Alt-BEAM Archive

Message #12124



To: beam@sgiblab.sgi.com
From: Bruce Robinson Bruce_Robinson@telus.net
Date: Sat, 18 Mar 2000 23:41:26 -0800
Subject: [alt-beam] Re: Aesthetics


Meabadboy@aol.com wrote:
>
> In a message dated 3/19/2000 1:37:45 AM Eastern Standard Time,
> andrew@best.net.nz writes:
>
> << I seem to remember that we used to use mercury to disolve silver
> and gold, this was then painted on plates, cups and other items
> as it was p[ainted the mercury would evaporate and leave the trace
> of gold or silver.. it is conductive (dont stick them in the microwave)
> >>
>
> yes mercury will disolve gold ~ but ~ it is extreamly toxic to humans

Way back in my youth, people didn't take short term exposure to mercury
too seriously. We fooled around with it in science classes (and smuggled
it out sometimes).

Five years later they were decontaminating those classrooms. Since then,
the dangers of exposure to low levels of mercury have really become
apparent.

Nowadays they use cyanide to extract gold from ore. It's a whole lot
safer.

> ... if you do decide to play with this ~ make sure there is ample
> ventillation in the work area ...

If you decide to play with this, don't :)

Bruce



12125 Sat, 18 Mar 2000 23:27:30 -0800 [alt-beam] Re: H-Bridge Question (Newbie) beam@sgiblab.sgi.com Bruce Robinson Hi, Dave & Kyle.

The H-Bridge as shown is NOT smokeless. Turn both sides on at once, and
you have a direct short through the transistors.

The circuit is good as shown for about 100 mA (you can push it somewhat
higher without too much worry). If you need more capacity, replace the
two 2N3906 transistors (top row) with PN2907's, and replace the bottom
row of 2N3904 transistors with PN2222's. That'll give you 800 mA
capacity for about 25 cents extra. You can leave the middle row as
2N3904's, because all they do is turn on the other transistors.

What circuit are you using to control the H-Bridges? The BEAM Online
bridge uses POSITIVE inputs. This is fine for a bicore, since one side
is always positive and the other negative (the bicore is pretty much
smokeless as well). It, but it is NOT suitable for a conventional
microcore, where an active neuron usually has a negative output.

If you're going to drive the H-bridge from a ucore, hexcore, etc, you'll
have to modify it as follows. Replace the middle row (2N3904) of
transistors with 2N3906 transistors, and turn them 180 degrees (flat
sides facing the right in the diagram). This makes the "off" condition
positive, and the "on" condition negative (or ground). CAUTION: if you
fire this circuit up with the inputs NOT connected to Vcc, you'll smoke
the bridge.

To get the full story on this circuit, look up the "Tilden H-bridge" on
the circuits page at the Beam Web Index:

http://www3.telus.net/rfws/bwi/index.html

I use the full-blown, smoke-proof version (figure 6) as a standard.
Overkill, sometimes, but I've never smoked or overloaded one. If you
don't want to invest in the 74HC139 chip to protect the bridge, Terry
Newton sent us a note a couple of days back on how to protect a bridge
with a pair of diodes.

Bruce



12126 Sun, 19 Mar 2000 00:06:08 -0800 [alt-beam] Re: "Living Machines" question beam@sgiblab.sgi.com Bruce Robinson Pat Wehren wrote:
>
> 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?

Hi, Pat. It's an interesting question, and I don't know for sure that
the impulses will converge on that solution EVERY time. But here's part
of the reasoning. Note that the statement in Living Machines refers to
the case where only two impulses are circulating in a 6 Nv loop.

If the delay times were all exactly the same, the pulses would follow
the same pattern indefinitely.

If the delay times vary slightly, then there will always be one neuron
with the shortest delay, and it will always be followed by neurons with
longer delays. So if you have two impulses with one intervening neuron,
when the second impulse hits the shortest delay neuron, it will time out
and PUSH the preceding impulse out of its neuron prematurely. On the
other hand, when the first impulse hits a "fast" neuron, it will move
ahead faster than the second one, increasing the number of intervening
neurons by one (at least for a short time).

This business of pushing a preceding impulse ahead will always be
present as long as one neuron separates them. It doesn't happen if two
neurons are between them. So in purely non-scientific terms (apologies
to all), the impulses are most comfortable when there are two neurons
separating them.

That isn't all there is to it, of course. I played with some timing
diagrams for a while, and found that the impulses settled into a pattern
of oscillation, with the separation changing from one neuron to two, and
back again.

I suspect there are a couple of additional factors that might cause a 2
neuron separation (and note that Tilden refers to it as a "fragile" 180
degreee mode lock). One factor is the possibility of external influences
such as feedback (load-induced voltage drop) shortening the short delay
from time to time. The other factor, I suspect, is some influence of one
neuron being felt two neurons away. For example, when an impulse gets
"pushed" ahead, does it leave behind some residual charge that slows
down the following impulse? Dunno.

Now that I've got some room on my breadboard, I think I'll try firing up
a 6 Nv loop with two impulses and leave it powered up for a couple of
days to see what happens.

Results to follow,
Bruce



12127 Sunday, March 19, 2000 1:13 PM Re: Decrease motor voltage beam@sgiblab.sgi.com William Cox
>Ummm.. Just give it less voltage....
>-William
>
>----- Original Message -----
>From: Corey Centen
>To:
>Sent: Saturday, March 18, 2000 4:15 PM
>Subject: Decrease motor voltage
>
>
>> Is there anyway to decrease the voltage requirement of a motor? It will
go
>> slower but require a smaller voltage. Does anyone know how?
>> ______________________________________________________
>>


12128 Sun, 19 Mar 2000 06:26:14 -0800 (PST) Re: Aesthetics Justin
> Print PCb's out, photostat em' onto trans, and since
> carbon conducts electricity, pathways can be made.
> components can be stuck by whatever means, like
> conducting epoxy, or sticking another piece of
plastic
> onto the connections and melting em' , i.e.
laminatin'
> em'.
> Would look nice if wrapped around caps, or motors :)
> any comments? (or criticisms?) :)
>
> Just one - I don't think it will work :-)
> First test - does photocopy toner actually conduct?
Does it do so
> without massive resistance? Neat idea, but I suspect
Reality has other
> ideas - photocopying would just be way Too Easy.
Perhaps also look into
> the possibility of conductive paint. I suspect it
exists but is probably
> not accessable/affordable. Ways to make our own
could bring results. I
> suspect a better method might be to put copper tape
onto plastic and
> etch it like a normal PCB. (Mind you, copper tape is
very thin, but it
> might not be thin enough. And there is also a
possibility that the
> etchant might attack the adhesive, though that could
be easily avoided).
>
> On the topic - some stuff that could be useful for
some beam apps - that
> conductive rubber used on the inside of calculator
buttons and similar.
> Anyone know where you can get that in sheets?
>

__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com



12129 Sun, 19 Mar 2000 08:12:34 -0800 (PST) [alt-beam] Re: Biocore beam@sgiblab.sgi.com Evan Dudzik yes, just hook up bot power and grounds and continue
right onto the second chip from the first, like it was
one big long chip... this oughta work...

PS- the terms are BICORE not BIOCORE
and it's a NEURAL net not a NEUTRAL net...

--- Cory Houck wrote:
> 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?
>
______________________________________________________
>


12130 Sun, 19 Mar 2000 11:20:16 EST [alt-beam] FRED circuit (beginner) beam@sgiblab.sgi.com "Corey Centen" Hi everyone,

I just completed the FRED photopopper circuit. I have one problem though. I
am feeding 9v into the circuit but either only one motor turns on or they
both just stutter...

Any ideas. When I put a jumper from the collector of NPN to ground both
motors work, but I am afraid this is bypassing some parts of the
circuit...Any ideas? I don't think it is a defective NPN, it worked on the
breadboard...

Thanx.
______________________________________________________



12131 Sun, 19 Mar 2000 11:35:15 EST [alt-beam] An Idea beam@sgiblab.sgi.com "Cory Houck" I had an idea you have a bicore wired up to a nv from the out side so the
bicore chip will controll the nv like a brain connected to a spinal cord so
it can chose which funtions to use and not use like eyes(photodiodes) ears(
com device) arms?? and legs>
______________________________________________________



12132 Sun, 19 Mar 2000 11:48:39 EST [alt-beam] Re: FRED circuit (beginner) beam@sgiblab.sgi.com "Corey Centen" Quick Update...I substituted the solar panels with a new 9v and now both
motors work perfectly. Is there a problem with putting too much voltage into
the circuit?

>From: "Corey Centen"
>Reply-To: beam@sgiblab.sgi.com
>To: beam@sgiblab.sgi.com
>Subject: FRED circuit (beginner)
>Date: Sun, 19 Mar 2000 11:20:16 EST
>
>Hi everyone,
>
>I just completed the FRED photopopper circuit. I have one problem though. I
>am feeding 9v into the circuit but either only one motor turns on or they
>both just stutter...
>
>Any ideas. When I put a jumper from the collector of NPN to ground both
>motors work, but I am afraid this is bypassing some parts of the
>circuit...Any ideas? I don't think it is a defective NPN, it worked on the
>breadboard...
>
>Thanx.
>______________________________________________________
>Get Your Private, Free Email at http://www.hotmail.com
>

______________________________________________________



12133 Sun, 19 Mar 2000 11:52:43 -0500 [alt-beam] HPV Tutorial! "BEAM" "Richard Caudle"
content-transfer-encoding: quoted-printable

Hey all!

I finally got my arse in gear and put the HPV tute together. You can find =
it on my page: www.geocities.com/frankendaddy/BEAM.html

Have fun!

Richard Caudle
www.geocities.com/frankendaddy
Home ICQ - Frankendaddy

---------------------------------------------------------------------------=
-----

Guardians of the sacred words: Nee, Ptang, and NeeWhon! =



content-transfer-encoding: quoted-printable



>
Hey all!
 

I finally got my arse in gear an=
d put the =

HPV tute together.  You can find it on my page:  
href=3D"www.geocities.com/=">http://www.geocities.com/frankendaddy/BEAM.html">www.geocities.com/=
frankendaddy/BEAM.html

 

Have fun!

 





Guardians of the sacred words: N=
ee, =

Ptang, and NeeWhon!         =





Home