Public Lab Research note

Gather and analyze long term solar and wind data for my locale

by wjw | January 25, 2016 01:09 | 1,623 views | 2 comments | #12610 | 1,623 views | 2 comments | #12610 25 Jan 01:09

Read more:

What I want to do

There is some good data regarding wind and solar provided by various organizations via the web. However, getting data for an area that is really local is not so simple. When the topography is like that of the South Western Driftless area in WI, (Grant County) making a decision about whether and how much solar and wind might be worth installing on a given site boils down to measuring at that site. Hills and valleys and the way they are arranged can have an impact on determining what is, or is not feasible to invest in.

I decided that I needed a plan, and that I was not going to be satisfied with that plan unless I had good data from right on site. So what I want to do (and have been doing) is building a simple, inexpensive and reliable way to gather data, analyse it, and hopefully disseminate it as well as make use of it for myself.

Site deployment will take place in March or April.

My attempt and results

I assembled a basic monitoring station which tracks wind speed, wind direction, rain accumulation, rain rate and temperature. Thus far it works pretty well on the home lab bench. I use a Raspberry Pi with an XBee S1 as my central gathering point. My sensors were obtained from SwitchDoc labs and Adafruit. The sensor array is connected to a Pro Trinket (arduino) which also uses an XBee S1 to send collected data to the RPI every 15 seconds. The sensor array is powered by a couple of LiPo batteries which are charged through a charge controller (also from Adafruit) from a 6v 2watt solar panel. Data received from the arduino to the RPI is stored in a SQLite database after being parsed by a simple Perl script that runs as a daemon on the RPI, reads the data from the XBee and stores it in said data base. Thus far(on the test bench) things work pretty well. I intend to add humidity and barometric monitors as well as they are very inexpensive.

Questions and next steps

I have placed some reasonable documentation on Github ( and would be glad to have anyone interested take a look. My main concern is with power generation. Solar panels, charge controllers and batteries are not my strong point, so any input would be appreciated. Additionally, I would be more than happy to share what I have learned thus far if someone out there is interested in doing this type of project.

Why I'm interested

My families farm is small, 234 acres. After some years of growing crops conventionally and being rather unsatisfied with the results, it was time to make a change. We will be allowing the soil to recover from the beating it takes from fertilizers, pesticides and herbicides. As such, there will not be much income, so finding ways to make the farm sustainable while it recovers is important. Of course one can't just let things sit. There are fences and buildings to maintain along with equipment. Some land maintenance is also required. All of this takes energy in one form or another. Along with all this is the conclusion that I have drawn that the small family farm must not die. Technology can and does provide for alternatives to the ever growing factory farm. There are efficiencies on a small farm which will never be matched on the larger farms. But we have to learn about them. To do so, we need to measure things differently or with a different goal in mind than we have had for the last 40 years. This project is one of many that I have in mind. It is just a start, but that is fine.

So in general terms, that is why I am interested in this project. I think more folks might also be if they sat down and thought about it for a bit.

I am also interested in finding out about wind and solar systems from folks that have them installed. In particular, small systems.


So, the time has come to get ready to move this from the bench to the real world.

Thus far...

  • Started with a new plated through-hole half size proto board which makes component mounting a lot easier.
  • Switched to XBee Series 2 for communications.
  • Switch to 3.6V, 1200mAh NiMh battery from standard 3.7V 700mAh battery
  • Raspberry Pi Perl code now uses Device::XBee::API to get data from XBee on Pro Trinket.
  • Added hot glue to protect connections on some solder joints as the wires are 25Ga and prone to break. Will order some PCB mountable RJ11 jacks and use phone plugs for connections later.
  • Tested code on Pro Trinket - seems solid, reliable, and fast.
  • Currently profiling Perl code on Raspberry Pi to ensure it is efficient. Seems to work fine now that I have made a few mods.
  • Am now using a USB to FTDI converter on the RPi instead of the GPIO pins. I have found it to be much more reliable and stable.

Left to do...

  • Need to add as MOSFET or Schottky diode between the charge controller and the Solar Panel to prevent back-feed when the panel is shaded or in the dark.
  • Add a voltage regulator on the output of the charge controller, as it will output whatever is available, ~7V when the solar panel is in play, or 5V if plugged into USB, or 3.7 volts if sourcing the battery. Not good. need 3.3 to 3.7 volts consistently for Pro Trinket and XBees.
  • Test range of XBees. Should be able to get 300 feet line of sight out doors. ~100 feet indoors.
  • Add the light sensor to the Trinket, modify code to send that data to the RPi along with the rest, modify DB schema to store light sensor values.
  • Build up the web interface for viewing data on the RPi.

This image show the RPi to the back of the post, the charge controller and 6V, 2 Watt solar panel to the left, and the protoboard to the right... all running at the time the photo was taken.


Overall setup (and my messy workbench). The tipping bucket rain gauge is behind the pole. There is a DS18B20 weather proof temp gauge just showing in front of the pole in front of the top bracket which hold the rain gage. Testing/calibration of the sensors has been done and is reliable. (except the light sensor which has yet to be incorporated.


I think I will have this done by spring and be able to gather data for the summer, fall and winter of 2016 and beyond...

Reply to this comment...

My initial testing has been reasonably successful. The hardware is workable, though I had to move the rain gage off the post due to the fact that, in windy conditions, it provides enough resistance to shake the entire assembly, rendering rain gage measurements unreliable. Beyond that: * the Pro Trinket works well. I did exchange the 3.3V for a 5V model though.

  • the DS18B20 temp sensor is accurate and reliable

  • interupt pins for annemometer and rain gage sensors are performing well

  • wind vane direction recording is solid now that the appropriate voltage values have been applied

  • I moved to a full size proto board to give room for other sensors later (barometric, humidity, etc..)

  • I have added a light sensor and am getting lux, broadband and IR readings from it

  • XBee series 1 radios worked ok, but I chose to get XBee Pro's (64mw) to ensure readings over long distances

  • Power is still a bit of a challenge:

  • The solar panel outputs 7+ volts which is too much for some of the sensors, which I prefer to run on buss voltage as compared to the 5V regulated voltage from the Pro Trinket
  • A DC/DC converter takes care of the voltage issue, but they are delicate
  • Inserted a schottky diode between the solar panel and the battery charger to prevent back-flow to the panel when it is shaded or dark out.
  • added a 3v voltage regulator along with a small buss for future use of 3v sensors
  • have a 5V regulator which will be added between the battery and the buss.
  • Will be using a 6V NiMH battery of 3400 mAh
  • Software has been solid. Will be very few mods from this point on. Probably only for added sensors.
  • Still some reading and adjustment to make on the XBee's. Theay are damn hard to work with!

There is about 5 hours worth of work/testing to do before I re-deploy to the back deck again. If all systems work for the rest of the month of March, I will final deploy at the Farm in April. At that point the last challenge will be to ensure the wireless network from the Farm house reaches the RPi and is reliable. The Perl script running as a daemon on the RPi to receive data from the Arduino via the XBee is running great, starts on boot, and takes very little CPU or memory. The simple web page showing a graph of the various weather values is roughed out, will add more graphs so data is easier to read: one graph per data type. Graphs are interactive, so users can have some fun with them. SQLite DB is good to go, and Perl plays nicely with it. The final step here in the shop is to mount the Arduino and company in a weather proof box, mount the solar panel, and physically place the ligtht sensor.

This project is nearly done. Time to start planning for the Pollinator Monitoring project. Lots of new stuff there, with some new power challenges to solve.

Reply to this comment...

Login to comment.