[conspire] Watering automation: an edge case cautionary tale

[Ross Bernheim]

Rick,

You have hit on the big difference between micro-controllers such as the
Arduinos and computers such as the Raspberry Pi. 

The Pi and many others are 3.3V whereas the Arduinos have many versions 
that use 5V. This is important when trying to interface to the rest of the world.

Most of the Arduinos have multiplexed analog to digital inputs. While these
are not high performance a/d converters, they are more than adequate for
most uses.

While there are ‘hats’ for the Raspberry Pi that can do a/d and other i/o,
the low price of Arduinos makes them a good choice to handle the 
interface between the Pi and external hardware in the real world where
you need a/d and real time control.

The limited memory of the 8 bit Arduinos and the lack of higher level
languages for them means that using them as smart peripherals to a
computer such as the Raspberry Pi means that the programming for 
the Arduino can be much simpler and smaller.

Ross 


> On Mar 7, 2017, at 1:09 AM, Rick Moen <rick at linuxmafia.com> wrote:
> 
> Quoting Paul Zander (paulz at ieee.org):
> 
>> Since OpenSprinkler is open source, you could, in theory modify the
>> software so the green house gets watered on a fixed schedule
>> independent of rain.
> 
> There's a less-convoluted way.
> 
> I have that perfectly fine Rainbird 8-station 'dumb' controller left
> over, lacking only a power cube (that I stole for the OpenSprinkler
> unit).  It can get reconnected, controlling water to solely stations in
> the greenhouse, and not connected to the rain sensor.
> 
> None of this is actually urgent, as I think consecutive rainy weeks with
> almost no sunny days is unlikely.  But I'm bearing the matter in mind
> for future work.  Another greenhouse-specific need currently going unmet
> is automated watering for seedling trays.  In the past, I've relied on
> everyone's good intentions for manual watering, with very mixed results, 
> i.e., lots of dead seedlings.
> 
> Getting an additional water feed _to_ the greenhouse is part of that
> bundle of problems, which bundle includes fixing the halfassed current
> water distribution.  At present, we have garden hoses going under concrete 
> paths and then connecting to solenoid valves, then going out more garden
> hoses to drip hoses.  It's all scarily bodged together and one hose leak
> away from turning half my yard into a pond.  Much of tht ought to be
> entrenched and upgraded to galvanised pipe.
> 
> And you know, one really can't pile much work onto an Arduino, because
> it's an astonishingly minimal computer.  Its big advantage isn't ability
> to support significant software (it can't), but rather good analogue
> interfacing abilities.  For anything complex, you'd want a real computer
> (say, a Raspberry Pi or better) sending instructions to the Arduino over
> TTL signaling.
> 
> There's a _lot_ of information out there about how to do practically
> anything (as to analogue interfacing, anyway) with an Arduino.
> Originally, I found myself deep down that rabbit hole:  I think I was
> reading an ingenious explanation of how to use simple household
> materials (like, a nail and a bit of bell wire) to make the Arduino able
> to monitor the level of moisture in the soil around a particular plant,
> when I realised I was in danger of spending the next three months
> designing and building hardware -- whereas, I really just wanted an
> assembled and working controller and working, off-the-shelf rain sensor.  
> So, that was when I mail-ordered an OpenSprinkler unit and bought a rain
> sensor from OSH -- so that it would Just Work[tm], but be hackable if I
> needed that.
> 
> 
> _______________________________________________
> conspire mailing list
> conspire at linuxmafia.com
> http://linuxmafia.com/mailman/listinfo/conspire