I have worked on a inferred version of my bubble logger, but as the sensor is of various quality and hold no way of calibration I have stooped further development of the inferred version. Secondly, the inferred sensors very easily breaks, and especially if you attached the VCC wrong they breaks, hence, they not really suited for DIY projects.
The main issue is as said the quality of the sensors and hence to low quantitative detection, but also the inferred sensor hold the issue it double count bubbles at the very low pressure where a tiny bubble either before or after the main bubble get detected. This can be overcome by sound detection as here we just hear one blop, but I did not find the way to control this in code for the inferred version, hence, I go back and concentrate on the sound version for now. I guess also the data presented of FG vs. calculated FG under the theory page also indicates I have better “luck” with the sound version so far. The data obtained with inferred detection is attached here:
—————————————————– Old Page below ——————————————————-
To build a Bubble logger you will need:
- A NodeMCU Version3.x costing around 3.5$
- A ds18b20 probe + shield temperature probe (including a resistor 4,7K ohm) and attached it on pin 02 (D4) of NodeMCU, in all a price of 3$
- A Optical Sensor (LM393 H2010 Photoelectric Opposite-type Count Infrared Sensor for Arduino) and attach it till A0 of NodeMCU,. Can be fund for 2$.
- Optional – A single side Relay on pin 13 (D7) of NodeMCU.
- Some single-double-triple female til female wires to connect it all, 1$.
The Box I use to store it all inside is a sistema 200ml box costing around 2$. Assuming you have a S-shaped airlock , the cost of a Bubble Logger is around 10-11 $ if brought from China.
The keen reader will notice the optical inferred senor is connected till the VUSB of NodeMCU there will feed 5V til this sensor, but the “LM393 H2010 Photoelectric Opposite-type Count Infrared Sensor for Arduino” sensors I looked at all allows for both 3,3V and 5V.
Calibration – sadly none!
Currently I have no way of proposing a calibration of the Opposite Inferred sensors, and as such the only way is to find a sensor there might be suited for rG estrimate (and hence SG decision) is to buy a lot and test them by counting bubbles over time and compare it to what Ubidots state. This should be done when the bubble-rate is not to high at around 30 BPM.
Quality of LM393 H2010 Photoelectric Opposite-type Count Inferred Sensor?
The quality of these are a bit so and so and they can range a bit in detection sensitivity therefore you will need to check them as above. I would recommenced people to buy double the amount you think you might need as there is failures or dysfunctional items, epecially in regards of the inferred senors. Of the 10 sensor I have checked, 2-3 gave more or less accurate counts (90% detection) in regards of blops pr. min (see/hear count at around 32-34 BPM), and the other 7-8 sensor was down to 50% of the count, so finding a stable sensor might be an issue. The test I made was to attached a fish tank pump and let it run into a closed container with an airlock+sensor attached at a steady 32-34 BPM and even I carefully attached the sensors the did not respond well. Hence, I have decided to not focus on the inferred version anymore.
In this regards the sound sensors seems way better and hence is better for rG/SG estrimation. Therefore, you will need to make you own polynomial if you wish to use the LM393 H2010 Photoelectric Opposite-type Count Infrared Sensor.
How to do it?
Please look at the various pictures for a bit more info, so easy!
Regarding the temperature probe, then the resistor is build into the cable in above pictures, and I just spliced together the cords and lastly used those heat warping plastic to secure it instead of soldering.