Alt-BEAM Archive
Message #05081
To: beam@sgiblab.sgi.com
From: Bob Shannon bshannon@tiac.net
Date: Wed, 07 Jul 1999 19:08:02 -0400
Subject: [alt-beam] Re: Something funny with the 1382 voltage triggers?
Steven Bolt wrote:
> On Tue, 6 Jul 1999, Bob Shannon wrote:
>
> ---8<---
> > Now that we are all brushed up on the circuit theory, and
> > failings of BEAM as a teaching tool, would anyone care to look at
> > the actual problem?
>
> I for one am not going to do any more of that. The 2-transistor
> latch/driver isn't interesting, it isn't much good in this
> application, and it is still usually shown with a 3904 as driver,
> which just isn't the right transistor to drive a motor with.
I know no easy way to tell you that your totally wrong this time Steven.
Please dont take this as flame bait, but there are some big errors here...
First, the 2 transistor latch/driver is the most popular SE for a damn good
reason.
Its the easiest to build and get working for the majority of builders. One
reson for this
is that there are fewer interconnections to be made well and correctly.
Suneaters simply have too many parts, and their power delivery to weight
ratio is very
poor when compared to the 1381 SE. The added complexity and weight rule
out
'better' SE's in many applications where space and weight are critical.
The idea that the 1381 SE is not good 'in this application' is highly
dependant on the specifics of the application. Its not good engineering
practice to make that choice without knowing all the design parameters.
The 3904 is perfectly good in this application, provided that the start up
current demand of
the motor does not bring it out of saturation. Looking at the voltage drop
from emitter to
collector when the motors I'm using fire shows that a 2N2222 would be
overkill in this case.
The 3904's are not 'broke', so I am not about to fix them.
Lastly the idea that the 1381 SE is flawed because of the .7 volt shut down
voltage needs
to be thought through in more detail. My motors run down to .81 volts.
They start under load
at voltages not a lot higher than this. As long as your not using a
calculator cell for power, the time to charge from .7 to .81 is less
important than the total energy from 2.7 down to .7 volts.
Provided I'm using that energy well of course.
By the time I've dropped down to 1.5 volts, I've already rolled clean over
a coin, and I have stored usable inertia in my suspended chassis. I'm not
done getting useful work done yet, and I dont want the SE to stop giving me
power now that I've overcome stiction and my rolling resistance is at its
lowest point. (watch the on-line video clips of Vore-n-more, and pay
close attention to its butt when the motors fire, notice how the energy is
simply transfered between the suspension components, its not easy to see at
that low frame rate, but its a very powerful effect. Without this effect,
dynamic walkers are impossible.
If you think the obstacle ability of a direct drive rover is nil, you have
not really tried yet.
And in this application, the 1381 SE clearly beats all others, the FLED
actually comes in next
due to its power to weight ratio (given that you 'shrink' the FLED's
plastic case a bit...).
The details of the application can become very important when you design
for maximum
practical efficiency rather than theorectical efficiency.
> If BEAM made any progress at all - as an entry level hobby - the
> 2-transistor latch/driver would have been thrown out long ago. It
> doesn't help people, it bothers them, and teaches little except the
> wrong things.
This we should leave as a matter of opinion. I think more people learn
from the more
popular SE's than they do from the 'better' ones.
> Can you do better? Sure. Separate latch and driver. Then make sure
> that your latch switches off properly (at well above 0.7V), or use a
> monostable instead of a latch; that's more appropriate for most SE
> applications.
Ok, prove it. Show me the numbers.
How many centimeters per minute, under what lighting, and over what
obstacles?
Or maybe my photovore can simply drag yours backwards faster than yours can
go
forwards? Lets test this directly. I have little interest in arguing the
point, but I am
very keen to test it emperically, maybe we can all learn something new
here.
Then tell us how bad the circuits really are. In some applications, doing
work in the sun is more useful that just moving in dim light.
I thought one of the beam ideals was the concept of a smart body? This
must include the whole design rather than selected abstract ideas of
efficiency.
Perhaps it would be most efficient if I simply gave you a set of motors to
play with?
------------------------------------------------------------------------
eGroups.com home: http://www.egroups.com/group/alt-beam
http://www.egroups.com
- Simplifying group communications
Home