Author Options:

How can I make a grain moisture content sensor that can withstand high temperature? Answered

We are working on a multi-grain dryer (controlled by Arduino Mega) that dries a corn to its required moisture content (mc). If the mc is near the required (ex. 12-14%), the dryer will stop automatically. We need to make a sensor that can withstand up to 140F while the drying is ongoing. Meaning, the sensor is inside the dryer reading mc every couple of hours. If there is any device, will the reading be affected by the heat? we are pretty low on budget now so we couldn't afford to buy any pricey device and I think its hard to ship to our country in terms of period since we are pretty much running out of time for our evaluation.


Jack A Lopez

7 days ago

I think the usual trick for measuring moisture content in grain, is by using a capacitance probe,


Which is to say you have something resembing a big capacitor; i.e. two parallel metal plates, two parallel metal rods, or two concentric cylindrical metal shells, or pipes. Then you fill the space where the dielectric goes, with your grain, or wood, or soil, or whatever.

It is the sort of thing that looks like it should be easy, like just based on the formulas for different shapes of metal plates,


However, I claim it is tricky, especially for amateurs in the field of electronic measurement, to actually build a working capacitance probe.

I think most of the reason for this, is because the capacitance of a useful probe geometry (e.g. plates with more than 1 cm of distance between them) usually winds up being a very small capacitance, and thus hard too measure.

For example, this page, from Hackaday, kind of explains the naive approach to building a capacitance probe,

and maybe explains why this does not work.

I have not yet found a tutorial for easy-peasy, guaranteed-to-work, build it yourself capacitance probe. But it might be out there. I'm still looking for it.

By the way, since you asked about temperature. The probe itself is just metal pieces, with a thin insulation layer covering them. These should work well at almost arbitrarily high temperatures.

However, because the capacitance probe is noisy, the maker of such a probe plus measuring circuit, might have decided to put the circuits, like ICs, op-amps and such, right on top of the probe, or inside it, if it is like concentric shells geometry.

The point of this story, is the measuring circuits probably do not like heat, but what are you going to do if they must be located right on top of the probe to get a less noisy signal?

So that is pretty much everything I know (or don't know) about capacitance probes.

My other suggestion is that perhaps you can use some other, indirect measurements, to try to account for lost water.

If you can somehow weigh (or mass) the whole batch, as it dries, it should be losing weight (mass) as it loses water.

Also if you could collect the water you're driving off, and condense it, and measure the mass, or volume, of the condensed water, then that would maybe indicate a stopping point.

Also measuring the relative humidity of the exhaust gasses might work. Like when the exhaust gas gets too dry, perhaps that indicates a stopping point.


10 days ago

Not sure if it is feasable for your project but:
Back in the day I had a similar problem on a machine, different production though...
For us the main problem was that the sensors failed far too early as the problem exceeded the specs for the enviroment.
Was about 20 sensors later until a combined brainstorming of people having no clue about it found the solution.
Final solution was construct a "bypass pipe".
Basically the air to be monitored was pumped through long pipe loop.
The sensor was mounted where the inside temp was low enough for the sensor to operate properly while still high enough to prevent any condensation.

For just humidity though it is also possible to use ceramic sensor plates.
Basically just some metal strips close to each other with the ceramic being like a absorbent sponge.
Only problem is that most of them are not linear and quite costly.