Alt-BEAM Archive

Message #11516



To: beam@sgiblab.sgi.com
From: "Phillip A. Ryals" Phillip@ryals.com
Date: Wed, 08 Mar 2000 08:33:38 -0600
Subject: [alt-beam] Re: MicroCore vs. BiCore


See, I've even seen this *with* a motor driver, so I guess I kinda fall into the
"recognize feedback" crowd.

I built a 6V M/S bicore walker with servo motors and a 74HCT245 driver. It was
in the 'chiu' configuration with the front motor horizontal to the ground. If I
put an object under one of the front legs, that leg would lift higher. I was
able to get it to step about 3 1/2 inches high on one side while still keeping a
useful gait.

So from my experiments with that one, I think feedback was definitely playing a
roll. If it wouldn't have detected the increased current draw of the motor, it
would have flipped itself over when something was placed under a leg instead of
stepping higher. I could also easily see the process shorten. It wasn't just
stalling the motor until time for it to move the other way... The length of the
pulse sent to the motor in that direction was detectably less than the other
direction.

So am I understanding what my bot is doing, or am I overlooking something?

-phillip

Richard Piotter wrote:

> Well of course you won't get decent feedback through a buffer or driver
> of any kind. Many of Tilden's bots use VERY efficient motors (like
> escaps), and sometimes minimal driver circuitry. There are, from what I
> understand, two forms of motor/Nv feedback are currently
> recognized/debated/rejected, depending on who you talk to. Supposedly,
> depending on how little isolation you achieve between motor and Nv, you
> can supposedly effect the charge time of the capacitor based on motor
> current draw. The Motor will draw more current under load. on high
> efficiency systems with closely tied motors and Nvs, I presume the motor
> will draww current, changing the time it takes to charge the input
> coupling capacitor of the next Nv (longer, I presume). How does this
> effect the circuit??? I'm not sure. The other suggested form of feed
> back occurs when the load, even through a driver drops the total system
> power. This would presumably effect the charge time of the whole circuit.

Home