Alt-BEAM Archive
Message #09838
To: beam@sgiblab.sgi.com
From: David Stafford write2dave@yahoo.com
Date: Sun, 6 Feb 2000 09:52:43 -0800 (PST)
Subject: [alt-beam] Re: The 240 microcore - was Bicore question
Wilf,
Thanks for stirring my synapsi. Great food for thought!
It seems that using branches like that you could develope
some sort of memory in your bot. Or at least more complex
if then parameters.
Dave
--- Wilf Rigter wrote:
> 74HC240 is a octal buffer so you get more gates and more
> drive in one chip
> compared to a 74HC14.
>
> It is however quite possible to make a suspended bicore
> with a 74HC14 but
> you must use two unequal capacitors ie 0.22uf and 0.47uf
> to force the input
> with the smaller cap to have a sufficiently large voltage
> swing to cross the
> upper and lower Schmitt trigger thresholds.
>
> The question of microcore versus bicore is almost a
> philosophical one:
>
> A master slave bicore with a reverser circuit can be
> constructed with one
> chip. The bicore operates in the saturated mode anyway so
> there's no need
> for a PNC. It is barely possible to drive very small
> motors directly from
> the output of bicores in which case there is no need for
> a motor driver
> stage.
>
> On the other hand, the microcore behaves more like a
> simple model of a
> neural network with "processes" propagating from one
> neuron to the next,
> each influenced by local sensor conditions .
>
> Neurons, neurons everywhere and look at all those pulses!
>
> Here is a little thought experiment: in an open ended 6Nv
> (74HC14) chain
> with LED indicators, what is the maximum number of
> processes (output low
> pulses/LED on) that can be active at a time? I have not
> tried this but I
> predict that if you try to "stuff" more that 3 processes
> in a 6Nv chain some
> interesting things may happen. Since each process is
> independently
> controlled by a sensor, for example a photodiode, it is
> possible for the
> first few Nvs to generate short pulses and later Nvs
> generate long pulses.
> Now what happens when short processes collide with the
> longer processes? Is
> each process still independent or do they interact? Like
> a microcore ring,
> I'm sure the Nv chain can be "saturated" when the maximum
> duration of any Nv
> pulse in the chain is dictated by the shortest process.
> To inject processes
> into the chain, the first Nv can be "stimulated" with a
> beam flip-flop
> triggered with a pushbutton. That way you can manually
> control the input
> pulse widths to see the effect it has on process
> generation and propagation.
> Longer chains of n Nv should be capable of maximum n/2
> processes but no more
> before saturating. A bottom up tree-like Nv structure
> would see a process
> propagate down the trunk before radiating along the
> branches. Kind of
> interesting stuff if you plan to control a complex
> sequence of motor motions
> for example in a multi-jointed beam worm. What about a
> top down tree with
> processes propagating down the branches toward the
> central neural chain
> (trunk). Add some logic such that two processes must be
> counted or be
> present at a junction where branches meet before
> processes can propagate any
> further. With only one Nv triggered ie by a sound or
> light "event", the
> process may die at the nearest junction but with many Nvs
> triggered,
> processes join up at junctions and can reach major
> branches or the central
> neural chain to activate some action ie fire a motor
> neuron.. How about some
> memory (Nu) at the junctions so that the junction
> remembers the last process
> and "sensitizes" the junction to lower the propagation
> threshold. How about
> bottom-up and top-down Nv trees connected at each end in
> a branched but
> closed loop structure like the circulatory system.
>
> Well that's the nature of a thought experiment, ideas
> happen, propagate off
> in different directions until one or more hit fertile
> ground and then you
> gotta built and test it.
>
> regards
>
> wilf
>
>
> > -----Original Message-----
> > From: Senior [SMTP:kyled@cruzers.com]
> > Sent: Wednesday, February 02, 2000 3:36 PM
> > To: beam@sgiblab.sgi.com
> > Subject: Re: The 240 microcore - was RE: Bicore
> question
> >
> > Hey Wilf... ;)
> > I've got the opposite problem:
> > No '240's,
> > Excess of '14's.
> > How bout a 14 bicore? PSH? :)
> >
> > Laters,
> > kyle
> >
> > SG wrote:
> > >
> > > WOOWW !
>
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
9839 Sun, 06 Feb 2000 11:52:18 -0800 [alt-beam] Re: 74HCI4 Microcore Question beam@sgiblab.sgi.com Bruce Robinson > Geoff Waters wrote:
>
> Hi- This has been bothering me for a while now, what is the difference
> between the bicore and the microcore? The microcore uses a 74HCI14 I.C
> and the Bicore uses a 74HCT240 I.C right?
Hi, Andrew.
The bicore consists of 2 neurons which are alternately active. Connect a
motor to them, and the motor runs first one direction, then the other.
Two motor walkers use 2 bicores (one per motor) in a master-slave
relationship. This keeps the motors synchronized.
A microcore is a loop of 4 neurons with one or two pulses circulating
around the loop. Depending how you hook up the motors, and how many
pulses you have circulating, you can either have a motion similar to the
bicore (both motors running at once), or you can have a quite different
motion:
Motor 1 forward - Motor 2 forward - Motor 1 reverse - Motor 2 reverse.
> I learnt about the Microcore yesterday, on Andrew Miller's
> site. He explains that, if you put LEDs on the ouput of each
> Neuron, you should get a "Saturated" state.
To clarify, the LEDs don't cause the staturated state, they let you see
it. This is a great way to learn about the microcore; if you start right
away trying to drive a motor, you have too many things that can go
wrong. If the motor doesn't move, you won't know if it's the microcore,
the motor driver, the motor, or a loose wire. Using LEDs lets you play
with the circuit and see exactly what is happening.
> Now this is where i'm confused, where do you connect
> the PIN , PNC switches, ...
To start, let's define a neuron, because it isn't always obvious when
you look at a chain of them. Using Andrew Miller's diagram as a
reference, let's "remove" a neuron. I'll use the middle neuron on the
bottom of the chip image.
It consists of a capacitor with the right hand lead connected to both a
resistor (which goes to ground) and a Schmitt inverter input. The input
to the NEURON is the left hand lead of the capacitor; the output to the
NEURON is the output of the inverter.
The place where you connect PIN and PNC switches is the point where all
3 components in the neuron connect to each other, which is the input to
the INVERTER. This is often refered to as the bias point.
> do you connect them to every Neuron, ...
Nope, not in the circuit he describes. You just connect them to the bias
point of one neuron. The pulses (or processes) move around a ring of
neurons, each neuron causing the next one to "fire" after a delay. By
connecting a PNC switch to the bias point of one neuron, you prevent it
from reacting to an incoming pulse, so it cannot cause the next neuron
to "fire". In effect, you are absorbing pulses as the move around the
loop. Eventually, they are all absorbed and you have a non-functioning
microcore.
Similarly, the PIN need only be connected to the bias point of one
neuron. It will inject a pulse, which will then travel around the loop.
And yes, there are PNC circuits which connect to every neuron, but they
are special cases for specific applications. They aren't required in the
Miller application.
> where do you connect the motor drivers and motors?
To the output of the neurons, which happen to be the outputs of the
inverters. Some points to note:
- you can connect a "driver" chip such as the 74--245 directly
to the neuron outputs. This is Miller's approach.
- you cannot usually connect an H-bridge directly to the neurons,
because it's possible then to turn on both sides of the H-bridge
at the same time, causing a direct short circuit through your
transistors. Heat & smoke!
- There are "smokeless" H-bridges out there -- Wilf has designed
a few -- and Tilden uses a decoder chip to eliminate the
problem.
- Bicores don't have this problem with H-bridges, because you can
only have one neuron active at one time in a bicore, so you can
only "turn on" one side of the H-bridge.
> How many neurons can fit on a single 74HCI4 chip?
Six, as in HEX. One inverter per neuron.
Hope all this helps. If I haven't been clear, fire away.
Regards,
Bruce
9840 Sun, 6 Feb 2000 11:58:49 -0800 (PST) [alt-beam] Voltage detector question beam@sgiblab.sgi.com "Leonardo Boulton F." Could anybody explain me how exactlly a voltage detector works?...
Thx.
L.
=====
I'm here... http://geocities.com/l.boulton
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
9841 Sun, 06 Feb 2000 14:07:14 -0600 [alt-beam] Instead of 74x14 for Microcore.. beam@sgiblab.sgi.com Kelly Petriew Hey again all,
As I was in the early stages of BEAM (still am), I happened across a local
electronics repair shop that sold me some chips that were "compatible" with
the 74x14's that we use for the Microcore. These chips are CMOS and much
faster though.. I breadboarded a '245 motor driver and a microcore using
this chip and then replaced it with the proper 74HCT14's that I bought from
Solarbotics.
Question:
Is there a need to pump the Microcore faster? What
advantages/disadvantages would this have? Is it more efficient?
The only thing I can think of is that your motor driver would speed up your
output to the motors and thus make the bot move faster (assuming the motor
driver can keep up). Bad if using a solar engine as the caps would
discharge quicker.. right?
The chips in question are CD40106BE made by Harris --> now Intersil
I went to their site at:
http://www.intersil.com/search/index.asp
and searched for that part number. All that came up were (heheh) Radiation
Hardened Hex Schmitt Triggers, the CD10406BMS. (Sounds Mean eh?)
So, are these things TOO FAST, or what?
PS: my motor driver used a Philips 74HCT245N and seemed like it kept up
just fine. The tiny walker I intended to build would be waaay too fast and
fall apart if I used that CD40106 chip...
Kelly
9842 Sun, 6 Feb 2000 12:29:44 -0800 (PST) [alt-beam] Capacitor Identification beam@sgiblab.sgi.com David Stafford
I've been dismantling some electronic junk for the parts.
I have a bunch of caps but I don't know the values. Is
there an easy way to check?
I know I could look at the amount of time it takes to
charge to a given voltage on a Scope. Then do some fancy
calculations to determine the farad value. But I don't
remember the equations (please tell me the equation to save
me the effort if you happen to have it handy), and I don't
have easy access to a scope. Any other suggestions?
Also if anyone knows how to decipher the capacitor markings
that would help me tremendously.
Thanks
Dave
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
9843 Sun, 6 Feb 2000 15:29:37 EST [alt-beam] questions, quetions, questions... beam@sgiblab.sgi.com StickG@aol.com (this message has been posted, but not sure if sent, so...)
Well hopefully, if I get some good answers to my questions on
this post, then I won't have to ask for a very ,ong time, and you guys
will have one less newbie to annoy everyone.
My first question is this: If I wanted to make a four motor,
"stryder" type walker, how would I use the head component to control
the other circuit part? (pin numbers or aquick scanned sketch would be
great, but if you can describe it well enough, then no need)
My second question: Along the same lines, how does one connect
a reverser circuit to a two or for motor bicore in order to make it
reverse or turn?
Is it possible to make a two motor walker phototropic with out
a head?
Now here's one to stir up some controversy, Which is the best
way to go in a walker circuit? Bicore or Microcore or even Hemicore? Is
there one that gives more complex behaviors?
Ok, that's it (Heh,Heh) Hope to hear from you all! thanks.
stick
Home