Methods I Rejected: Capacitance measuring microchips

Methods I Rejected: Capacitance measuring microchips

Soil Moisture Sensor Project

Microchips that measure capacitance

It is 2017. Surely there are microchips that can measure capacitance. They are used in touchscreens! So I added them to the list of methods to review.

Just in case you are wondering “Brendan, a touch screen can already tell the difference between air and your finger. Why don’t you just use the microchips behind the touch screen to measure soil moisture? It’s easier than reinventing the wheel”. I did think of that, but I don’t want to stick my phone in the ground.

The lack of people saying “Hey, look over here! I done it! I stuck my phone in the ground to measure soil moisture. Give me your money!” quickly made me suspicious that it was not a viable option.

Remember that when you add water to the soil, you will change the dielectric constant of that soil. Dry soil will have a constant less than 10 and wet soil less than 80. Since you want to measure how moist your soil is, not just whether your plant roots are drowning or are sitting in a desert, you need to be able to measure what is in between these two extreme values.

As it turns out, the chips that control touch screens are good at distinguishing between the air and your finger. But that is pretty much all they can do. They detect “all or nothing”, not what’s in between. The sensors in the touch screen can detect the large change caused by your finger compared to air, but not the small change caused by the pencil. The capacitive touch screen takes advantage of the fact that the dielectric constant of your finger – being mostly water – is closer to 80, but the dielectric constant of air is only 1. This is why your finger works on the screen when the pencil does not.

Microchips made by Analog Devices and Texas Instruments are able to detect what’s in between, but they are either too expensive or are inaccurate in soil. The drawbacks to these devices is that the chips by Analog Devices run on a relatively low frequency (16 or 32kHz for the AD715x or AD774x) and the chips by Texas Industries are relatively expensive (between $6 and $11 each for the FDC2x1x that run at 10MHz).

According to people who think they know what they’re talking about and some who have qualifications which support their claim that they know what they are talking about, you need to use high frequencies (hertz is the measure of frequency or how many times per second something happens and is signified by Hz) to overcome the effect of stuff in the soil.

These options either offer potentially low accuracy or, like time domain reflectometry, are too expensive to be used in a sensor network. Onwards to low pass filters or measuring how long it takes to charge up a capacitor.

Go back to review of previous sensors, or main page.


Leave a Reply

Your email address will not be published. Required fields are marked *