Group Actions
Engagement
  • Author Author: gpolder
  • Date Created: Date Created
  • Views 2029 views
  • Likes 1 like
  • Comments 10 comments
Related
Recommended

Old meets new, the 1-Wire Weather Station on the SPARK Core. (part 6)

gpolder

Now we have a running 1-Wire Weather Station on the SPARK Core, as described in Old meets new, the 1-Wire Weather Station on the SPARK Core. (part 5), we now come to the next step, as the station needs to be powered outside. We need to put the weather station somewhere in the backyard, where no external power is available. Commercial wireless weather stations mainly use a low power 433 or 868 link which transmits weather data once per minute or less and runs for months, if not years on a pair of AA batteries. Compared to this the power need of the SPARK Core is huge. Therefore I decided to use a solar panel for powering.

A small panel of 5 Watt should be enough. Some electronics between the bare solar panel and the SPARK Core are needed since the Core needs to be powered with 5V DC. Also a battery and a charger circuit is needed for operating when no sunlight is available.

 

Parts list:

  1. Solar panel 5W
  2. SPARK Battery Shield (mustache board)
  3. Richtek EVB_RT7275GQW Evaluation Board, modified for 5V.

 

Solar Panel

The solar panel I planned to use for this project is a small 5W panel with a Voc of 21V and Vpm of 17V.

image

 

SPARK Battery Shield (mustache board)

For powering the SPARK Core the SPARK Battery Shield (Spark Docs | Shields and kits)  is used. The battery shield is a LiPo battery charger and voltage regulator combined into one pcb with a weird form factor. You can use it to power your Core with any 3.7V LiPo battery and charge it at the same time via the USB port. The shield is built around Microchip's MCP73871 battery charge management controller and TI's TPS61200 boost converter for up converting 3.7V to 5.0V. The battery shield comes with a small 3.7V 450mAH battery.

image

Input for this board is 5V mini usb, designed for connecting to a 5V power outlet. This will not work with the voltage provided with the solar panel, so a second circuit is needed to convert the varying voltage output of the solar panel to a steady 5V. A traditional 7805 voltage regulator should work, but then to much power will be lost into heat. Luckily I recently did a review of the Richtek EVB_RT7275GQW Evaluation Board , which  is perfectly suitable to do the job.

 

Richtek EVB_RT7275GQW Evaluation Board

Originally the evaluation board is designed for 1.05V output. In my review (Richtek EVB_RT7275GQW Evaluation Board), you can find a description of the modification to 5V output by replacing one of the resistors by a potentiometer, as can be seen in the picture below. An USB connector was added to connect to the battery shield. A 100Ω resistor connects D+ and D-.  This is according to the USB battery charging specification, which requires a resistance of less than 200Ω between the data pins to indicate the USB port can be used for charging.

image

 

First tests

After these modifications All components were connected and the solar panel was placed in the window for some initial testing.

image

image

As can be seen in the picture above, the power and charge leds lights up, indicating that everything works as expected. The same is the case for the voltages measured at the input and output of the Richtek evaluation board.

As you can see in the image above the tests are taking place in winter conditions, meaning a long period of dark, and a short daytime period and low solar elevation. Furthermore we have lots of dark cloudy days during winter.

For the final setup the 450mAH battery will be replaced with an 1300 mAH version. Hopefully sunlight during daytime will be enough to charge this battery fully, and optimization of the software on the SPARK Core will make it run a full night on this capacity.

 

Stay tuned for further experiments and results.

Please let me know when you have any questions.

  • in reply to gpolder

    Gerrit

    I had forgotten this conversation ...

     

    All things with solar are 'variable' since the angle v where the sun is, the amount of sunlight, etc can vary.

    You have a 5w panel and if you assume for calculations that it's 80% efficient you get 5x 0.8 = 4.8w power.

    Your panel delivers 17-21v and the current at 5w @17v is 294 mA. (which is the absolute best you will get)

    So the most current you have during the time the sun is on the panel is 294mA, which needs to both power your spark, and the remainder goes into the battery.

     

    I found this that may help and applies the 80% efficiency rule in the time to charge (but gives the other results as well)

    image

         source http://www.csgnetwork.com/batterychgcalc.html

     

    While the sun is out and your spark is 'running' the charge current will drop below the 290mA.

    This chart applies to any battery system but lets assume your charge/load system sticks within the guidlines for the LiPo you have.

     

    I use a water as an example to explain power.

         Imagine you have a bucket and a dripping tap.

         The tap only drips during the day when the sun is out (the pump is off at night)

         You've added a full bucket under the tap, and every 10 mins you take a quantity of water from the bucket.

     

         If during the night you take out more than the dripping can replace, you will drain the bucket below the point of being able to take the water.

     

    If you can, I would suggest measuring the current the spark takes to do whatever you want it to do.

    This would be the constant load and might be 100mA as an example.

    100mA constantly is 100mA/H and would drain the 1300mA/H in 13 hours.

     

    If you only ran it for 1/10th of the time (60/10 = 6mins) you would draw 100mA/10 = 10mA/H and your 1300mA/H battery would last 130hrs.

    In this scenario since you run it for 14 hours (overnight) at 10mA/H, you only need to replace 14x10 = 140mA/H

     

    image

    which is effectively 0.5 hrs charge time (at full current)

     

    I hope that helps.

    Come back if I have confused or missed something ...it's entirely possible

     

    Mark

  • in reply to mcb1

    Mark,

     

    I don't get it completely.

    Taking your numbers, the right math is: 4 watt over 10 hrs is 40Wh.

    when you consume this in the remaining 14 hrs this results in 40/14 = 2.84 W.

    Which means that you can draw 0.767A, or 767Ma during this 14 hours.

    Am I right or not?

     

    In practice things are different.

    I expect the 5W is peek, I will get much less, say 1W.

    Durring daytime the Spark will run on solar power and the battery will charge.

    Since my battery is 4.8 Wh, this is possibly enough.

     

    The challenge now is to run the Spark during the 14 hours of no sun power on 4.8Wh, which is 0.34W, or about 100mA.

    I already did some tests, I could run it for 5 Hours. By shutting down the Spark completely each 10 minutes I think I can extend this time significantly.

     

    I did a lot of test with different setups, batteries and code, so my next post is a little delayed.

  • in reply to gpolder

    So maybe its better to reduce the charging current, in order to make longer use of the solar pannel when lighting conditions are worse

    Thats an interesting thought.

     

    You could try simulating what happens with a power supply.

    I think you'll find that the current will drop on its own in order to ensure the charge voltage is high enough.

     

    You may also be surprised at the optimum angle required.

    I believe there are plenty of resources to find what angle based on your location gives the best on the web.

     

    As with most storage, you need to put in about 140% to get 100% out. (approx 71% efficient)

     

    With your 5w panel you can only generate power over 8-10 hours and if you assume an 80% efficiency (5 x 0.8 =  4 watts)

    4 watts over 10 hrs = 40 watts which will result in approx 2.84 watts battery capacity.

     

    The trick is to ensure your 24hr load is less than 2.84 watts (2.84W / 3.7V = 0.767mA) for 1 Hour or 31.9mA for 24hours.

     

    (I hope my maths is right)

     

    Mark

  • in reply to mcb1

    Mark,

    thanks for the info on the adafruit board. Actually the moustache board I'm using now uses the MCP73871 battery management controller, which is more advanced than the MCP73831 used by the adafruit board. The moustache board charging current is set to 500mA, but I can change that by replacing one of the resistors. Output voltage of the solar panel was around 5.5V with current weather conditions, while the Vpm* of the panel is 17V. So maybe its better to reduce the charging current, in order to make longer use of the solar pannel when lighting conditions are worse.

     

    Gerrit.

     

     

    *) Maximum Power Voltage or Maximum Power Point Voltage (Vpm) The voltage at which maximum power is produced by a solar panel.

  • in reply to bwelsby

    Brian, yes, I have to investigate in that. It is possible to put the unit in sleep for large periods. I also am thinking of putting the led of, this also will save energy.

    stay tuned.