than relying on GPS or an optical system for information about the location of the sun.

Be sure to check out the video below of the solar tracker in action. Even without the panel (or the sun, for that matter) the tracker is able to precisely locate the panel for maximum energy efficiency. And, if you’d like to get even MORE power from your solar panel, you should check out a maximum power point tracking system as well.

32 thoughts on “Two-Axis Solar Tracker

  1. A Raspberry Pi? Well, at least it’s not an Arduino.

    Oh dear, HAD scraping the barrel bottom again.

    Only crap panels need to be pointed directly at the sun.
    You gain about 20% with a tracker, which in this setup is probably enough to run the motors & controller. Yay.
    When the wind breaks this, you’ll get 0%. There’s a reason twin axis is uncommon, gosh what could it be.
    Single axis is good enough, especially if it’s portable. You gain a couple of percent with a two axis system, big deal. Ask an astronomer where to point one axis.
    There’s a reason most commercial tracker are optical.
    This is simpler than optical? Oh that’s right, HAD editors don’t know what words mean.

    1. Lol, someone having a bad day? Also HaD editors stole my lunch money, used way more than their fair share of carbon today. Jay should bury his head in shame, how dare he not adhere to some unknowable standard for hackery set forth by the committee?!? Maybe he did this for fun? Maybe you could be nicer? Maybe he’s learning like everyone else and would be better off if people were friendly and helped him out instead of being so snarky?

      1. The project is fine, it describes itself as a different take on solar tracking, using words like “hobby” along with “bulky and very expensive”. Well well, using $100 linear actuators does that.

        The ‘editors’ have added nuances that the project author doesn’t claim – that’s the problem. If I was the author I’d be a little annoyed at that write-up.

        The fact that it won’t really perform much better (couple of % as best) than a single axis tracker is another matter. Single axis trackers are easy to set up too: point at sun, adjust vertical tilt, done. Adjust tilt monthly.

        As an aside, you can use the panels themselves as the optical sensor. Take two panels, and put a board perpendicular between them. If they’re not pointing directly at the sun, the board will shade one panel leading to a lower voltage. Swing the panels until the voltages equalise. Enjoy.

        1. This is a portable panel on a portable tracker. Having to polar align something everything you intend to use it sucks. That’s precisely why you see a lot of portable units out there have 2 axis. Also your assertion that it will use as much power as it saves is only true if you’re truly bad at your electronics design in which case I feel very sorry for you. Though that may explain why you seem to get a kick out of shitting on other people’s designs.

          1. It’s running a Rasperry Pi on a panel that appears to be no more than a 10 Watts.

            The Pi itself uses 0.52 W to idle, and the average production of a 10 Watt panel over a 24 hour day is about 1.5 Watts, so the control system uses up approximately a third of all the energy it makes. That completely negates all the gain.

          2. Of course you can put the Pi to sleep when the sun goes down, but that’s still going to be around 15% of your output lost.

            The gain you get from a tracking panel is on the same order as the energy required to run the system.

  2. Is there really that much benefit to have a second axis on a portable solar panel?

    I want to take a solar panel when I go camping. I was thinking that if it can track on one axis and I set it up with a compass or by simply looking at the morning sun then it should work fine. I wasn’t going to use an Arduino or even a 555. Just two small extra solar cells out of garden pathway lights to drive a transistorised DC motor server circuit. With the extra cells slightly offset to the main cell to see when there is some advantage to moving the main panel.

    1. If you have the needed space, is usually much cheaper and less troublesome to just add an extra panel than a tracking system. Most times results in even a higher power output and better realibility.

      That is why solar tracking systems are normally not used unless they are absolutely neccessary like in concentrator type systems.

  3. Wow guys, way to bash on a project and HAD. No one said this was gonna change the world. You guys set your standards too high. It is a cool project, done in a interesting way, and well documented. Nice job on the project mate. HAD keep doing what you are doing.

    1. No kidding! Why do people gotta poop all over someone else’s work?! If you see weaknesses in the design, be grateful for your insight and maybe share constructive suggestions. Many times the essence of “HaD” goes beyond just arriving at the desired utility. Sometimes the reward is in the exploration and taking the long way.

    2. You’re right, they are being a bit hard on it. But I think it’s mainly due to the HAD commentary rather than the project itself. This line in particular: “This is much easier (and cheaper) than relying on GPS or an optical system for information about the location of the sun.”
      Saying that this method is easier and cheaper than an optical system is simply not true. I’m assuming this what sent them into a rage. I’d like to point out that the author of the project did not make this claim. If the editors strike that part out, all will be well.

      A few seconds on google showed me this analog circuit: http://www.electroschematics.com/8019/diy-solar-tracker-system/
      Products similar to this can be bought if you don’t want to make it yourself.

      This system is certainly more advanced than a plain analog follower. It would be cool to see features like “Hail Avoidance” by automatically going vertical when the weather report looks bad.

      1. Quote: “It would be cool to see features like “Hail Avoidance” by automatically going vertical when the weather report looks bad”

        Yeah and some video motion detection … wanna see those solar panels batten those hail stones around like tennis balls!

        </sarcasm>

      2. Yes, when we studied analog electronics in school, a solar tracker was one of the examples of servo feedback. In theory, at least, you could build a two-axis tracker with a couple of 741 op-amps. Of course, it would be, in the splendid American phrase, as dumb as a box of rocks, but as long as the sun was shining it would point the solar panel in the right direction.

  4. It’s worth running the numbers before building this, especially due to the complexity.

    From here:
    http://www.solarpaneltilt.com/

    …we can see a fixed panel gives us about 71% of maximum, and a panel that is tilted twice a year gives us about 75% of maximum. You can translate these figures directly into how much energy you would get in return for the money you spent on the Pi and motors.

    This design uses green LEDs as photovoltaic cells:
    http://www.redrok.com/led3xassm.htm#led3xc3

    It compares the output of two of them to determine which way to drive the motor. The single motor is tilting the solar panel which is mounted on a single axle. The axle is oriented north/south, at an angle appropriate for the latitude.

  5. *kicking my vat of my brain jar of hyperthymesia*

    This isn’t a full criticism and hopefully someone will chime in to back up the correlating evidence.

    Someone DID a sun/solar tracker previously, dirt simple. I say dirt simple because it was valid and proven. The strongest IR emission in the sky was the sun, now I am uncertain if they had referenced the Python OpenMV lib to do it but in essence, it was a extrapolation of ensuring brightest point was at 0,0 on the x and y.

    (There might have been a brief mention of using OpenMV on a Android Phone to fire a automated Supersoaker at squirrels with water. Backyard Defense with Python Nod. )

    TL:DR; Using a cheap android phone, some servos someone placing 2 photographic IR filters (same alignment) on the phone lens camera and the simple .apk software was able to send -1, +1 or +1, -1 via logging wrote/dumped to a open ftp (or tftp – the details escape me )log file in the directory and the back-end would perform the servo adjustment correct itself every minute to focus the cross-hair on the brightest point. I think the resolution was like 320×240. but it was enough to have the thing pointed correctly at the sun. I can remember a BW image and a white large dot on a grey background and a “wow, that is elegant as hell”

    Now, I’m uncertain how accurate this was or how many times it locked up but I remember thinking, ‘Damn, I could make a solar forge with 9 mirrors and 9 Fresnel lens’ if I could figure out the additional code-base to merge those beams’ to a fixed point crucible. “I could make and pour Obsidian glass FFS!”.

    Hrmmm.. Based on what E-bay shows there is someone that seems to have obsidian blanks pre-prepared for knife-making.

    The patterns seem very modern and not to mention there is a strong possibility of having some graphite and Bentonite clay and Graphite molds being used.

    Bottom line, we could refined this with either a $h!ty sub $20 phone or 9 or 12 photoresistors in an array.

    “M-Morty explain to them wha*(urp)*t I’m talking about”
    ಠ_ಠ
    “20% percent right as usual Morty”

  6. Common LEDs can be used as light sensors. Two of them with a panel between so one will get shaded. Build a simple sensing circuit that is made to run a motor when the East side LED’s output drops, then stop when they balance. Also have it reverse the motor when both LEDs have no output (because it’s night) to move the panel back to the East position, ready for the sunrise. Of course you have a switch to stop it.

  7. The problem with small trackers are… panels are so cheap. At least in comparison with the cost and added complexity of a tracker. Hail avoidance, yes, thats a plus. but the added structural strength needed to protect against wind? Hail is probably better kept at bay with some chicken wire (I suspect less than 5% additional losses). One thing the tracker does give you is a longer solar day, this reduces your battery workload slightly.

    If you assume the hardware costs at least $40 (and were no even including the electronics). Then you would have to have a panel where 40$ represents 20% of the panel cost. Otherwise, just buy more panel right? ie a panel of at least $200. Where I live that buys you a panel of over 250W. Its huge, and the mechanics will then certainly cost more than $40.

    All that being said, things that move sure look good.
    Murray.

    1. And this is why you dont see anyone using trackers anymore. I pay $1.49 per watt of panel so the cost of a tracker that can handle being outside in the weather all year long is more than buying 25% more panels.

      I am buying 200 watt panels for $310 each. Plus trackers you lose the efficiency of putting all the panels on the roof to give you a huge chunk of saving in shading your whole roof and creating a heat chimney effect under them that will draw in cool air up and across the roof. I saw a HUGE decrease in AC use from that.

    1. There’s no difference between tracking the sun and tracking a star, they move exactly the same.

      Photographing a star required a controller even dumber than the box of rocks [Stephen] mentions. Your inclination (vertical tilt) depends on latitude and time of year (which can be ignored for solar). Horizontal motion is done with a motor geared down so far it will spin the telescope exactly once per day. They used to be clockwork, these days just little gearhead steppers run off a 9v battery (for low end scopes).

      Scopes that jump from star to star work by being told where one star is, from that they calc where the rest are.

      We’re not bashing the project; the builder will figure out it’s a bit of a waste of time eventually (drop the tracker and get another panel) but they will learn something. What we dislike is the breathless and clueless embellishments added by the HAD ‘editors’.

  8. I posted links to two sun position algorithms/papers/code, but the moderators will not post it.

    Google “sun position algorithms”. Look at nrel (dot) gov. There is a spanish national lab with the initials PSA – www psa (dot) es (slash) sdg (slash) sunpos (dot) htm – a simpler algorithm that is as good as the tuned nrel code

  9. To bring in enough power we either need to improve the efficiency of our panels or find ways of getting more from our current solar panels. Every panel you see in your day to day life is in a fixed position, most likely facing south at a 45 degree angle.

Leave a Reply

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