Hi Guys,
Let catch up with where I am today with my little project.
I've built up a box to house the project, its like the motherboard which runs with an Arduino Mega 2560 board. On there I have quite a few things already connected up and running, like the SainSmart LCD with touch control, on board temp sensors, real time clock, and connections out for ambient temp & humidity readings. Also there are a bunch of sockets for "expansion cards". These will be used as I come to develop the EC reading circuit, pH reading circuit, CO2 etc etc.
I also made a special power supply for the thing, it includes +-5V and +-12V split rail supplies for op-amps ill be using to make measurements..
I got some of the basic stuff up and running already from the sensors I started with.
Then I came to start playing with the EC probes that I got for this project. I wanted to find a source of good probes that other might be able to repeat with, so I got a couple of types in, one from RS by Lutron for about 30 quid, and a Chinese graphite probe for about the same. With either of these, as with another probe you guys might acquire in future, the specs are not so accurate.
The probes have temp sensors inside as well as the conductors for the E.C reading, so first is to get them to read accurately. I plugged them in to the Arduino and used the standard math for a 10K thermistor but I found the readings were out by 2 or 3 degrees!
So, I concluded I would need to be able to calibrate those temp sensors inside the EC probes, before Im going to get readings with exceptional accuracy, which is what Im after. I want the performance to match the expensive bluelabs stuff, better it and be on its way lab-grade.
Double checking the accuracy of the digital or otherwise temp sensors I already had, they were poor. TMP100 , the I2C device on the PCB, gives +/-1deg absolute accuracy. There are lots of other on the market and you can find 1/2 degree , 1/4 degree , but only a handful that are better and reasonably priced.
You average hi-low temp equipment will also have an error, and with five or six different temperature measuring things I could muster, I grew annoyed since everyone said the temperature was different. In the worst case there could be 2 or 3 degree difference between readings from different units.
so, I started playing with some better sensors. The TSYS01 from IST , accurate to 0.1deg C. But to use it I had to build me a home PCB reflow oven so I could mount the bugger on a board and connect to it. I found it was sensitive to the mounting, which would make it difficult for you guys to reproduce if you wanted to.
So I searched out more sensors and came across the TSIC 506F, which is a through hole pre-calibrated sensor, accurate to 0.1deg.
I also built up a special program in Visual studio to connect to the arduino and log the temp sensor outputs, this I can use to view the temp signal, log it, or run a temp sweep with several sensors and see the difference - this will be useful to calibrate thermistors and those EC probes with
To summarise a bit further, after testing various stuff I arrived at a good solution, that others might find also useful if they like tinkering with this kind of stuff.