Difference between revisions of "Counting Poo/Ideas"
m (James moved page Counting Poo/Ideas to Project:Counting Poo/Ideas: Not sure why this didn't move when I moved the main page) |
m (Danspencer101 moved page Project:Counting Poo/Ideas to Counting Poo/Ideas over redirect) |
(No difference)
| |
Latest revision as of 11:50, 5 February 2019
Ideas
This page is for logging and discussing ideas for how to improve and progress the project.
Multiple Sensors, One Aggregator
Added by JF, 13Jan14
Consider designing the data-gathering controller to accept input from multiple latrines. When several latrines are located together, this may reduce the financial and power costs, relative to one controller per latrine.
I don't know how often this would be useful. Are there areas where multiple pits are close enough to allow this? Would it require a low-power short-range wireless network of sensing nodes instead of wires? If so, does that make it too expensive/complex?
Design pit sensor as "remote node"
Added by JF, 13Jan14
Instead of just a "dumb sensor" in the pit, incorporate a small microcontroller (e.g. ATTINY85) into the design.
For a small extra cost, this adds the capability to re-program/re-configure remotely without lifting the pit slab. Also reduces analog signal path lengths, which is always a good thing.
It also means that the design of the data collector can be generic, as the interface to each sensor will be the same regardless of its actual implementation. The interface can be a simple serial protocol.
Observe this diagram: https://s3-eu-west-1.amazonaws.com/fowkc.elec.wip/CP_RemoteSensor.jpg
In this example, communications is over an SPI bus, with the remote node as an SPI slave. The reset line is also bought out to allow remote programming on the same wires, using ISP. An ethernet cable might offer a suitable method for communications. Four cores are used for MOSI, MISO, CLK and RST. Two cores are used for 5V and two for GND.