Build Log: Monitoring the Direction of the Hudson River Flow using just an XBee Cellular and Sensors

Thread Starter

eduncan911

Joined Nov 14, 2011
29
This is a build log of a school project my daughter came up. The project overview, concepts, and ideas are all posted on our blog post (and follow up blog posts).

I wanted to start this thread as this project keeps mutating as we start to build and code the project. As kind of a "we tried this, and then we tried that" kind of log. Basically, a public place for our journey and notes on the project.

The first change... No more Raspberry Pi!

Originally, we scoped the project to connect an XBee Cellular directly to a RPi for all communications. We would use the GPIO on the RPi for all inputs, process the inputs, and return the data. We have several DHT11s around and know how to get started; so, it sounded like a piece of cake.

However as we started hooking up the XBee3 and connecting it to the network, we started to dive into the docs and I realized just how flexible the XBee module is itself - with all of its own GPIOs.

Then finally, we ran into the documentation about MicroPython and all of our Adafruit experiments just came rushing forward!

> Daughter: Can we just connect the sensors to these pins on the XBee and program the same things? Seems easier than programming two things.

Sometimes, a child's mind brings us back to reality. :)

Further reading the docs reveals a Deep Sleep mode! So another theory came about: can we power down all sensors, and power them up during wake? The documentation says Pin 13 goes HIGH when the XBee is not in sleep mode. However, it lacks any information as to what the max AMPs this output supports. The idea may be to connect them directly to any sensors, and/or to a transistor to switch more power-hungry devices on/off depending on sleep mode - and drastically increasing battery life!

So we've dropped the RPi in favor of connecting the GPS and all sensors directly to the XBee for an overall ultra-low powered setup.
 
Last edited:

Thread Starter

eduncan911

Joined Nov 14, 2011
29
Another change: Need to change the VRM specs

Per the datasheet, turns out the little MCP1700-3302E I ordered is limited to 250 mA, which would have worked fine for a standard XBee.

However, the XBee3 Cellular device seems to want to pull up to 550 mA (bluetooth disabled). I missed that in the docs before.

Researching other options right now... I really liked the simplicity of the MCP1700 for things like this.
 
Last edited:

Thread Starter

eduncan911

Joined Nov 14, 2011
29
TMP36 is not performing, Switching to SHT31

This past weekend, we tried to connect an TMP36 we ordered from Mouser to a RPi to compare to an DHT11. From what I gather, the temps were not consistant ranging from 85 F to 91 F - it was only 72 F in the room.

What was cool about this past workout session with my daughter was that we got to use my EspoTek Labrador oscilloscope! I'll post some screenshots in a bit; but, this was way cool for her to see how voltages change.

I stumbled across this post that compares different temp sensors, as we also have some DHT11s (that hasn't been too accurate in the past).

https://borealperspectives.org/2018/12/17/the-great-temperature-sensor-tryout-part-1/

Looks like the XBee supports I2C through the machine.I2C import:

https://www.digi.com/resources/documentation/Digidocs/90002219/#reference/r_class_i2c.htm?TocPath=Machine%20module|Class%20I2C%3A%20two-wire%20serial%20protocol|_____0

Takeway:

So we're going to order an SHT31-D temp and humidity sensor from Adafruit, as they are just down the road a bit. :) (Really wished they had a walk-in store. Didn't find it at our local Microcenter).
 
Last edited:

Thread Starter

eduncan911

Joined Nov 14, 2011
29
Another change: Need to change the VRM specs
This was a good lesson. Before school today, I showed her the millions of combinations available to electron circuits. I said, "We have to use something called a Filter to narrow down our search." I first went to DigiKey and we narrowed down the LDO VRMs, but didn't find any in stock. I then headed over to Mouser.com and narrowed down our specs:

Input: 2V+ (planned operating is 4.8V fully charged) for low battery operation
Output: Fixed at 3.3V and > 750 mA

https://www.mouser.com/Microchip/Se...0zl2wZ1z0z63xZ1z0wa2eZ1z0w8raZ1z0wbyqZ1z0wd73

Also, by limiting the options to Microchip, I found a few that uses the same 1 μF (which I ordered a bunch of for the previous VRM).

We just had enough time to dig into one Data Sheet before having to head to school. Maybe it was TMI, but I showed her the table of inputs, some graphs to analyze, etc. The reason I am mentioning this is because on the way to school, she asked, "Daddy, can we connect it to the osciscope-thingy and see it?" :) This made me smile. She really liked to play with that breadboard/USB version I mentioned earlier.

Anyhoot... Here are the two I've narrowed it down to:
The deciding factor for me, besides the max amps (which I'll show her later today after school), was the voltage relative to the temp. Since this project will placed on the Hudson river throughout the freezing temps, I wanted the lease voltage drop off relative to temps.

MCP1826:
Screen Shot 2019-10-01 at 9.45.11 AM.png

TC1264:
Screen Shot 2019-10-01 at 9.45.23 AM.png

If I am reading this correctly, the MCP1826 at full 1.0A load has a Vdrop of about 0.27V (off of 3.3V) for around 3.03V available around 0C. The other has a Vdrop of about 0.45V at around 0C. Also, the max operating temps had the volts staying higher at higher temps (e.g. in the sun, baking) on the MCP1826.

So that's our winner for today. Now, to find another GPS unit as my old SUP500F doesn't seem to be powering on (maybe because it was in the hot/cold attic for several years).
 
Top