If you need to squeeze every available watt out of a solar panel, you’ll probably want to look into a solar tracking system. Unfortunately, they are usually quite large, heavy, and expensive. As an alternative, [JP Gleyzes] has put together a DIY solar tracking system that aims to address these issues.
Starting with a 100 W flexible solar panel purchased during a Black Friday sale, [JP] first created a simple frame using 20 mm PVC tubing and a few 3D printed brackets. It mounts on a wooden base with a printed worm gear rotation mechanism, powered by a stepper motor. The tilt is a handled by a lead screw made from a threaded rod, connected between the wooden base and the top of the solar panel, and is also driven by a stepper motor.
For even more efficiency, [JP] also created an MPPT charge controller with companion app using an ESP32, modified 20 A buck converter, and current sensor module. The ESP32 also controls the stepper motors. The optimum angle for the solar panel determined using the date, time, and the system’s GPS position. [JP] had also created a simple Android app to calibrate the panels’ start position.
This project is a finalist in the Planet-Friendly Power challenge of the 2022 Hackaday Prize, and all the details to build your own are available on your project page. Looking at the size of the system, we suspect future iterations could be even smaller.
Amazingly well made project !
Is it worth spending 55€ for a 100W rated panel (compared to a fixed optimized setup) ?
At 20c€ a kWh in France, 55€ represents 275 kWh. So given a benefit of 20% more energy (so 20W for this panel), it would need to harvest 13750 hours of sun, on an average of 6 hours a day in France, it would require ~6 years.
My panels give 130 Watts each at their best, and I was looking for such a detailed project to improve my solar production … What scares me is the wind that may take my panels out of my roof, so I would personally add a wind sensor to retract them … But given this well documented project, it shouldn’t be that hard !
Thanks again for sharing !
I cannot agree more : it’s currently way cheaper to place more fixed solar panel than to move them.
The only exception is when place is scarce ( like boat, satellite, etc…) but then environment is so harsh on the fixture, placing more solar panel still make sense.
For solar home setup (4kWp) solar panel are only half of the total hardware installation, the other half is rail, wiring, panel, inverter, etc… Solar PV module are now pretty cheap, only 160€ for 405Wp
If place is scarce you mostly use higher efficient panels
55$ ils for the whole system (not the panel)
This price includes the mppt controller and the tracker 😎
I think everyone is missing the main point when someone does a tracking system ! It’s not the size of the panel that is the issue. A tracking system is intended to increase the output of whatever panel is used and the design is NOT limited to just that one size of panel, it can be adapted to ANY size panel or array ! So once a system is working, take that design and mod it to work to move whatever array you wish to move. Figure out how to do it as cost effectively as possible and do it. Extending the solar window is still the most effective process. It puts the output at near peak as long as possible, instead of 5 hrs. it moves to 9 average which is almost double, in some place it can be longer, like 10 to 12 or even 15. Plus it requires less material and space as far as the solar system not counting the tracking system.
+1 Bret
I quoted this comment into my “market analysis” : https://hackaday.io/project/185105-low-cost-solar-panel-solution-mppt-sun-tracker/log/210311-a-light-market-analysis
I will check out the link. I currently have a MPPT system installed for my RV which is my full time residence. But as was mentioned by someone else, my space is limited so I need to extract every watt, watt hr. I possibly can for what I have. I agree since I know I can purchase panels for ~1 $/w and by the time a tracking system is assembled and the cost of said tracking system is finalized, another panel would make up for the difference however since adding another panel would also require additional cost for the interconnect pieces, not much but still more than what I have already invested in for the present arrangement, then the additional hardware to mount that panel, so in the end, I would be ahead to spend another say $300 dollars for a completely finished 200 watts, I still feel having another 2 to 3 hrs./day at what the cost of building a tracking system would set me back and then I don’t have to worry about the space limitations near as much, or having to acquire more and different connection necessities. I personally have had many years (18) of off grid power production and could go on and on about what I have done over those years but it goes way out of the scope of this post so I won’t. People can analyze all their lives and still not dictate everyone’s circumstances. It comes to what the person doing the system feels they need and what they are required to work with. If you have room and can afford more panels by all means that’s the way to go because no moving system is always the simplest and most effective as far as keeping things working, mechanical stuff fails every time and usually does so in the most inconvenient place, like facing the wrong way from the sun, like the solar water heating system I was around some years ago. I give those against the tracking system hands up because like I said none mechanical stuff is always more dependable. I spent nearly 50 years as an auto and truck mechanic so I know something about that too. Right now I’m on the fence about the tracking system for my array, I want it because it will make a bit more power over the stationary system but it will be something to have to deal with if it fails also. The cost of the linear actuators I will require is high in comparison to another 100 to 200 watts but compared to the cost of installing those other panels, I think it balance out pretty close if I build the system myself. I will NEVER make the choice to purchase a ready made system of any design for any reason over a stationary arrangement because of the cost/added wattage, it surely doesn’t add up economically ! It becomes an issue of being able to say, see that, I made that and made it work!
looks good as long as it never rains.
This looks so flimsy that a gust of wind is likely to relocate this device from where you installed it to where you can’t find it.
This appears to be an XY system. It may be cheaper.
However mechanically simpler is to mimic the axis of the Earth’s rotation. An ‘Equatorial’ mount.
While true the earth does wobble on its axis, still most of the benefit is received by tracking the sun across the sky, and you could adjust it once or twice a year for the amount North and South.
XY? No, it looks like azimuth-elevation, but your point is still valid. Still, the amount of power required to change both azimuth and elevation is more than made up by having a similar mechanism. For an equatorial mount, you still have to be able to change elevation to adjust to the season. So you still need that second axis of freedom, even if you don’t give it a motor.
I think of “Earth wobble” as Precession, not the progression of seasons.
The apparent movement of the sun to the north (in the northern hemisphere) in the summer and south in the winter is not due to precession, but to the axis of the Earth’s rotation not being aligned with the axis of its orbit around the sun. It does also precess, as all gyroscopes under external forces do, but this is at a much, much slower rate, currently around 25,000 years. This is what is sometimes called “wobble”, and it causes the constellations to move in relation to the seasons, and for different stars to become the “pole” stars.
It’s indeed an Azimut/Elevation system
I don’t want to badmouth JP in any way here – what he did here is neat and inventive.
That being said, I’m always curious about how much power goes into running the tracking system on actively-tracked arrays. In JP’s case, it’s probably on the order of, say, 10 watts or so for the control electronics and drivers and motors to hold torque on what look like one 17 and one 23 sized steppers.
That’s got to be a good fraction of his output gain over leaving them in one place.
On more “commercial” systems which have to be rugged enough to be installed and left to run on their own for years in places that will see them exposed to some very bad weather and high winds once in a while, the mechanics have to be much more robust and the losses probably much higher. And, of course, any mechanism at all is much more expensive than a simple fixed frame.
I never really see good data on this. Is it really worth it with modern panels, or is it just more practical to optimize for one season or time of day where you expect to need power most, and just take the loss the rest of the time?
When you consider how often you actually need to move the panel, I would think that the positioning system could be powered up for a few seconds every half-hour or so. The rest of the time, the power supply for the steppers can be shut off completely, and the microcontroller put into a very low power mode without losing track of time. So if it pulls 10 watts for 10 seconds out of every 30 minutes, that’s 10 W * 10 sec * (10 / 1800) = 0.55 W. This is way conservative – in reality, the motors don’t need to move far enough to need to be powered more than a couple seconds each half-hour. Also, on cloudy/overcast days, there is little advantage to moving the panel, since the light is coming from a whole hemisphere, so it only needs that half watt when the sun is out and the panel is providing 100 W or so.
“…the motors don’t need to move far enough to need to be powered more than a couple seconds each half-hour.”
Except at sunrise when the panel must rotate a (relatively) large amount from where they were pointing at sunset. :-)
I can’t believe my reply is in moderation! Oh, I see, I used a word that is .. um, problematic?
“conservative”?
B^)
And yet it allowed you to say it. Maybe it was the math. The Internet hates math.
I do agree with most of your comment!
However power consumption was key specication for me also.
The motors are idle most of the time!
The MCU is deep sleeping most of the time.
Trust me you don’t lose that much energy !
I built a single axis solar tracking 2400 watt gridtie system in Madagascar 4 years ago. Surprisingly, it has withstood 2 cyclones in the last year without damage. The tracking system had two mini solar panels and moved the panels towards the panels the stronger light source. I used Chinese linear actuators for the project.
I say Bravo! And I should give some precision. In 1974 at the CNRS that is french organization for science research we where in the process of designing cheap heliostat for the solar CSP project Themis, the first in the world to be truly operational. We proposed a heliostat with the same principle as presented here. That was clearly the cheapest way to industrialize solar tower project. You may find our patent under the name De Flandre, Matarasso, Traisnel now in the public domain.
May this proposition very clearly presented here give a new life to that design!
Merci Pierre !
BTW, I am working for CNES (French Space Agency). The heliostat was a project I made 2 years ago Mostly same code as this sun tracker. If you look for “heliostat” into the ESP32 code you will find the 2 lines which differ !
Hi all,
Thank you for all your comments !
I came to the same conclusion for the ROI of this project. It must remain in the target price of 1/3 of the whole system to get the same ROI time. Which should be more than 4 years… So very very long.
However I wanted a fully portable system :. “set it and forget it” and I got it !
A key spec for me was too get the most power from this system. Together with the mppt controller out is also achieved 👍
Regarding the heliostat, I first made one (almost sale code) before jumping to this project. Amazing coincidence!
It’s really cool how “complex” projects can be implemented nowadays !
If you like “trackers” you should also like my other project : https://www.thingiverse.com/thing:4625420
Let’s keep fun in our life and take benefit of open source solutions !
JP
Have look to my “market analysis” here : https://hackaday.io/project/185105/log/210311-a-light-market-analysis
Imagine a spiral on the surface of a sphere, have the normal of the PV array surface traverse that path while you monitor the power output, the highest level shows the position of the sun. You can probably do that with a few 555 timers and a ROM.
There it is.
I was wondering how he calibrated the system regarding its physical position and the software’s idea of its position. While he initially had limit switches in the design, instead he just starts off the system by manually adjusting it to face the sun (with the help of some phone apps). I suppose that’s effective enough since great precision isn’t needed here.
Yep you got it!
I wrote the App, it’s already available for download. A newer version will come very soon.
details here : https://hackaday.io/project/185105/log/210034-automating-the-solar-tracker-initialization
I already won that contest. I raise and lower the axis of rotation by hand once a week, and every morning I wind up a spring in an old clock that rotates along with the sun all day , then do it again the next day. It’s done before I finish my coffee in the morning. The first one used a Bungie cord and a clock.
nice but you should be there every morning :-)