Re: [mowbot] Mowbot list

robin nospam at acm.org
Mon, 23 Sep 1996 21:14:59 +0100

Monta Elkins <Monta nospam at vt.edu> wrote:
> You might want to reconsider; price and ease of construction.
> my 68hc11 from motorola was $68.11 512 bytes of ram
> no power supply, no input devices, no LCD disply,
> no stereo sound (whoops).
But Mowbot will have a strange power supply (12V lead-acid probably)
with lots of fluctuations as the motors kick in and the battery runs
down, so the regulator on the GameBoy might be unsuitable or, as other
things will need regulated power anyway, redundant; I don't know what you
were counting because the HC11 has buckets of inputs (timed, counting,
edge detecting, analogue) and the Mowbot doesn't need an LCD display
(or stereo sound:-). Nor I don't know what you got for $68, but round
here the adequate A8 is 12 pounds ($18) and that's much more expensive
than the Motorola price list (a mere $8 I think!).

> Development tools cost extra.
Not so. There are dozens of (admittedly not superb) PC tools out there,
and I do all my work under Linux with the excellent, free, GCC.

> You can't compete on cost/performance with some
> as mass produced as gameboy.
You can't if you want to play games. But that's not what we want it to
do, so the expense is in the peripherals we'd have to add.

> And availability... Ever try to actually order
> Motorola parts in small quantities; it ain't easy.
I can get one-off A8s by tomorrow morning mail order from Maplin (a
hobbyist electronics catalogue). I can't believe you lack a similar
supplier! I admit that getting the fancy parts is difficult, but we
don't need anything fancy.

Once again though, the goals must come first. If it should turn out that
we are going to need to store a map of the explored area to 1cm resolution
in 3-colour+IR+ultrasound, the processing is out of the HC11 league.
At the other extreme, if we decide a random path search is good enough,
the ``processor'' could be a couple of discrete logic gates.

> I've got a 110 volt string trimmer.
> I can't give you quantitive answers now, but I can give you some
> comments.
>
> Noise. Not a lot.
> Damage to non-grass:
> unsure.
> I've seen gas powered do a fair amount of damage to
> a human leg. abrasions, cuts etc.
> I suspect damage to non-living items would be minimal.
> Hopefully the cutting could be placed under the vehicle
> such that contact with anything non-grass would be very
> difficult.
> Motor was relatively small (guess less than a pound).
> I'll look for current requirements - guess about 1 amp at 120v.

This is much more powerful than I'd anticipated. I was hoping for
something more like 0.5A at 12V. Maybe as much as 1A.

> OTOH, Mowbot maybe able to cut at night too.
> 24 hours a day for a week, would seem to give us
> quite a bit or leeway.
Hopefully we won't need it (see later). I was rather taken with the
idea of a nocturnal Mowbot---charging in its solar kennel during the
day, and coming out only when it's dark. My mysteriously well-kempt
lawn would give the neigbours something to puzzle over!
> robin nospam at acm.org wrote:
> > Let's make up some numbers and see what comes out. These really are
> > made up though, so if anyone has any figures at all they're better
> > than mine: c=0.01m?, g=1E-6m/s??, w=0.1m, v=0.1m/s, e=0.2 (generous)
> > then the largest area covered is:
> > a=cvw/ge
> > =(0.01*0.1*0.1)/(1E-6*0.2)
> > =500 m^2
> > Amazingly, this is just a bit bigger than my back garden, and I didn't
> > even have to go back and fiddle the figures!
>
> offhand those figures seem reasonable.
> After during a hard rain I'd guestimate g=1E-5m/s [growth rate]
> But I might give w=.16m [cutter width], c=.02 [max cut length],

As I realised when I worked out the time between cuts, my estimate of
maximum cut length c divided by growth rate g is too small. The
above numbers mean you'd have to cut the grass every 2.8 hours!
So taking Monta's c=0.02m and dropping g a bit to 1E-7m/s gives
a more reasonable time bewteen mows of
t=c/g
=0.02/1E-7
=200000s = 2.3 days
and a much more generous area:
a=cvw/ge
=(0.02*0.1*0.1)/(1E-7*0.2)
=10000 m^2
Wow! That's nearly 2.5 acres!

Robin.

--
R.M.O'Leary <robin nospam at acm.org> +44 973 310035  P.O. Box 20, Swansea SA2 8YB, U.K.