A robot to water trees and plants at known locations.
This robot wanders around watering trees and plants at known locations, filling itself at the water station as necessary.
This project is actually in two parts - the watertruck and the waterstation.
A watertruck - A modified GizMow, adapted to carrying
A waterstation consisting of
The water container must be fixed to the rover, with the pump externally mounted so that the flow meter can be mounted to it. The flow meter must be mounted vertically with flow going upwards (that is, according to the specifications, but it's not actually clear how critical this is). The water level meter can be mounted at the top of the container, to enable testing of whether the container is full.
The work flow is
The rover goes to the successive watering points, and runs the pump, and monitors the flow-meter, depositing the required amount of water on each plant.
If the flow-meter rate stops (or becomes very low), then we assume the water has run out and we refill.
There are many fail-safes built into the water-station software.
It will be normally closed (ie no flow), and close if power fails (and stay closed even when power comes back on). There are ESP8266 relays which claim to require continuous open commands from the ESP8266, but then they arrived, that's not what they did (at least they were cheap).
A 'squirt me' request will only turn on the water or 10s.
If a 'squirt me' request is not followed up by another request within 10s, the station goes into lock-down (turns off water and won't turn on again for at least 5 minutes).
If continuous 'squirt me' requests span more than 10m, the station goes into lock-down (turns off water and won't turn on again for at least 5 minutes).
If a 'shut-off' request is recieved, the station goes into lock-down (turns off water and won't turn on again for at least 5 minutes).
Any water requests which are recieved while the station is in lock-down causes the lock-down timer to restart.
There are also some other mitigations mechanisms
The water station is not taking water from the main house tank, so that even catastrophic spills will only drain a supplementary tank.
If the rover detects that the fill is taking too long, it times out - that is, it just stops sending 'squirt me' messages, and raises an alarm.
The rover must move to a position where it can fill from the waterstation. How will the rover (and/or the water station) know that the rover is in the correct place to receive water? If the rover is not within the radius of the container mouth of the correct position, the water will spill.
Intended to be used with a fish tank. Can be used to detect if the tank is full.
Designed to work with a solar panel, should be good enough to pump out of tank into ground.
In principle it should be possible to just gravity feed this (and count the pulses to allow for reduced flow as the tank empties), but a pump will make the dispensing faster.
SEN-HZ21 pulse counter. It produces a pulse every 1.5mL (0.0015L) or so. [Around 0.00152 L/pulse @ 5L/min and around 0.00145 L/pulse @ 9L/min].
It can be used to detect that the tank is empty (no flow), but also used to regulate exactly how much water is dispensed on each location.
slosh needs to be prevented in the tank. Slosh will cause spillage, and may even cause the rover to capsize.