Modular Farm Blog: Post 4 - Automatic Nutrient Dispenser Part II

Hi vertical farmers,

In the previous post we discussed ways to implement our nutrient doser prototype and simulated its operation while waiting for the delivery of solenoid valves.

In this post we will talk about the nutrient doser prototype, the challenges in its implementation and the solutions found.

The Implementation

We decided to make a magnetic based implementation of the doser because our attempt with an optical solution revealed susceptibility to ambient light.

So we attached eight hall sensors (AH3503 from Nanjing AH Electronic Science & Technology Co.,Ltd.) to a cable-trunking and we fit on it the buffer tube of our doser as can be seen in the following pictures.

Figure 1. Front view of the sensor showing hall sensors. The tube fits above.

Figure 2. Rear view of the sensor during soldering process showing power and signal cables.

As you may remember we are using 4 reservoirs containing 2 concentrated nutrient components, water and a pH corrector (typically an acid). Sequentially the buffer reservoir is filled with each of the 4 liquids while its volume is controlled via the magnetic sensors. The sensors were placed linearly to have a resolution of 2 milliliters of liquid in the tube. Once the required liquid volume is reached the respective electrovalve stops filling the buffer reservoir and the discharge electrovalve discharges the liquid directly into the main reservoir.

Inside the tube there is a floating container filled with air and a neodymium magnet.

Hall sensors are connected to the eight channels of EZR32WG ADC. The following diagram illustrates the connections of the sensors to the board.

Figure 3. Connections to EZR32WG.

Challenges

This approach revealed some difficulties.

As the magnetic field decreases strongly with the distance we had weak initial measurements. So we needed to file the tube in order to decrease the distance between hall sensors and the magnet.

Another challenge arises from signal processing. As the fluid rises into the tube and the magnet passes by the sensor we get a signal of magnetic field measurements similar to a sinusoid (see figure below). This signal needs processing to result in useful level information.

Figure 4. Doser diagram illustrating the signal, where V is voltage and P is magnet position.

The weight and shape of the magnet are also cause for concern due to signal strength and the need to float on the fluid.

Demonstration

Follows a video demonstrating the nutrient doser prototype in operation.

You don't have permission to edit metadata of this video.

Edit media

Dimensions x Small Medium Large Custom

Subject (required) Brief Description Tags (separated by comma) Video visibility in search results Visible Hidden

Parent content

Modular Farm Blog: Post 4 - Automatic Nutrient Dispenser Part II Design Challenge Project Summary #001: Modular Farm **Updated 12/01/2015** Design Challenge Summary, Week Ending August 29th

Poster

Upload Preview

Ok

Cancel

The level detection circuitry and sensors are functional but since we could not get the electrovalves on time it was not possible the build the full automatic nutrient dispenser prototype. Also the pH readings and control needs to be implemented. We will come back to this topic later. On the next post we will introduce our artificial LED lighting system concept, the circuitry and hopefully a working model demonstration.

Feel free to put any questions or share your thoughts and suggestions on our design and methodologies.

Keep in touch!