OpenDendrometer

OpenDendrometer board v0.2 has been received, assembled, and tested. All device functions are working properly. It reads the signal from the tread depth gauge (TDG), logs the TDG measurement and a timestamp from the RTC to the SD card, goes to sleep, wakes up if either the 4 minute period is over or a Bluetooth connection is made, and transfers data over Bluetooth upon request. The only unexpected finding was that the MCP1700-1502E/TO 1.5V regulator is actually providing around 1.3V. to the tread depth gauge. Fortunately, this lower supply voltage doesn't seem to effect the operation of the tread depth gauge. 

Next step is measure battery current draw during a few test operations. Ultimately, the goal is to confirm that the reduced power consumption strategies are working and to put together a battery life estimate for the device.

I am super happy with this board revision! Not only has it been my first board revision, but also the first time I haven't fully breadboarded before committing to fabricating the PCB. In addition, I have been wanting to develop a low-cost, low-power, Bluetooth, Arduino data logger design for years now. It is so great to finally see that come to fruition. 

Version 0.2 of the OpenDendrometer board has been finalized. Boards are out for order with JLCPCB. 

v0.2 changes include:

• MCP1700-3302E/TO 3.3V regulator with low quiescent current instead of 3.3V regulator provided on Arduino Pro Mini
• MCP1700-1502E/TO 1.5V regulator instead of voltage divider off of regulated 3.3V
• pinout to accommodate the HM-10 Bluetooth module
• 2N7000 MOSFETs on SD card and HM-10 Bluetooth module power in order to disconnect when not in use to reduce power consumption
• SD card module flipped over so it is easier to insert the SD card
• 2N7000 MOSFET on Bluetooth "state" pin for microcontroller wake up interrupt
• battery JST connector moved closer to the edge for less interference with tire depth gauge wires

Most of these changes are part of the strategy to more than quadruple the battery life of the device. Interestingly, they will only add about $1.50 to the board cost. Current estimate of time between battery charging with the new strategies is 4 to 6 months.

I am planning to use the following strategies to reduce power consumption of the next revision of the OpenDendrometer:

1. Remove the power status LED limit resistors from the Arduino Pro Mini and the RTC 3231 module.
2. Use a MCP1700-1502E/TO 1.5V voltage regulator for the digital tire depth gauge power rather than a voltage divider off the regulated 3.3V.
3. Remove the 3.3V voltage regulator from the Arduino Pro Mini. Use a MCP1700-3302E/TO 3.3V voltage regulator instead. 
4. Use a 2N7000 mosfet to cut power to the SD card when not in use.
5. Use a 2N7000 mosfet to cut power to the Bluetooth module during specific hours of the day. 
6. Log data to the SD card in batches. Fill RTC 3231 4k EEPROM, then dump to SD card.
7. Put the HM-10 Bluetooth module into a sleep state (via AT+SLEEP command) after disconnect.

The big push to achieve a working OpenDendrometer prototype in time for the 2022 Hackaday Prize finals was very successful. Though the project did not secure any top 5 finisher funding through the prize this year, it did receive $500 as a top 50 finalist. Fortunately, $500 goes a long way on this project.  

I fully intend to use some of that $500 to continue development between now and the end of the year. I have decided on two focus areas for development:

1. new hardware strategies to reduce power consumption
2. full implementation of wireless data transfer over Bluetooth

These changes will require a revision of the OpenDendrometer PCB. 

I was able to make significant progress towards adding wireless data transfer with Bluetooth this weekend. The image above is the breadboard unit used for Bluetooth testing.

Earlier in the week I had some issues getting the low-cost AT-09 Bluetooth modules to work. I was unable to connect to them with my phone. They also didn't respond as expected when I tried to configure them with AT commands. Based on comments online, seems like others have had similar issues with the AT-09 modules. To get away from this, I ordered HM-10 modules. The HM-10s have been working great without issue.

I have been using "Serial Bluetooth Terminal" for Android to connect with the HM-10 on my phone.

The screenshot below shows some of the functionality is working so far. Blue text "H", "B", and "D" commands are from the phone. Green text is what is received from the Arduino controlled HM-10 Bluetooth module. The "D" command transfers over all data from the SD card to the phone.

Excited to get Bluetooth functionality added so I can pull data off the OpenDendrometer in the backyard from the comfort of my living room.

It happened to start raining during the latest field test. Great to see the OpenDendrometer prototype dry and fully functional after 24 hours of exposure to rain. 

The true power of the OpenDendrometer is the realization that most of us have so much left to discover and understand about the life of plants. That those seemly unmoving creatures outside your window are actually very dynamic systems and that tools exist to observe their behaviors. 

It is weird to say since I have spent so much time developing it, but literally all the OpenDendrometer tool does is measure air temperature and small changes is displacement over time. Yet the data continues to capture my imagination about the life of plants -> periods of water stress and abundance, long term growth rates, changes that mark the seasons, and other stories that I haven't discovered yet.

If my work on the OpenDendrometer has made you even mildly curious about using commonly available electronics components to understand the life of plants, then this project has been a success.

Great looking data from the OpenDendrometer prototype over the last 28 hours! Showing the expected negative displacement at night (i.e. tree swelling) and positive displacement during the day (i.e. tree contraction). In contrast to the previous OpenDendrometer breadboard development unit, this prototype is showing dendrometer data with much less noise and larger changes in displacement. I suspect the primary cause of these improvements is due to design changes that increased rigidity between the tire depth gauge and the tree.

It really paid off to start logging the temperature data from the DS3231 RTC module as well. It is interesting to see the response time between air temperature and the dendrometer measurements. 

It was dark outside when I went to collect the data and update the Arduino sketch. A spooky glow from the Arduino and DS3231 power LEDs.

Next step of the plan is to collect data over a seven day period.

I secured the new OpenDendrometer prototype to the tree in my backyard this afternoon. The plan is to collect data over the next couple of days.

This new mechanical design includes some features that increase the rigidity between the tire depth gauge and tree. These should improve the device's ability to detect micrometer changes in tree diameter. Changes include: no longer using a 3D printed enclosure (so no material softening at 30°C), two threaded rods to anchor to the tree (previously was only using one), and a metal shaft collar at the probe tip (previously was a 3D printed tip).

Excited to see the data!