| |
In lugnet.org.us.lugola, Mark Bellis wrote:
> In lugnet.org.us.lugola, Kevin L. Clague wrote:
[snip]
>
> Kevin L. Clague wrote:
> > I don't know if you consider my stuff beautiful, but I have made a lot of
> > interesting things using pneumatics:
> >
> > My first pneumatic biped:
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=38069
> >
> > My second pneumatic biped:
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=46671
> >
> > Two pneumatic adding machines:
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=38075
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=38283
> >
> > Some pneumatic creatures:
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=38103
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=38496
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=48796
> >
> > Some intro images:
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=46353
> >
> > Some basic computing circuits and pneumatic engines:
> > http://www.brickshelf.com/cgi-bin/gallery.cgi?f=44968
>
> Very nice models!
Thanks!
>
> I have made all the standard electronic logic gates from pneumatic parts and got
> over the leakage problems.
>
> The OR/NOR/AND/NAND gate uses 6 valves and 2 cylinders - each input cylinder
> pushing 3 valves. The tub wiring is a bit of a rats' nest! The same gate does
> all 4 functions depending how you wire it. Reverse the output to go from OR to
> NOR or AND to NAND and reverse the inputs to go from OR to NAND or AND to NOR,
> just like electronic gates! The cunning bit is that the air takes 2 parallel
> routes when the inputs are 01 or 10.
Mark Terrabain invented a single piston, single switch AND gate, but it only has
one pressure output, so you have to retract the pistons with some external force
like rubber bands.
I invented a single pison, dual switch, single pressure output OR gate.
I've got dual pressure versions as well.
I can get AND down to single piston with three switches. OR can also be done
with one piston and three switches.
Mark Terrabain, got a half adder down to two pistons and a large number of
switches. He is amazing.
>
> The XOR/XNOR/Parity gate uses one cylinder and one valve for the first stage and
> 2 cylinders and 4 valves for the second. The key principle is to swap the air
> connections without any leaks occurring - the 'exhaled' air has to be routed
> right back to the first stage. A Parity gate uses more 2Cyl+4Valve stages in
> the same way as the second stage, though I haven't built one yet.
You are talking about DeMorgan's law here.
I can make an XOR gate with one piston and four switches. I can make an AND/XOR
with one piston and 7 switches.
>
> A flip-flop first appeared in the alternative model of 8868 Crane Truck. It
> uses 2 cylinders and 2 valves and goes through the sequence 00,01,11,10 (gray
> code!), where 0 is down and 1 is up. The crane truck was what first got me
> interested in pneumatic logic.
Yes. I've used this one a lot. It is a fundamental building block of self
actuating pneumatic sequencers.
>
> Further pairs of cylinders and valves can be added in the flip-flop sequence. I
> drew the diagram for a six-legged robot that would move one leg forad at a time
> then all six back together. The main loop is a seven-set flip-flop and each leg
> is a 2-set flip-flop with the last movement ynchronising with the main loop.
> The seventh stage in the main loop reverses the whole system for the
> simultaneous backward movement of all legs together.
but.... did it work?
Up until about 5 months ago, my pneumatic sequencers were acheived by trial an
error.
It took me quite a while to get my first caterpillar to actually work.
I had an epifany on a flight to Boston, and now I have a design methodology for
pneumatic sequencers. My Inchworm is a prime example of that effort. It worked
the first time.
>
> I used flip-flops and an XOR gate to make a 10-step pick and place robot. The
> robot extends, grabs a pen, retracts, turns at the elbow, turns at the wrist,
> extends, drops the pen in a pot, retracts, turns back at the elbow, turns back
> at the wrist and loops round again continuously. This is an ideal application
> for the air compressor as six pump cylinders soon result in aching hands!
> Notice that the extension/retraction happens twice in the sequence. This is
> what the XOR gate does. Two large cylnders in parallel with the 2 small ones on
> the grab move the 4-valve part of the XOR gate. The limit for 1 cylinder moving
> valves is about 3 valves.
Have you figured out how to make a MUX using just switches? It is easily done
with just four switches. The four switch MUX is the key component in my XOR
gates.
Yes. I used to have to use the dual extension/retraction method.
The design methodology I used in my Inchworm can eliminate the need for all the
dead time recyling the flops.
>
> I did the logic for a four-joint Octopus arm that would place its arm on an
> object, curl its arm around it, pick it up and uncurl its arm in reverse
> sequence. There are 4 cylinders and 4 valves on the arm and 12 cylinders and 20
> valves on a 32x32 board at the side!!! The logic for reverse sequences is
> horrendous (especially avoiding leaks).
Sounds complicated. Have a peek at my inchworm's design methodology. It may
save you some complexity. If you give me the binary sequence that you want for
the arm, I can try to give you an optimized circuit.
>
> I've also succeeded in getting pneumatic cylinders to stop in the middle (for
> Excavator steering etc...). This requires a set of 3 valves driven by 2
> cylinders, where the valves are not all in the same position at once (ie.
> there's a smaller offset between their levers). They're sprung together with
> small white belts. This system is under further development and when I get a
> backhoe excavator I'll cross it with the 4x4 jeep to make a 4x4 4-wheel
> pneumatic steered excavator with self-centring steering and backhoe!
You can easily get a three state mechanism by using two cylinders hooked
together at the base. The three states are both expanded, both contracted, and
one but not both expanded.
I used this to work on a pneumatic biped that only shifted weight forward, not
side to side. It could never make the transition to the other foot, so I took
it apart and need to start over. It would hold one foot directly under the
center of gravity, while the other foot sequenced through bringing the up foot
from the front to the back so it could be weight bearing.
>
> Sorry, no pictures yet but there will be when I get a digital camera! Meanwhile
> I'm happy to discuss things and help out.
>
> Mark Bellis
It is great to meet more people interested in pneumatics!
Kevin
|
|
Message is in Reply To:
 | | Re: Lego pneumatic systems overview?
|
| (...) For models with a lot of cylinders I use a car tyre air compressor. I bought the slightly more expensive one tha has a variable pressure limiter and set the limiter to 20 PSI initially. This wasn't quite enough. 25 PSI was good for all but the (...) (21 years ago, 26-Sep-03, to lugnet.org.us.lugola, lugnet.technic)
|
21 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
