🚀 go-pugleaf

In article <1hq0ilINNbji@fnnews.fnal.gov> behnke@FNALF.FNAL.GOV writes:
>In article <00965764.D96DDB80@Msu.oscs.montana.edu>, uchrgrh@Msu.oscs.montana.edu writes:
>>I would like to build a solid state device that would turn off a water
>>pump in my aquarium if the water level drops  below a certain level.
>>I am willing to look at prefabricated stuff but would prefer to build it
>>my self.
>
>Seems that a simple micro-switch attached to a float would work fine.  The float
>would be attached to the lever of the micro switch.  The pump motor is in series
>with the micro switch.  When the float level gets below a level YOU consider
>safe, the switch opens and stops the pump.  For a analogous system, take the
>cover off of the tank of your toilet and watch how the action works when you
>flush it.  If needed, one could get fancy and have an extra tank of fresh H2O.
>When the water level in the main tank drops below a specified level, a second
>pump is started to pump fresh water into the tank to keep the filter pump
>operating.
>
>Just an idea.
>
>Mike "This 24GB database will be the death of me" Behnke
>
>My opinions are my own, and MOST of them don't mean a thing !
>
>mb

HEY!!!!  That would be simple, functional, and reliable!  What the heck are
you trying to do here?!

I was going to answer this in a few weeks.  I wanted to get the results of
the modeling and reliability studies first.  Since I'm on the spot though,
I'll give the preliminaries.

First, for reliability, you definitely need multiple sensors.  Three
should be sufficient.  These will each be monitored by separate
microcontrollers and the results sent to a master contoller via a duplexed
serial link.  Some more work will have to be done to determine the optimal
protocol for the link and so on.  Most probably have to develop an entirely
new one.

The master controller will compare the results sent from the three
subsystems and determine whether the water level is within desired
parameters.  Periodically, the master controller will run diagnostics on
itself and each subsystem.  If any part of the system fails diagnostics, a
signal will be sent to a dial-out modem which will automatically call a
computer at another site.  Periodic "I'm alive" calls will also be made.
If a scheduled call is missed, the computer will assume the system is down
and initiate emergency response notices to appropriate people.

There is obviously still quite a bit of design to do here, but this is just
a start.  Any other ideas would be appreciated. ;)