Alt-BEAM Archive
Message #12056
To: beam@sgiblab.sgi.com
From: Bruce Robinson Bruce_Robinson@telus.net
Date: Fri, 17 Mar 2000 19:42:58 -0800
Subject: [alt-beam] Re: list within a list???
David Perry wrote:
>
> thats kinda of stupid - some people like Bruce Robinson (who only
> joined a little while ago) haven't built one robot...
Lot's of parts of robots, though :)
> Yet from pure knowledge absorbtion he has become an expert ...
Um, not quite. The one area I can claim a reasonable level of knowledge
is with Nv / Nu nets; I build a lot of them. Once in a while hook them
up to a motor and make something move. Sometimes it actually moves the
way it was intended ... sort of.
On the idea of coding messages, it may work, it may not. Using the [B]
for beginner won't work for long, because beginners will keep signing
onto the list and won't know the protocol. By default, their subject
lines won't be coded.
Asking "who lives near ... ?" seems a legitimate question ... how else
are you going to find BEAMERS that live in your area?
And as for a second list, there is one. Spinal Column on the BEAM
Heretics site. It's been around for about a year as I recall -- advanced
stuff only, please. It kinda flared up, and then died away. I notice
many of its contributers still contribute to this list. I can't see why
yet another list would fare any better ... try reviving Spinal Column
instead.
Meanwhile, gotta get back to making a circuit board for Victor, my
virtual robot :)
Bruce
12057 Fri, 17 Mar 2000 22:56:29 EST [alt-beam] Biocore beam@sgiblab.sgi.com "Cory Houck" Is their a way to hook up 2 chips {74***240) to get a 7 or 8 neutral net
(please correct me if im using the wrong terms) so you can have like a siw
legged walker with a head on it that will track light?
______________________________________________________
12058 Sat, 18 Mar 2000 03:28:41 GMT [alt-beam] Re: Mark Tilden Tact(Mr.mark T please reply) beam@sgiblab.sgi.com "Jason -"
>From: Dave Hrynkiw
>Reply-To: beam@sgiblab.sgi.com
>To: beam@sgiblab.sgi.com
>Subject: Re: Mark Tilden Tact(Mr.mark T please reply)
>Date: Fri, 17 Mar 2000 10:29:56 -0700
>
>At 09:08 AM 3/17/00 , Jason - wrote:
>>I am just wondering what material does mark T use to construct the legs of
>>his walkers.....it seemed so configurable and can be easily be bent.
>
>He usually uses 8 or 10-gauge solid-core house wire. When he has the shape
>right, he replaces it with 1/16 or 1/8" nickel welding rod.
>
>Regards,
>Dave
nickel welding rod? are they easy to be bend? and where could i get
them........what about those tactile sensors........is there any hints that
i could use to make really tough tactile sensors other tha using piano
wires.....cause i cant find it here....thanks
______________________________________________________
12059 Sat, 18 Mar 2000 00:04:59 EST [alt-beam] Re: Charging NiMH batteries beam@sgiblab.sgi.com Meabadboy@aol.com In a message dated 3/17/2000 9:47:20 PM Eastern Standard Time,
robertm@idcomm.com writes:
<< Don't know if it's the batteries, the charger or just my luck :) >>
suggestion `
the handy board site uses a charger circuit for its nicad bat ~ u can D/L a
PDF ` adobe acrobat file ~ and copy that circuit ` verry efficent and cheep
to build ~ takes up little space
the Bat is a Radio shack bat ~ has worked fine for me now for over a year
http://el.www.media.mit.edu/groups/el/projects/handy-board/techdocs/index.html
down load the "Handy Board-Specific Information" file it has all the
schematics of what u will need
Steve
12060 Fri, 17 Mar 2000 23:02:33 EST [alt-beam] Hobby Battery packs beam@sgiblab.sgi.com "Cory Houck" Can I use the battery packs from my R/C car for a walker I was wondering if
they would be to heavy and they only take 15 min to charge.
thanx,
______________________________________________________
12061 Fri, 17 Mar 2000 22:04:11 -0800 [alt-beam] Re: directional indicator for Bicore heads. beam@sgiblab.sgi.com Bruce Robinson Dennison Bertram wrote:
>
> Here's another one from the old files:
>
> I have yet to test this idea, but back when I was working on bicore
> heads with Depth Perception ablities, I was wondering how I could
> create a circuit which would tell me the directional angle that
> the Head was turned in. So here's the idea, which I'm curious to
> see what others think about it.
> Each motor out line from the Bicore is connected, through a resister
> to a Capacitor. When the motor out line is "high" It charges the cap,
> however when it's "low" it discharges the cap. Now, if this was
> attactched to a high frequency bicore head, you should be able to
> roughly judge the direction the Head is pointing in by measuring the
> average voltage across the cap. ...
It works, Dennison.
This is not too different from an idea I have been working on to monitor
the directon of a two wheeled robot. In my case I have a pair of
encoding disks on the wheels that send up/down pulses to the "counter
cap". As there is a danger of a wheel stopping in the "on" position, I
placed a couple of Nv's between the wheel encoder outputs and the cap.
This ensures that each pulse is "exactly" the same length.
The design calls for a full charge/discharge cycle to represent 400
degrees of arc, with a reset function if the robot turns a full circle
in either direction.
The biggest problem I had was sampling the voltage on the cap without
changing it. I finally set on using a LM324 Op Amp configured as a
voltage follower. I can measure the output of the Op Amp without
affecting the capactior at all.
The system almost works:
- it can hold a charge on the cap for 10 minutes with only
about a 2% change in voltage.
- I can monitor the Op Amp output an pull off any directional
position I want.
- The one thing that doesn't work (yet) is the full turn resets.
- For long term use, the system lacks a means of realigning itself
periodically. I have considered a digital compass attuned (say) to
North -- each time the robot faces due North, the compass resets
the cap.
The robot platform exists, complete with encoders (yes, I DO build
hardware). The direction circuit is on my breadboard, and I am
assembling a small driver board as we "speak".
As far as adapting it to a head, I think your idea is right on. Driving
the circuit off the bicore outputs is probably the best approach,
because the head movement will be approximately proportional to the time
a bicore is active. If your head has a stop that prevents it turning
beyond 360 degrees, then you won't have to worry about reseting the
charge after a full circle. Best of all, you can recalibrate the setting
each time the head moves past a certain point on the body.
In my case, the idea evolved from a "chirp counter" I devised for
Richard Caudle's HPV project (and which Wilf improved upon). What we
seem to be doing fundamentally is using an Nu neuron (or a portion
thereof) as a counter. There must be lots of BEAM applications for an
"approximate" counter of this sort.
Bruce
12062 Sat, 18 Mar 2000 00:38:59 -0700 [alt-beam] Re: Mark Tilden Tact(Mr.mark T please reply) beam@sgiblab.sgi.com Dave Hrynkiw At 08:28 PM 3/17/00 , Jason - wrote:
>nickel welding rod? are they easy to be bend?
Not bad. Depends on what you happen to find. I've bought stuff that's
softer than music wire to bend (nickel vs. steel - you guess which is
tougher!), but I've also stumbled on a few pieces that were a real bear to
bend, even more so than steel. But even the weakest is nice to work with.
> and where could i get them
Look up "welding supplies" in your yellow pages. Explain you're looking for
a high-nickel content welding rod, and they'll set you up. Just be ready to
pay for it - it ain't cheap!
>........what about those tactile sensors........is there any hints that i
>could use to make really tough tactile sensors other tha using piano
>wires.....cause i cant find it here....thanks
You can't find music wire? Check any hobby/craft store for a K&S display.
They're everywhere.
Good luck,
Dave
---------------------------------------------------------------
"Um, no - that's H,R,Y,N,K,I,W. No, not K,I,U,U, K,I,_W_. Yes,
that's right. Yes, I know it looks like "HOCKYRINK." Yup, only
2 vowels. Pronounciation? _SMITH_".
http://www.solarbotics.com
12063 Sat, 18 Mar 2000 00:40:57 -0700 [alt-beam] Re: Hobby Battery packs beam@sgiblab.sgi.com Dave Hrynkiw At 09:02 PM 3/17/00 , Cory Houck wrote:
>Can I use the battery packs from my R/C car for a walker I was wondering
>if they would be to heavy and they only take 15 min to charge.
They're sub-C nicads - a bit large for most BEAM walkers, but I've seen
them used quite successfully. Depends on the scale of your machine.
Regards,
Dave
---------------------------------------------------------------
"Um, no - that's H,R,Y,N,K,I,W. No, not K,I,U,U, K,I,_W_. Yes,
that's right. Yes, I know it looks like "HOCKYRINK." Yup, only
2 vowels. Pronounciation? _SMITH_".
http://www.solarbotics.com
12064 Sat, 18 Mar 2000 00:05:29 PST [alt-beam] "Living Machines" question beam@sgiblab.sgi.com "Pat Wehren" The paper says that if the times of the neurons are roughly the same, they
will converge on 100100. I can't for the life of me see why. If one
turning on turns the next one off, which pops back on after t, the only
way the processes wouldn't propogate normally was if they were actually
next to each other. What's with the "at least two neurons away"? WHAT am I
missing?
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
Home