Alt-BEAM Archive
Message #02701
To: "'Chiu-Yuan Fang'" chiumanfu@home.com
From: Wilf Rigter Wilf.Rigter@powertech.bc.ca
Date: Tue, 27 Apr 1999 17:46:05 -0700
Subject: [alt-beam] Re: walker sensors
Hey Chiu,
The idea is to induce temporary photo-phobia (or better yet photo-amnesia)
during collisions and that avoids conflict between the "turning away from an
obstacle" neuron and the "seek the light" neuron. As in technology imitating
life a beam bot cannot serve two masters/neural neurons (grin).
Also, I noticed that in the attached reference email, the emitters of the
2N3906s are connected to 0V (heheh!) .
This means that the 2N3906 base-collector junctions are used as diodes. Just
replace it with a 1N4448 for the same results. There are many other
questions about this circuit but perhaps Sir Real is still with us and could
comment. In all fairness, his lordship acknowledged the experimental
(surrealistic?) nature of his design so it's ok but perhaps simply not the
simplest solution.
enjoy (cause I sure am)
Wilf Rigter mailto:wilf.rigter@powertech.bc.ca
tel: (604)590-7493
fax: (604)590-3411
> -----Original Message-----
> From: Chiu-Yuan Fang [SMTP:chiumanfu@home.com]
> Sent: Tuesday, April 27, 1999 3:54 PM
> To: NSX -
> Cc: beam@corp.sgi.com
> Subject: Re: walker sensors
>
> Why don't you just try it?! I'm using them on my 5 motor walker and they
> work
> exceptionally well. Needs more tuning though. TRY IT and see for
> yourself!
>
> Chiu
>
> NSX - wrote:
> >
> > hie i would like to know whats the eye multiplaxer in chiu fang
> > website
> > >
> > > http://www.geocities.com/SouthBeach/6897/mail3.html
> > >
> > > thats in the influential mail section number three...i would like to
> > > know more bout it and how could i interface it into my
> > > walkers.....thanks
> > >
> >
> > ______________________________________________________
> > Get Your Private, Free Email at http://www.hotmail.com
>
> --
> Chiu-Yuan Fang
> chiumanfu@home.com
> ICQ=5614919
> http://www.geocities.com/SouthBeach/6897/beam2.html
[Wilf Rigter]
Subject:uCore Biasing Network stuff.
Date:Fri, 16 Oct 1998 23:07:11 -0400
From:Sir_Real
Organization:Cybermedia Productions
To:BEAM
OK Im gonna try my hand at ascii again I hope it comes out alright.
This is a circuit (or part of it) that uses the same latching circuitry
in the microcore FAQ that is used to switch the multiplexer back and
forth between outputs.
this circuit will switch GROUND from being connected to R2; to GROUND
being connected to
R3 this is good for changing your turn angle. and if you have a
phototransistor across R2 (can be done but not recommended) it switches
from photodetect mode to object avoidance mode. basically it just
bypasses the photodection when the object sensor hits an object until a
few seconds after the sensor clears the object. So if the light is on
the left and the robot is turning left and there's an OBJECT on the
left, as soon as the object is detected the bot turns off photodetect
mode and starts turning right. I am only drawing up ONE SIDE on the
circuit, the other side is exactly the same and you should know that all
the sensorws get crisscrossed mechanically on the bot. Left side sensors
go on the right side of the bot etc...
And if I am wrong about this please correct me.
Here goes
uCore bias network.
NOTE: P
T = Phototransistor
(don't ask me how to bias it i just plugged it in and it worked. I was
too lazy to check it's polarity. :) I didn't work good in the other
direction so I put it back :)
The emitters of the transistors go to ground.
1M Nv from the
microcore
o---------o-----R1---------------o-----------------------|>o---
| | |
P \ \ R3
T / R2 / 1M
| \ 1M \
| | |
o----o----o o-----o(In #1) o
| | |
| n \ R4
o-------P p----Gnd / 1M o-----o(In #2) see circuit
below!
\ |
Q1 | n
3906 o----P p-----Gnd
Q2
3906
Theory (sort of)
By default R1, R2 (with phototransistor in parallel) are tied to ground.
When the latch below switches states The R1/R2 network gets shut off
from ground and the R3 and R4 network are attached to ground.
1. Object detection latch
Vcc
|
|
:| N.O. switch (this is the object detect switch)
|
| U1
o-----o------|>o-----o (output) To input of next circuit
| | this signal also goes to
Q1 and
/ | defaults to HIGH during
normal
\ R1 = C1 operation
/2.2M | 1uF electrolytic
\ |
| |
Gnd Gnd
2. Inverter (while output of latch is high, output of this Inverter is low)
So basically these 2 circuits combined make what I call a flip flop. :)
maybe thats not REALLY what it is but thats what I call it. :)
In from circuit above >------o
|
\
/ 2.2M
\
|
p
(output)to Q2 o----N n------------Gnd
defaults to LOW |
\
/ 20K
\
|
+ve
Ok I have to go find more ways to switch paths between resistances and such.
I hope this helps someone out there I know I sure have gotten alot of
help from people. So I'm just trying to do my part . :)
Cyas.
Subject:Short out a resistor to bypass it.
Date:Sat, 17 Oct 1998 00:00:01 -0400
From:Sir_Real
Organization:Cybermedia Productions
To:BEAM
You can short out a resistor to bypass it using a PnP transistor as long
as the emitter is tied to ground like so. (haven't tried any other
configurations yet.)
Nv
o-----------------|>-----
|
\
/ R1 o--------------o logic LOW and R2 is in the circuit
\ | logic HIGH and R2 is shorted out!
| n
o-------P p-----o Gnd
|
\
/ R2
\
|
o
Gnd
I don't know how efficient this stuff is but it doesn't seem to be
draining the battery overly fast. Actually I can't see where it'd take
much current to switch a transistor on at all.
Man ya learn something new everyday. I used to hate working with
transistors because you had to figure out how to bias them and do all
these calculations and stuff but this is cool. I haven't picked up my
calculator all week. :) Not that I'd remember how to use it or anything.
heh.
Bill.
------------------------------------------------------------------------
eGroup home: http://www.eGroups.com/group/alt-beam
Free Web-based e-mail groups by eGroups.com
Home