My grandparents raised chickens, so I've actually been in a house more
than a few times.  First thing to do is make some correct simplifying
assumptions..

1.  What you're trying to do is improve the decision on when to take the
chickens.  To go from a random sampled best guess to something more
concrete.  Only problem is, a person around this kind of thing will
become very skilled at making this decision.  So you have to decide if
you need a cheap system that does this ok, or an expensive one that does
it as exact as possible?  You're very unlikely to get both, and it'll be
hard to even just get cheap and any better than done by a person..
2.  Realize that you'll never get perfect, period.  Hell, as hot a
houses are normally, by the time you weigh that last one, the first has
wilted a pound or two away already.  Only way to get exact would be kill
them while you weigh them.  (Not that bad an idea, you could charge $10
an hour to let people hit the electrocution button, there are plenty of
twisted people out there.  But you need live delivery to the plant
though..)

  First thought came to my mind was as William said, food on one side,
water on the other.  But you still need many paths and scales, wouldn't
do to have a lot die because 5k chickens can't get through the chute at
once.  Yes, some chickens will be over/under ideal.  But that happens
anyway no matter what you do.  Doesn't matter anyway, with 5k chickens
you'll tend to get similar distributions from lot to lot.  If there's a
glutton chicken or a few, there'll tend to be a similar amount in all
groups.  Yes, sometimes you might still take them a little early or
late, but the best you can hope for is an improvement towards ideal
anyway.  Any automatic system that gets most birds will be more
generally accurate in the long run.

  So forget about what the actual weight of each bird is.  Just
integrate the total force applied to the sensors each day etc and keep
track of that.  You don't need absolute data at all, a bird doesn't have
to be still to get that, and it's still proportional to weight.  But
you'll still have to watch out for things like how they tend to huddle
when it's cold and rest when it's hot, but run like fools when it's
moderate.

  Still not sure how much whatever you do will gain you though.  Feed
taken, water taken, and temperature integrated with how old they are in
days will probably give just as accurate results vs all the
maintainability issues with an electronic solution.  I know there's some
variability in how old they are when delivered, but the curve will
quickly match up to the general curve and you'll know.  And it's much
easier to automate the food and water supply vs temperature etc to get
optimum chickens instead of trying to figure out if your chickens are at
optimum after the fact.





Ryan Pogge wrote:
>
> the simple scale idea has been tried, it just didn't work weel.
> its not that only one chicken stands on it but rather only a few..
> and to get a good sample you need to use LOTS of scales. also its hard to
> get only one to stand on it, and also to get it to stand fully on it and not
> bounce the scale around.
> Ryan
err.... that is the bad part we are talking more than 1,000 houses
containing 5,000 birds each ouch :-)