Self-Powered Sun Tracker Takes a Cue from NASA Solar Probe

Getting a solar array to track the sun has always been an interesting problem, and it has led to some complicated solutions. Controllers that use GPS and servos seem to be much in favor these days, but as this NASA-inspired sun tracker shows, the task needn’t be overly complex.

It’s pretty obvious from the video below that [NightHawkInLight]’s solar tracker is just a proof-of-concept for now, but it certainly shows promise. It’s based on NASA’s sun-skimming Parker Solar Probe, which uses sensors at the rear of the probe to maneuver the craft to keep sunlight from peeking around the sides of the shield. [NightHawkInLight]’s design simplifies that scheme even more, by using solar cells as the four sensors. The cells, mounted behind a solar shade, are directly connected to small gear motors that control azimuth and elevation. When a cell sees the sun, it powers the motor that moves the panel the right way to occlude the sun again, thereby cutting power to the motor.

[NightHawkInLight] mentions the obvious problem of what happens when the sun comes up and the array is pointing the complete opposite direction after the previous sunset, but we’re still not sure his solution – a larger array with tracking cells mounted further apart – will work. We’re also not sure how it will scale to larger arrays that need bigger motors to move. We’ve seen such arrays handled with more complicated trackers, of course, but we hope the simplicity of this design can be made practical for real-world use.

30 thoughts on “Self-Powered Sun Tracker Takes a Cue from NASA Solar Probe

  1. Just add another panel on the back of the “turn east” panel pointing the opposite way i.e., backwards and wired into the circuit in the “turn to the east side”. That way the backwards panel will “see” the sun behind it in the east in the morning after closing pointed west the previous sunset and will start turning the array to the east. Once the front sensor starts to see the sun it will shade the backwards facing panel and continue on to face the sun. The backwards panel may have to be tilted slightly to keep the array turning east when it faces due south. Don’t want a dead spot there although the sun will move west in the sky and pass the dead spot eventually.

  2. I think an angled mirror on the back of the main solar panel would redirect the array at sunrise, I don’t see why it wouldn’t scale up, treat this as the master unit and the rest of the array as a or multiple slaves.

  3. I remember looking into this a long time ago, and there was a hydraulic version that used sun-heated fluid to do the same job. This is a great project, but one that used no electronics at all would expand our horizons a bit. (there might be a pun there)

    1. The sun-heated fluid thing (basically wax-motor vs spring) is a neat idea and very simple/reliable, but it does have some (linear) tracking error that varies throughout the day, though that could probably be eliminated by rotating the sensing part at a higher gear ratio than the solar panel.

      The hydraulic/wax-motor approach has the neat property that it automatically returns to east overnight (cold), with the unfortunate side-effect that it starts turning eastward when a shadow passes.

    2. Same, but with actuators to compress instead of heat. Also looked into a 4xCDS cell on the panel corners that used a lens hood of sorts to shade the cells.

      The innovation here is the greatly reduced costs of panels that may allow direct power whereas over a decade ago I never got my own ideas to work due to concerns over long-term battery costs and no $$ to take the ideas any further.

      As for the dual axis, isn’t that a bit extreme, unless you absolutely must get as close to 100% conversion as possible. Find a chart and pick the angle optimal for all seasons for your location and lock that in. Only worry about the east/west axis from there.

  4. Some older spacecraft used light sensitive diodes in an angled baffled arrangement to sense the sun. Wider baffles at the edges, finer baffles when near. Surprising how few diodes you need to get close. On spacecraft the computer handled the motors but I don’t see why you really need the computer for this application.

  5. This is very nice, but there are two obvious methods to solving this.
    One, set a default “end stop” to the array. When the array hits it, set a timer for say 30 minutes, then re-orient the array to the sunrise position.
    Two, solve the problem. You can determine your lat and long, and therefore you can determine what the optimal orientation is for every single day. So run your array using a calendar timer. Even with something ridiculous like 50 adjustments per day, you only need a lookup table of 18000 entries to determine where the array should be at any moment. You could even burn it into an EEPROM.

    1. I love the idea of pre-computing the angles and burning them into EEPROM, but it adds a layer of complexity that the OP is presumably trying to avoid: a micro to read the eeprom and a (stepper/servo) motor that knows its position.

      Because once you’ve got the micro in the system, you can just calculate out the angles. It’s not timing-critical, right?

    1. Another clever idea is: 2 LDR’s + NE555 controlling servo.
      2 LDR with 45 degree angles chanding PWM signal of NE555
      and this will change servo angle.
      This is also verry clever and simple solution but not self powered.

  6. Back in the late 70’s Mother earth news has a DIY single axis solar tracker ( cant finD it on line) using 1) photo transistor, 3) relays and an SCR, and 2 limit of travel switches. You had to fabricate a shadow box for mounting the transistor and it did require 12 VDC power but was really simple. I built a version that didn’t require the SCR which was for making the unit return to the morning start position.

  7. You don’t need[*] 2 axis control to track the sun. Do what astronomers do, and use an equitorial mount, then you only need to move on one axis [old timers needed nothing more complicated than a clock, but that’s a bit open-loop]. Astronomers long ago figured out that an alt-azimuth mount is unnecessarily complicated when it comes to tracking celestial objects.

    [*] You do need to adjust the angle of the equatorial mount to cope with the change in solar angle through the seasons, but that only needs to be done one a month or so.

    1. That’s wonderful. Then I’d only need one drive mechanism instead of two and then adjust the “elevation” by hand now and then. To be fair though, with an alt-azimuth mount, you can mount miniguns to the side and frighten off sales people.

  8. Why don’t we see Fresnel lenses used to concentrate light onto PV panels and to reduce the range of tracking movement required. For some locations, would an appropraite lens catch enough more energy such that within a day it comes close to what is obtained from a tracking same sized panel as the lens is illuminating?

    1. They’re out there, but concentrators have little point unless you’re willing to use higher-efficiency PV junctions and spend the money to cool them. Higher efficiency junctions cost more per area, so concentration makes sense there, but the lenses are not free either. It’s a tradeoff: do you want 1m^2 of cheap PV panel, or 1m^2 of lens and 0.1m^2 of expensive panel and a massive heatsink, and how long does it take for the extra complexity and cost of the latter to pay off?

      https://en.wikipedia.org/wiki/Concentrator_photovoltaics

      Also lenses are very sensitive to incident angle. If the beam coming out of the concentrator misses the little junction then it’s pretty counterproductive, so you often need *better* sun tracking with a concentrator than if you’re using fixed mounts.

  9. Why move the photovoltaic panel at all? Build a coelostat/heliostat/siderestat with a mirror that moves to reflect/concentrate the light on a stationary panel. There is all kinds of software available these days, you could probably control it with an old cell phone or a raspberry pi device which could be powered/charged by a small separate solar cell if not by a conection to the main panel.

  10. When i was younger I made a pin hole camera out of a film canister put four LDRs in the bottom with some black card tubes around them to increase selectivity. It worked pretty well as a directional light sensor. I even experimented with coloured filters but I could never get my hands on more LDRs back then.

  11. The one thing you don’t have in space ( clouds, fog ) is the one thing that defeats this design on the earth’s surface. Clouds and fog will diffuse the “sun” signal and cause the mechanism to lose track. I worked at a solar power plant in 1984 that used a similar sun tracker design except it was 4 platinum RTD temperature sensors. The problems with maintaining a proper sun track caused other more severe secondary failures due to an uncontrolled focused sun beam that was about 1500 degrees. Unless you have a way to keep on target – eventually the sun position moves outside the correctable error band and you have to manually correct position to be on target. The problem gets really severe with fog.

  12. To track satellites we used a system that had four antennas surrounding the main antenna. When all four antenna have the same received signal, you are on target. If an antenna has a higher signal, you move toward that antenna until you get a balance. This has been used for years and you could do the same with a solar panel. If you balance the signal across four aiming cells, you must be on target or you are 100% shaded. Either way you stop moving.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

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