HDC1000 Temp/Humidity sensor sensitivity

Happy to help.

First of all here is the specification for the HDC1080. It has all the answers!

www.switchdoc.com/wp-content/uploads/2017/01/hdc1080.pdf

That being said, from the specification, the specs for the temperature are:

+/- 0.2C or +/- 0.4C at the lower resolution.  The worst case would be about ~0.7F so that would be close.

To make this simpler, cd down into the SDL_Pi_HDC1000 directory and run


sudo python testHDC1000.py

If the default resolution for the temperature is 11 bits, then your data makes sense.   Look at the testHDC1000.py code and you can see how to convert the temperature to 14 bit accuracy.   Then add that to the SmartPlantPi code.   The beauty of open source.

BP

I second that request!

Unfortunately I have been unable to get more granular data with the HDC1000 sensor - looking around I'm not sure anyone else can either.

You pull two pieces of data from the sensor - d0 & d1, let's call them.  These two pieces of data are always the same, and so when you do that math on them you find that you will never get readouts <1 degree F apart.  Admittedly I am no programmer! So I concede I could be missing something, and I did learn a bit *trying* to get better data from it, so it wasn't a waste of time, but your own freeboard is the same - big steps in readouts.

However I am going to order this:


www.seeedstudio.com/Grove-Temperature%26Humidity-Sensor-Pro-p-838.html


This is the AM2302 AKA DHT22, which I already have and it outputs very granular measurements (0.2F apart).

I'd just be curious if you picked the HDC1000 over the DHT22 for any specific reason that might not be apparent to a beginner like me?

Thanks again for your help and advice!! I've learned so much working on this project.

Thanks so much for checking yourself! I spent a lot of time on it but I learned a lot. Looking forward to the update.

I merged your changes and it works great! Thank you!

Nice Picture! That shows that I nailed the problem yesterday! Whew! I wrote most of that code so it was my responsibility to nail that bug to the ground.

Remember you promised a tutorial on grafana!

BP

Wellll I'm not such a great writer, nor a good programmer (yet!) - but I happen to administer Linux systems as part of my job so I have some experience there, so for me, setting up InfluxDB and Grafana was not too challenging, but I used this to guide me:

www.andremiller.net/content/grafana-and-influxdb-quickstart-on-ubuntu

That is about a year out of date but instructions are very similar.  So first caveat is that this is not running on my Pi, it's running on a server up at AWS (t2.micro, does not need to be beefy).  I think it could run OK locally on the Pi.  It seems to be using only 177mb of RAM (both Grafana and InfluxDB) - but it may need better I/O than an SD card can provide.  But for our purposes it's not so much data so I think it'd work.

For my Python code to get the data up there, I hacked away at SwitchDoc's code until I got it right


First I need to import the right python module for InfluxDB, and tell it about my server:



from influxdb import InfluxDBClient
client = InfluxDBClient(host='myinfluxdbhost.example.com', port=8086, database='smartplant')

Then I add the job that will be scheduled to run.  This crafts the JSON necessary for sending along to InfluxDB.  The "Value" I reuse from the existing variables that have polled the sensors, so that was effortless.  The "Measurement" field represents the "Series" in the InfluxDB database.  Each sensor needs its own "measurement" so that the data is separated into the different "series" in the database.



def publishStateToInfluxdb():
  timestamp = int(time.time() * 1000000000)  # need microseconds!
  if (DEBUG):
    print('Publishing data to InfluxDB, time: %s' % datetime.now())
  json_body = [
    {
      "measurement": "sunlight_vis",
      "tags": {},
      "time": timestamp,
      "fields": {
        "value": state.Sunlight_Vis
      }
    },
    {
      "measurement": "sunlight_ir",
      "tags": {},
      "time": timestamp,
      "fields": {
        "value": state.Sunlight_IR
      }
    },
    {
      "measurement": "sunlight_uv",
      "tags": {},
      "time": timestamp,
      "fields": {
        "value": state.Sunlight_UVIndex
      }
    },
    {
      "measurement": "airquality",
      "tags": {},
      "time": timestamp,
      "fields": {
        "value": state.AirQuality_Sensor_Value
      }
    },
    {
      "measurement": "temperature",
      "tags": {},
      "time": timestamp,
      "fields": {
        "value": state.Temperature*9/5+32
      }
    },
    {
      "measurement": "humidity",
      "tags": {},
      "time": timestamp,
      "fields": {
        "value": state.Humidity
      }
    }
  ]
   client.write_points(json_body)


Finally, I just add a job to the scheduler:


   # send state to InfluxDB
   scheduler.add_job(publishStateToInfluxdb, 'interval', seconds=10)



I'm sorry, I know this is all ugly, but I can attest that it does work!  I hope some of this helps but please let me know if I can answer any questions.

To clarify, all this code went in SmartPlantPi.py.  I don't believe I needed to adjust any of the existing code, just added to it.  So I still just run SmartPlantPi.py with everything else it does, except now it also sends the data to InfluxDB!