Connection hit a new low today - 4800 bps. No, I did not drop a zero or two. Office is not wired with twisted pair and the ambient electromagnetic background is noisy. Very noisy.
It was like a time warp to 1979.
I worked out how to chain my sence whisker gearing together in such fashion as to act like simple neurons. This immediately makes me think of changing the design from a wheeled crawler to a hexapod walker.
As a quick count, I should have a touch sensor on the belly, and one on each shin, and one of each foot.Plus probably a couple to sence 'muscle' contractions.
From a construction standpoint, the only fiddly piece is the mutilated ratchet ; I shall have to devise a method for constructing them en masse.
In terms of a walker, I am rather struck with the appearance of RHex.
Searched 'Rod logic' and synonyms.... found a depressing lack of content. Mainly noise, a lot of references to Drexler, a couple to Fowler.
(Fowlers the 'inventor' of central heating as we know it, and also of rod logic. Circa 19th Century.)
Doesn't look like anyone has done much work on them. Beyond discussing volume versus electronics volume (a popular theme - Mickey Mouse beats Donald Duck kind of calculation), or the thermodynamic implications thereof.
Only links I could find of value were:
Clever lego work:
http://goldfish.ikaruga.co.uk/logic.html
Or academic work:
http://www.zyvex.com/nanotech/mechano.html
Everyone seems to think that RAM is the toughie for any practical implementation.
Would seem simple to adapt the gates to read and punch paper tape. I wonder how much tape my crawler could carry. Or perhaps, it should string tape from the ceiling, and crawl up and down it as it computes, like some sort of algorithmic spider.
What are the implications of a rod with adjustable studs?
(I'll avoid the joke about the stud with the adjustable rod.)
(Fowlers the 'inventor' of central heating as we know it, and also of rod logic. Circa 19th Century.)
Doesn't look like anyone has done much work on them. Beyond discussing volume versus electronics volume (a popular theme - Mickey Mouse beats Donald Duck kind of calculation), or the thermodynamic implications thereof.
Only links I could find of value were:
Clever lego work:
http://goldfish.ikaruga.co.uk/logic.html
Or academic work:
http://www.zyvex.com/nanotech/mechano.html
Everyone seems to think that RAM is the toughie for any practical implementation.
Would seem simple to adapt the gates to read and punch paper tape. I wonder how much tape my crawler could carry. Or perhaps, it should string tape from the ceiling, and crawl up and down it as it computes, like some sort of algorithmic spider.
What are the implications of a rod with adjustable studs?
(I'll avoid the joke about the stud with the adjustable rod.)
Yesterday was the fifth anniversary of Mir's re-entry. Easily scorned as old and battered technology ; easily ignored as a relic of the Cold War and hardly relevant in the new 'Culture War' we find ourselves enmired in.
But.
The distinction between 'ship' and 'station' is somewhat artificial. The Mir was a ship that carried its crew around and around the Earth for millions of miles. If it had a spirit, it was battered and creaky - but game.
How much work could have been done in this amount of time if a private company could have bought or leased it? As much - or more - research than has has been done on the ISS in this time, I suspect.
How much would it have cost to buy?
How much for a unmanned launch to rendezvous and boost the Mir to a higher, longer orbit?
How much of a spur to private spacecraft would it have been, for it to have been circleing overhead, waiting?
How much a spur to amateur science, to have this acsess to orbit?
But, like Skylab, it was built by the economies of destruction - and in its time, fell to the same economies.
Ah well.
Mir, we lift a glass; to Mir, the circling, and the fallen.
High point of my week;
Discovering that in E1M8, after grabbing the Pentacle, the Shambler's Lightening Attack can be used to levitate up
to the top platform.
Hard part: Landing. The @#$%& Shamblers won't stop attacking, and I tend to float to the ceiling until the Pentacle runs out.
'E1M8 --what?'
Quake One - the First - The Best and the Only...of course.
Discovering that in E1M8, after grabbing the Pentacle, the Shambler's Lightening Attack can be used to levitate up
to the top platform.
Hard part: Landing. The @#$%& Shamblers won't stop attacking, and I tend to float to the ceiling until the Pentacle runs out.
'E1M8 --what?'
Quake One - the First - The Best and the Only...of course.
Deregulating the sences...
While at work - which entailed standing and staring into space for six hours - I conceived the linkages nescessary for the bump sensors.
Using a ratchet, a pawl and an escapement, I can amplify a gentle push into a axle rotation with considerable torque.
I'm pleased that this rotation trigger eliminates the need for the touch sensors to work according to a clockbeat - even if it depends upon a movement beat. I don't mind a movement beat; if anything it adds aesthetic pleasure.
The real world does not work according to a precise clock, so sences that depend on any form of shuttering should suffer from 'event artifacts' or 'moment artefacts' to coin phrases.
It is probable that our perception of movement when presented with 30 still frames per second is a form of 'event artefact' ariseing from the interplay of our pulsatile neural system and the continuous world.
Co-Operative Vs. Lockstep
Timing can be critical.
If the sensory pulse is used to clutch and unclutch the drivetrain, than in the interest of least effort and minimizing wear, the shifts should be done while the drivetrain is not under load. If the mainspring regulator is pulsatile rather than continous, then sensory switching can be done during the 'null' portion of each beat. Of course, this means that the senses will beat counterpoint to the 'muscles'.
The question: is it more elegant to use one system clock, and than a pair of divide-by-two mechanisms to generate the 'motor' clock and the 'sense' clock ; or should there be a alternating mechanism, so that one beat of the 'motor' clock triggers one beat of the 'sence' clock, which in turn triggers the 'motor' clock?
Co-Operative Vs. Lockstep : Phrased like this it sounds like Dogma rather than Craft.
Timing can be critical.
If the sensory pulse is used to clutch and unclutch the drivetrain, than in the interest of least effort and minimizing wear, the shifts should be done while the drivetrain is not under load. If the mainspring regulator is pulsatile rather than continous, then sensory switching can be done during the 'null' portion of each beat. Of course, this means that the senses will beat counterpoint to the 'muscles'.
The question: is it more elegant to use one system clock, and than a pair of divide-by-two mechanisms to generate the 'motor' clock and the 'sense' clock ; or should there be a alternating mechanism, so that one beat of the 'motor' clock triggers one beat of the 'sence' clock, which in turn triggers the 'motor' clock?
Co-Operative Vs. Lockstep : Phrased like this it sounds like Dogma rather than Craft.
It should be possible to;
Use an electroplating rig to remove solder from old circuit boards, freeing up the components without heat and nasty solder fumes. Or perhaps a bath of dilute sulfuric acid, and an extra electrode?
Using the Pb + H2SO4 = PbSO4 + H2 + 2e reaction to generate (miniscule) amounts of power, and to recycle the parts.
Will dilute H2SO4 eat PCB material?
Will have to try this one out. Even if it doesn't work, the idea of spending an afternoon with Oil of Vitriol and Lead
sends a tingle right through my Mad Scientisty bits.
woooo chemical formulae. Brings me right back to my misspent youth. Ah, the smell of burning eyebrows...
Use an electroplating rig to remove solder from old circuit boards, freeing up the components without heat and nasty solder fumes. Or perhaps a bath of dilute sulfuric acid, and an extra electrode?
Using the Pb + H2SO4 = PbSO4 + H2 + 2e reaction to generate (miniscule) amounts of power, and to recycle the parts.
Will dilute H2SO4 eat PCB material?
Will have to try this one out. Even if it doesn't work, the idea of spending an afternoon with Oil of Vitriol and Lead
sends a tingle right through my Mad Scientisty bits.
woooo chemical formulae. Brings me right back to my misspent youth. Ah, the smell of burning eyebrows...
Clockwork Robots
I use 'Robot' rather than 'Automata' because the idea is an all-mechanical 'Bump-n-Go' explorer style robot.
Useing the bumper sensors to trigger clutches rather than microswitches is an interesting problem. Mechanical amplifiers will probably have to be installed in the sensors, in order to take the impact energy and ramp it up to a level that can drive clutches.
A mechanical amplifier and a diaphramn would allow some level of reaction to loud sounds.
Useing sunshades and bimetal strip style thermometers would allow the robot to sence local tempurtures, and engage in sunseeking 'Photovore' behavior. Ultimately, a parabolic mirror and a small Stirling engine might be able to wind the robot's mainspring back up.
Fluidics would probably be the easiest way to do this, but I am very tempted by the idea of brass clockworks ticking and humming, the sunlight flashing from the sharp edges of gearteeth.
Clockwork being impulsive by its very nature, each robot would have its own heartbeat. It could also be stopped, and be dead until a given sensory stimulus bringing it back from the dead. At least until the lubrication seized from the passage of time.
A electricity-free robot , crawling around the room, seeking sunlight and avoiding loud people. Sounds like the defination of a housecat.
Useing the bumper sensors to trigger clutches rather than microswitches is an interesting problem. Mechanical amplifiers will probably have to be installed in the sensors, in order to take the impact energy and ramp it up to a level that can drive clutches.
A mechanical amplifier and a diaphramn would allow some level of reaction to loud sounds.
Useing sunshades and bimetal strip style thermometers would allow the robot to sence local tempurtures, and engage in sunseeking 'Photovore' behavior. Ultimately, a parabolic mirror and a small Stirling engine might be able to wind the robot's mainspring back up.
Fluidics would probably be the easiest way to do this, but I am very tempted by the idea of brass clockworks ticking and humming, the sunlight flashing from the sharp edges of gearteeth.
Clockwork being impulsive by its very nature, each robot would have its own heartbeat. It could also be stopped, and be dead until a given sensory stimulus bringing it back from the dead. At least until the lubrication seized from the passage of time.
A electricity-free robot , crawling around the room, seeking sunlight and avoiding loud people. Sounds like the defination of a housecat.
Subscribe to:
Posts (Atom)
