#robotics Logs
May 02 2017
#robotics Calendar
09:23 AM rue_house: 555
09:25 AM rue_house: 2 1n4148 diodes
09:25 AM rue_house: a 1k resistor
09:25 AM rue_house: a .01uF capacitor
09:25 AM rue_house: mmm and a 5k variable resistor
09:25 AM rue_house: fan + to +12
09:25 AM rue_house: fan - to pin 7...?
09:25 AM rue_house: yup
09:25 AM rue_house: capacitor between 2 and ground
09:25 AM rue_house: 6 to 2
09:25 AM rue_house: 4 and 8 to positive
09:25 AM rue_house: 1 to negitive
09:26 AM rue_house: diode 1, cathode to 2
09:26 AM rue_house: diode 2, anode to 2
09:26 AM rue_house: diode 1 to 1k
09:26 AM rue_house: 1k to 3
09:26 AM rue_house: diode 2 to 5k variable
09:26 AM rue_house: 5k variable to 3
09:26 AM rue_house: --done--
09:27 AM rue_house: if no workie, spin around d1 and d2
09:27 AM rue_house: veverak, ok?
09:37 AM veverak: sounds co
09:37 AM veverak: cool
09:41 AM causative: curious: is there something like hydraulics, but with wires in place of fluid? I'm thinking of a central spinning disc driven by a motor, and wire ends are pressed against it by solenoids, and the disc drags the wire end a short distance by friction, pulling the wire
09:46 AM rue_house: http://ruemohr.org/~ircjunk/robots/arm10/p1080873.jpg
09:46 AM rue_house: causative,
09:46 AM rue_house: http://ruemohr.org/~ircjunk/robots/arm10/p1080872.jpg
09:46 AM causative: what am I looking at?
09:47 AM rue_house: cable drive
09:47 AM rue_house: wire in pipe pulling on mechanical parts of a robot
09:47 AM rue_house: off to work, bye!
09:48 AM causative: the questions I would have about that is, does it use one central drive motor to power multiple wires?
09:50 AM causative: looks like multiple motors, I was thinking there could be a weight advantage to have only one motor
09:52 AM causative: and the wires engage/disengage from it
10:05 AM Anniepoo_: causative, that's rue's robot
10:06 AM Anniepoo_: but I think the cables just move the weight towards the base of the kinematic chain
10:08 AM causative: I see what looks like several motors in rue's robot
01:37 PM jandor: hi guys
01:37 PM jandor: and ladies
01:38 PM robotustra: hi
01:39 PM zhanx: Afternoon
02:07 PM anniepoo: hey jandor
02:14 PM jandor: anniepoo, hi )
03:15 PM anniepoo: snail shell's all fiberglassed - busy with other stuff now, will try to pull it in a couple days
03:17 PM robotustra: want a house keeping robot
03:17 PM robotustra: or two
03:17 PM robotustra: or even 3
03:17 PM Snert_: I'd be happy with just a central vacuum
03:20 PM robotustra: Snert_: your requierments are too low
03:21 PM Snert_: state of the art has to get better before I could be happy with a robot in my home.
03:22 PM Snert_: it's has to look good. Like a piece of furniture that's supposed to be there.
03:22 PM anniepoo: hmm.... imagining 'central' like the phone company - One ginormous pump somewhere in the city, and pipes to everybody's house =8cO
03:22 PM Snert_: or totally hidden.
03:23 PM anniepoo: I like my roomba
03:23 PM Snert_: central vac is where the vacuum cleaner is mounted in the garage and piped out to usually 2 different places in the house.
03:23 PM Snert_: very quiet.
03:23 PM anniepoo: yes
03:23 PM anniepoo: they were popular in the 70's
03:23 PM anniepoo: too many instances of stuff getting stuck
03:24 PM anniepoo: but you can still get them
03:24 PM robotustra: roomba is dumb, I like my neato
03:26 PM Snert_: whi is roomba dumb?
03:27 PM robotustra: because it has an intellect like a pingpong ball
03:27 PM anniepoo: I think they didn't want it to be prone to brittle failure.
03:27 PM robotustra: neato has lidar
03:27 PM robotustra: and covers all floor uniformely
03:27 PM anniepoo: lots of 'things' happen in environment - people and pets walk around, if it bumps into a cardboard box the box moves,
03:28 PM anniepoo: confused by things like curtains, plants that are flexible,
03:28 PM anniepoo: people pick the unit up and set it down
03:29 PM robotustra: https://www.youtube.com/watch?v=Cq5W17t9ygU
03:30 PM anniepoo: go for it
03:30 PM anniepoo: I know a couple in SF who have a 'bring your own hole' miniature golf party once a year
03:31 PM Snert_: interesting. But I really can't think that roomba fucked up their programming that badly. Something else accounts for the random motion directions.
03:31 PM anniepoo: when they remodeled their house they installed a few tubes for built in holes
03:31 PM anniepoo: no - it's not buggy programming - it's intentionally random
03:32 PM robotustra: roomba can do good too, but it takes more time
03:32 PM anniepoo: randomness is often a good thing in programs - it tends to 'iron out' the worst cases
03:32 PM anniepoo: we use random data structures at work
03:34 PM robotustra: I also use RNG sometimes
03:36 PM robotustra: random means khaos
03:36 PM robotustra: kaos means enthropy
03:36 PM robotustra: enthropy means death
03:58 PM discardedes: anniepoo: hows that? that sounds interesting.....
04:00 PM anniepoo: back
04:00 PM discardedes: "<anniepoo> randomness is often a good thing in programs - it tends to 'iron out' the worst cases"
04:00 PM discardedes:
04:00 PM anniepoo: yup
04:00 PM discardedes: how do you use it?
04:00 PM discardedes: can you explain?
04:00 PM discardedes: that sounds interestig.
04:00 PM anniepoo: example - randomly linked list
04:00 PM discardedes: ok
04:00 PM discardedes: so its not sorted?
04:01 PM anniepoo: imagine normal linked list
04:01 PM anniepoo: but I want faster than O(n) access to middle
04:01 PM discardedes: ic
04:01 PM discardedes: so you are searching through it, one by one,
04:02 PM anniepoo: so when I make a cons cell, I pick a random number 0.. log2(n)
04:02 PM discardedes: and random can make it so that if there are regularities in what is being accesed that its random from the perspective of looking through the list....
04:02 PM anniepoo: where n is size of structure
04:02 PM discardedes: ok
04:03 PM discardedes: am I off base? or about correct in my assesment, you think?
04:04 PM anniepoo: and I make my cons cell have links not only to the next cell, but to the 2^1, 2^2 ... 2^k next cell
04:04 PM anniepoo: where k is floor(my random sample)
04:04 PM discardedes: ok
04:05 PM anniepoo: my access pattern is now much like a tree
04:05 PM discardedes: so.... yeahh.... id have to think about that
04:05 PM discardedes: you are sort of greating a system of links for one list that resembles a tree i guess
04:06 PM discardedes: you are sort of *creating* a system of links for one list that resembles a tree i guess
04:06 PM discardedes: how do you use the random number?
04:06 PM anniepoo: yes, and the insert time isn't bad - you can just remember the links for each 'level'
04:06 PM anniepoo: and patch them up
04:07 PM discardedes: ic.....
04:07 PM discardedes: thats interesting
04:07 PM discardedes: never would have thought to try such a thing
04:07 PM discardedes: what about that compared to say, making K constant?
04:07 PM discardedes: (out of curiousity)
04:08 PM anniepoo: if you make K constant, you have a rather inefficient structure
04:09 PM discardedes: do you understand why?
04:09 PM anniepoo: yes
04:09 PM anniepoo: if K is low, you have a linear access-like strucutre like a list
04:09 PM discardedes: yeah
04:09 PM anniepoo: if K is high, you waste memory, because the 'big jump' is available from every cell
04:09 PM discardedes: hmmmm......
04:10 PM anniepoo: this structgure is a time/memory tradeoff
04:10 PM discardedes: so, we might say that there is some optimal K, possibly, theoretically, but the usage of randomness makes it always pan out favorably?
04:10 PM discardedes: anniepoo: ic
04:10 PM discardedes: like memoization
04:11 PM discardedes: this is very interesting to me
04:12 PM discardedes: for example, how chaos might be used to *create* order.
04:13 PM discardedes: like, the more random numbers you have the more the spacing is regular.
04:13 PM discardedes: if i generate a billion random numbers from 1 to 10,000; i think i should have a pretty regular spacing
04:13 PM discardedes: so.... its like..... there is order in randomness.....
04:15 PM anniepoo: yes
04:15 PM anniepoo: this is what casinos do
04:15 PM discardedes: yeah... thats very cool that you have a way to use that which you understand.
04:15 PM anniepoo: they have a very regular income -
04:15 PM anniepoo: because they have thousands of players
04:16 PM discardedes: yeah.... from the perspective of casino odds, i think i understand how it works.
04:16 PM discardedes: that if there is a strong theoretical basis for given odds, they can always skim on the money flowing through.
04:16 PM anniepoo: Insurance, gaming, and such systems use the law of large numbers
04:16 PM discardedes: yeah.... its much differet to think in the aggregate like that
04:18 PM discardedes: but compare that to say, historic macro-economics, which is another interesting way to think. but doesnt seem to rely on randomness, but that such things must be true if everyone has to play by the same rules, or with the idea that we know the rules that we are all playing by.
04:19 PM discardedes: i guess it might boil down to, even within randomness there is probability.
04:20 PM discardedes: so the rates at which occurances happen has some kind of fixedness about it
04:20 PM discardedes: idk
04:21 PM discardedes: just throwing some thoughts out there
04:21 PM anniepoo: no, macro-economics depends on randomness
04:21 PM discardedes: yeah?
04:21 PM anniepoo: your model that says things are improving for business depends on a bunch of noise cancelling out
04:21 PM discardedes: is that micro-economics?
04:22 PM anniepoo: maybe my particular business is going to fail because I run a restaurant on a road and the state builds a new bypass
04:22 PM discardedes: like, economics from the perspective of within a firm?
04:22 PM anniepoo: but on the average, such things even out - your paving company prospers because the state is building a highway
04:23 PM discardedes: oh, ic
04:23 PM discardedes: yeah.... makes sense
04:23 PM discardedes: seems obvious enough atm
04:23 PM discardedes: :)
04:24 PM anniepoo: committing a crime is na individual decision - but reliably, tough economic times produce petty crime
04:24 PM discardedes: yeah.... but it wont predict that this or that person is going to commit a crime.
04:25 PM anniepoo: correct
04:26 PM discardedes: which is what makes it interesting..... it is a way to generate order in a.... sort of oblique way. Or like, probability (an randomness) is a way to think about and generate order in an oblique way.
04:27 PM discardedes: its not dirrect, in the sense of, these are the peices and im going to put them together and show you how the theory is airtight. Its more like, over an aggregate, this is the aggregate that floats atop the sytsem, both engagged and disengaged.
04:29 PM discardedes: similar to... you can if you look at its direction, you cant know its location, and if you look at its location you cant know its direction.
04:29 PM anniepoo: if you throw one die, it's impossible to guess it's value
04:29 PM discardedes: you cant know the specifics, but you can know the generalities
04:29 PM discardedes: yeah
04:30 PM anniepoo: if you throw 1,000,000 the sum is very close to the theoretical value
04:30 PM discardedes: yeah
04:31 PM discardedes: and if i know that and bet on 1/7 odds on a 6 sided die for every game, i can make some money!
04:31 PM discardedes: :-O
04:31 PM discardedes: bbiab
04:47 PM discardedes: anniepoo: me thinks that this might be good to consider for anything with a very high number of cases, each of which is anywhere from tricky to impossible to predict.
04:47 PM discardedes: and, also.... where there is a high volume of occurances to account for
04:47 PM discardedes: so... i can see why it makes sense for example that a robot could benefit by using randomness within the programming.
04:48 PM discardedes: a robot is constrantly getting input... so it makes sense that it would, play the odds, and try random things within the odds....
04:49 PM discardedes: an ai robot that operates in an environment, anyway, would have that characteristic
04:50 PM discardedes: and not only is the input high volume, but also anywhere from tricky to impossible to predict.
04:54 PM synja: precision in uncertainty
04:54 PM discardedes: yeah... its an interesting idea
04:54 PM synja: i use stochastic variables in my ai and it works wonders
04:54 PM discardedes: it also hints that the higher you could make the througput, the more stable you could make the thing operate.
04:55 PM discardedes: wow... thats cool to hear synja
04:55 PM synja: randomness is a last ditch effort
04:55 PM discardedes: ok
04:55 PM * discardedes look up stochastic
04:55 PM synja: basically just random
04:56 PM discardedes: ha! ok
04:56 PM discardedes: :D
04:59 PM discardedes: trying to make some colliodal silver for my first time today
04:59 PM discardedes: not exactly robotics related, but ehhhh......
05:00 PM discardedes: learned how different the conductance (or resistance) is between distilled and filtered water
05:00 PM discardedes: BIG difference
05:04 PM discardedes: i didnt measure the resistance difference, but it seems like you could basically consider distilled water as an insulator.
05:04 PM Tom_L: it is
05:04 PM Tom_L: non conductive
05:04 PM discardedes: yeah.. that was very interesting to learn
05:04 PM Tom_L: no minerals in it
05:05 PM synja: ^
05:05 PM discardedes: i tried running electricity througgh the well water and it fricken made a burn mark in one of the rods.
05:05 PM discardedes: and... the water was cloudy within a few minutes, and i was wiping liquified silver from that rod
05:06 PM discardedes: to dry it off...
05:06 PM discardedes: which appaertly indicates... TOO MUCH CURRENT
05:06 PM discardedes: at only 4.75 volts too
05:07 PM discardedes: switched to distilled water, its been going for about three hours now, and the water is still crystal clear, which is supposed to be a good sign.
05:07 PM discardedes: supposed to mean that the particals are smaller
05:08 PM Tom_L: that or it's not doing anything at all
05:08 PM discardedes: well... i can taste it afterwards for one kind of test
05:08 PM discardedes: should taste metalic
05:08 PM discardedes: another might be to try killing microbs with it
05:09 PM discardedes: like... say, put some in water with algae and see if it kill it
05:10 PM discardedes: using an fish-tank airpump to mix the water, which was a clever idea that has two other applications i want to use it for.
05:21 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/etching/newtank2.jpg
05:22 PM discardedes: Tom_L: definately has a metalic after-taste
05:22 PM discardedes: Tom_L: wish i could view jpg from my terimal
05:22 PM discardedes: :(
05:23 PM discardedes: i didnt taste it when i took the tea-spoon, but im tasking it now
05:23 PM discardedes: could have to do with the distilled water though too, no?
05:23 PM discardedes: ive heard distilled water has a metalic taste
05:25 PM discardedes: i think another thing about distilled water is that it sort of "wants" to have disolved metals in it?
05:25 PM discardedes: like, it has a higher potential to disolve metals or somesuch
05:28 PM discardedes: lets say, you put sugar in water, and put some yard dirt into it. bet something would grow in it eventually. then you could see what effect the solution had on it.
05:28 PM discardedes: (very scientific, eh) :-P
05:35 PM Tom_L: anniepoo, https://www.monoprice.com/product?p_id=15365&gclid=CJetz7Gb0tMCFQMJaQod2xMEqQ
06:33 PM logicmoo is now known as dmiles
07:36 PM rue_house: Tom_L, thats an ancient picture
07:37 PM rue_house: causative, I'm back
07:38 PM rue_house: causative, I didn't use clutches, there too much trouble and more expensive and heavy, and if that multiplex the motors, you cant do proper compound moves
07:38 PM rue_house: veverak, did you make the pwm generator?
07:40 PM veverak: rue_house: found out I don't have 555
07:40 PM veverak: nor diodes or capacitator
07:40 PM veverak: but had arduino without usage and 2xTIP120
07:43 PM rue_house: any 556?
07:45 PM veverak: nope
07:45 PM veverak: rue_house: but hey, got finished working pcb now
07:48 PM veverak: 1 day job to fully working thing
08:03 PM rue_house: you got your pwm going?
08:04 PM veverak: yep
11:52 PM rue_house: maybe next time I wont offer help like that, wasted thought
11:57 PM Tom_L: aww
11:57 PM rue_house: its a cool circuit tho, I keep finding uses for it