IoThermo Caps

#Why?

My mom runs a family business to produce nice and tasty jam, mustard and some barbecue dips.

Now, they want to know how long they need to boil the jam in the glass to kill the germs.

Too little time may not kill all germs and there will be must. Too much time will harm the decent look, the nice color from strawberries or raspberries will grey out.

So they need to take a look inside the glass. Measuring temperature directly during the boiling sounds great. Technically most challenging option, but that's where the fun starts.

So I started to put my brain to this topic. As first, strolling the web for any projects like this one. Well, commercial solutions were ready to buy, 2k€ per measuring unit and 5k€ for central Data-Log-Station.

#WHAT?!

Guess I can do this cheaper, even with calculating my working time.

So I came up with a 50€ data-log-station (Raspberry Pi B+ including SD and case) and about 20€ per measuring unit (ESP8266 with DS18B20 sensors).

Setting up the Raspberry Pi as a data-log-station was easy using Raspbian and additional software: Node-RED and MQTT. I will not do to any details of this software because there are great documentations out there in the web. Usage of this software will be explained with the measuring units.

Every measuring unit will consist of one glass with WiFi electronics and LiPo package and three connected caps with DS18B20 sensors glued in the center. The addressing of the caps will be done with the 1Wire-Bus addresses of the thermometer.

#HOW?

The ESP will read all temperature values and send a message via MQTT to the Raspberry. The message will look like

"Topic: CAPDEVICEADDRESS Payload:32.03" (yes, its summertime and attic with windows to the south..)

This messages will be handled by Node-RED: Translating the CAPDEVICEADDRESS to a human readable name (Cap-A,..) which is also drawn on top of the cap. (Its great to have a girlfriend who paints her nails, she will have all materials for this job)

Next job, saving this to a CSV-File. Picking CSV-Node from NodeRed, watch tutorial, done.

The coding for the ESP was mainly copy-paste from other projects. The sketch will be uploaded here in a few days, it needs some face lift and comments.

Prototyping runs perfectly. Just push the room temperature to the Pi, directly from the PCB (ESP8266 is not that breadboard friendly). I build myself a programming station for the ESP8266, directly including some pins for the DS18B20. Pictures will follow.

The processing list for the ESP is designed for stable runs, not for a decent power management is:

#setup: Connect to WiFi

#setup: Connect to MQTT broker

#loop: Send Message with Mac Address and RSSI

#loop: Read all temperatures of all connected sensors

#loop: Send for each sensor a message with device address and value

#loop: Wait for five seconds

With reading the RSSI, the measuring units sends some valuable data for process stability at all. A connection loss inside the boiling machine (5 meters long steel tunnel) needs to be detected in prototyping status, so a solution can be found. The results are still pending, first test in the boiling machine will be on 2018/08/07.

A nice dashboard view in Node-RED is displaying the temperatures and the RSSI in two charts. The main data science will be done later in excel with the log files, but to gain some early results, the dashboard is great and usable.

I am totally looking forward for the testing!

#FAILURES

Well, I did not research every detail and especially not the needed current. I did some testings with CR2032 batteries and they failed heavily. Such batteries are great for low power, but I did not build my projects as a low power application. I am quite sure that it will work with this kind of battery, but then I need to optimize my code... My LiPo from a drone is working so good, I have the possibility to charge and a run time at about six hours. Great deal.

As long as the idea with the small CR2032 batteries were active, I designed everything to fit inside a cap, so no cable routing outside of the cap. Good idea, but a lot problems: Electronic needs to be jam-resistant, so I need to cover the electronics with some epoxy. Not good for maintenance. Also, the battery is placed under a metal cap where 90 degree celsius hot water will be applied. the specification for these batteries are with a maximum of 80 degree. I was curious what will happen, but it turns out I don't need to go that risky way.

It is always good to keep an open mind during such projects. It can save a lot of effort in a wrong direction...