Subject:
|
Re: The latest rage in pneumatic computing
|
Newsgroups:
|
lugnet.technic, lugnet.robotics
|
Date:
|
Mon, 23 Jun 2003 14:08:01 GMT
|
Viewed:
|
1921 times
|
| |
|
|
Kevin L. Clague wrote:
> I built it last night and it works great!
>
> I was concerned about your use of switches that is backwards from normal. I
> transitioned through all the possible values and it never lost system pressure.
> Great job!
>
> Now I have to study the design so I understand how you did it. There's are a
> few good design patterns in there.
You know, I can really empathize with you on the fact that it might be
cryptic. You don't wanna know how long it took me to get the wiring on
it just right so that I wouldn't be committing any pneumatic violations.
I'll try and explain how I did it.
The first thing I thought I should do was to make a demultiplexer, that
is, transform the 8 possible states for the cylinders into a means for
directing airflow down exactly one of eight possible paths. These 8
paths are the exit valves on the black switches connected to Cin. What
I then wanted to do was "OR" particular groups of these together to
create the sum and carry outputs. There happen to be 4 combinations of
A, B, and Cin that result in a non-zero Cout, and 4 combinations of A,
B, and Cin that result in a non-zero Sum. Thus, it would require 3
two-input OR gates for each of them to collect those values into a
single output. Remembering what you told me last time about pneumatic
violations, I realized that since air was only going to be coming out of
*ONE* of the exit valves connected to the piston on Cin, that I might be
able to use just a regular pneumatic switch to function as a two-input
OR gate, as long as I made very, very, very sure that whichever side of
the switch air *might* be coming from, that would have to be the same
side that the switch itself would be set to, so that it could direct the
air flow into what would normally be the entrance valve.
The hardest part of drawing this was making sure that every single
relevant switch was set in the right position to collect airflow while
still balancing them across all 3 pistons. I wasn't even absolutely
sure when I first started drawing it that I would even be able to
accomplish it (I already knew it would be possible if I didn't worry
about balancing the switch load across the pistons though). I've been
wondering if this lack of thorough up-front design is the reason that
the diagram is so hard to follow.
Anyways, I'm pleased to hear that the implementation actually works in
practice and not just in principle.
>> Mark
|
|
Message is in Reply To:
 | | Re: The latest rage in pneumatic computing
|
| (...) I was concerned about your use of switches that is backwards from normal. I transitioned through all the possible values and it never lost system pressure. Great job! Now I have to study the design so I understand how you did it. There's are a (...) (21 years ago, 23-Jun-03, to lugnet.technic, lugnet.robotics)
|
9 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
Active threads in Robotics
|
|
|
|
|
|
