Sun Tracking for Solar Power – Part 3 – The Results

Sun Tracking for More Solar Power – Part 3 – The ResultsIMG_7884

SwitchDoc Labs has built a lot of solar power systems over the past few years.   Project Curacao,  SunAirPlus,  WeatherPi and the recent Solar Powered ESP8266.  We have fooled around with sun tracking systems, but we have never built one all the way out and then gathered the data to figure out if it was worth it.   Now that the sun has returned to the pacific northwest, it was time to do it.

In this series of postings, we are going to show you how to build a simple solar tracking system using a Raspberry Pi and a stepper motor.   The purpose of this project is to verify experimentally the gain in power from a solar panel from using tracking versus a fixed solar panel.

All the graphs in this series of posting are done using MatPlotLib on the Raspberry Pi.

There are four parts in this series of postings.

Results of the Sun Tracking Experiment

Figure 1 and Figure 2 show the current (mA) measurements for the tracked and untracked panels on March 31st and April 1s.   To calculate power, we need to pull the solar panel voltage and multiply it by the current (Figure 3 and Figure 4).

SunCurrent-9536
Figure 1 – Current (mA) on March 31st

 

SunCurrent-11274
Figure 2 – Current (mA) on April 1st

 

Solar Power Results

Figure 3 and Figure 4 show the power results from the March 31 and April 1st run.   There were intermittent clouds during the day on March 31st (Figure 3), hence the up and down of the data.   Doing the math on the power generated by the two panels on March 31st we end up generating 21.88 Watt Hours from the tracked solar panel and 17.52 Watt Hours with the untracked panel which gives us a 24.9% increase by tracking.   Pretty close to our estimate of 30%.   The test on April 1 (Figure 4) gave us a similar number of a 23.8% increase.

Figure7
Figure 3:  March 31st Power Curves
Figure8
Figure 4: April 1st Power Curves

One issue is evident from looking at our test data.   Looking at figure 3, you can see that the tracked panel peaked at about 11:30 where the untracked panel peaks at about 12:45 which is actually about local noon (when the sun is highest in the sky).   We are off about ~16 degrees or with our tracking software.   Easily corrected in the software to get even better numbers.   Just eyeballing the graph in figure 4, we could see getting another few percent of improvement by fixing that making it even closer to the 30% expected.

Regarding Power Usage by Stepper Motor

We connected up a Grove INA3221 I2C Current Measurement board to the Stepper motor controller (Grove Mini Motor) and we drove the Adafruit 5V Stepper motor we used in the SunTracker system.  We just love building prototypes and measuring things with the Grove connectors. It’s all very simple (we did modify a 50cm Grove cable to allow us to put jumper pins in the VDD line so we could measure how much current is going through the cable).

Based on our resistance measurement of the wiring in the motors (55 Ohms per winding – the Specification said 42 Ohms), we expected an idleIMG_7775 2 current draw about about 180mA. We measured about 170mA with the windings energized. If we used the motor this way all day, we would burn up ~11Wh. That would be far more than the ~4Wh we got from Tracking.

However, that is not how we are using the stepper motor. We turn 2 steps every six minutes. There no real torque on the motors from the panel just sitting on top of it (ignoring wind!) so we can turn the windings off for a huge power saving. Assuming we turn the motor on for 2 seconds out of 360 seconds, instead of burning 11Wh, we burn 0.061mWh which is less than 2% of our savings. So, in this system, it looks like a good trade-off. The I2C Motor control itself only takes about ~0.6mA.

We could tie down more of our assumptions (like measuring the maximum wind torque on the solar panels), but just feeling how hard it is to turn the motor when it is off, we think with this system we are covered.

Conclusion

Strictly from a power point of view, it is clearly worth it.  28% more power from the same cell.  There are trade-offs to be made and to consider.  You have added cost and reliability issues by adding the stepper motor and the mechanical linkages.   Why not just add a second solar panel?   The problem of that is again how solar LiPo chargers work.   You are limited to how fast you can charge a given LiPo battery.  SunAirPlus limits the charge to 1000mA for example.   If you put more solar panels on the system, you will increase the charge during the lower parts of the curve, but at the peak of the day, it doesn’t matter how much power is available to the charger, anything over 1000mA is wasted (and you will see the solar panel voltage go up as we did in Part 2 of this series).

This was a fun project and it was great to see that we could get pretty close to the theoretical value with some simple hardware and a Raspberry Pi.

Next?

We will be publishing a time lapse video of the experiment.   Not  much science in the video, but very cool.

Figure9
Figure 5:   Sun bearing down on the panels

5 Comments

    • The fixed panel current exceeded the tracker because of a mis-alignment of the tracker (off by about 12 degrees or so – about an hour of sun time). Good eye!

      Until we saw the graphs, we didn’t notice this in the data. It popped out in the graphs!

      SDL

  1. I seem to recall seeing a project which uses low cost light sensors on the edges of the panels to determine when and how far the panel tracks based on the differential between the sensors. I would think that would reduce the challenge of seasonal adjustments, horizions and mis-programming/calculations.

    This is a nice project and nicely written up.
    Thanks

  2. How do the power curves compare when the power required for the stepper motor and motor controller is subtracted from the Sun tracked power curve? Essentially, how do the net power curves for the untracked solar cell and the tracked solar cell compare?

    • Chip,

      Good question. We intended this series to look how much more we could generate. We are now doing a set of measurements using our Grove INA3221 to see how much current the stepper motor system uses in operation.

      Will publish it next week.

      SDL

3 Trackbacks / Pingbacks

  1. Sun Tracking for Solar Power - Part 4 - The Video - SwitchDoc Labs
  2. Sun Tracking for Solar Power - Part 2 - The Software - SwitchDoc Labs
  3. Sun Tracking for More Solar Power - Part 1 - The Hardware - SwitchDoc Labs

Comments are closed.