As Dave posted in January, to measure the capacitance we set up an Arduino with prewritten code which gives an auto ranging capacitance value, combined it with code to measure temperature using a thermistor and managed to get both readings to display on an LCD screen.
I did quite a lot of reading on these sensors and many seem to be coated in some sort of moisture permeable polymer. I reckoned it may be this that slowed the response time and caused other problems. I found a sensor designed specifically for measuring moisture in oil but the data sheet made it out to be so delicate it hardly seemed suitable for the sort of heavy handed use most of us give things. It was also very expensive and whilst the manufacturer was quite helpful, he couldn’t offer any code to run it in conjunction with the Arduino (although he said many of their customers used Arduino), instead offering their own, ready built and expensive electronics.
With a little more reading and I found a paper on testing new veg oil for FFA using a capacitive sensor (rather grandly termed Dielectric Analysis). Interesting reading …
http://www.mdpi.com/1424-8220/15/10/26457/htm Both water and FFA apparently effect capacitance readings so they had to heat the oil for some time to drive off the water in order to obtain FFA readings.
They made their sensor rather than using a commercially available item. Theirs comprised eight plates 100 mm × 112 mm with a 1mm gap between them, so basically a non adjustable air gap capacitor which, when dunked in oil, the oil becomes the dialetric.
So pretty easy to make …
Above is the first, quick attempt which comprised eight plates 25 x 25 with 1mm spacing cut from an old KTC tin. It seemed to respond quite fast and appeared repeatable, so I thought I’d go bigger and better. I found 28mm dia brass dog tags for sale in China at just over a penny each including shipping. A length of M5 nylon studding, some M5 nylon washers and a couple of lengths of 2.5mm twin and earth, earth conductor (which just happened to be a perfect fit in the dog tag holes) completed the shopping list.
A jig was printed to centre drill the dog tags and take a chunk out of the circumference diametrically opposite the existing hole …
I used 51 tags only because that was the usable number I ended up with from the batch I drilled! The whole lot was assembled predrilled hole to edge hole so the copper wire could be slid through the predrilled holes connecting every other plate. The wires and plates were then soldered which was a bit tricky as the plates act as large heat sinks, hence the tardy soldering …
Checking the solder hasn’t strayed is easily achieved by testing the two wires are open circuit with a multimeter. This sensor worked in air and capacitance altered when you “herrrred” on it.
This is what Dave is currently testing.
We'd welcome any comments, suggestions and improvements anyone can make. For instance we're wondering if this set up might monitor water in methanol and it should be able, as detailed in the paper, to determine the FFA content of dry oil.
