#robotics Logs

Nov 20 2015

00:18 orlock rue_more: nope, just the standard arduino i2c stuff
00:18 orlock rue_more: Not doing anything much with it at the moment, its more to get the hardware to a point that i know it all works together so i can put it in an enclosure
00:19 orlock eventually the eeprom will be for storing blinky-light patterns
00:19 orlock (and wifi access details i guess)
00:24 rue_more oh its arduino
00:24 rue_more your librry or the arduino one?
03:45 deshipu http://24.media.tumblr.com/tumblr_lxx3az2FW41qdf6a3o1_500.jpg
03:54 SpeedEvil http://imgur.com/gallery/xFkhofN
04:19 Jak_o_Shadows I wonder how much the maths changes when you go to rotary actuated/sensored stewart platforms
06:36 akem hi
06:38 akem keukpa so i got questions about shoulders and fingers, not much infos on the website not fully fonctionnal cauze im on my cellphone
06:55 anonnumberanon hi
07:02 akem whats up anon?
07:07 deshipu akem: you still haven't asked your question
07:07 deshipu akem: don't keep us in the void
07:08 akem yea deshipu hello
07:10 akem keukpa what motors for the shoulder, and how many for 1 finger, cauze i did not find the details, not sure about the internal of 1 finger, the 3 hinges, what you use
07:10 akem i ve a nice one on youtube
07:12 deshipu akem: well, it all depends on what the robot has to do, how strong and how fast it has to be, and what budget you have
07:14 Snert budget being the biggee.
07:14 akem deshipu i know but about his particular 3d printed robot project
07:15 deshipu akem: which one?
07:15 akem since he did it already,the ones i saw on the vkdeo
07:15 akem wait
07:18 akem http://m.youtube.com/watch?v=W62Wfz1xqYg
07:19 akem the hand i will try to find it back arduino and puppet like wired fingers
07:19 deshipu that is a lot of plastic
07:20 Snert darn, no shoulder detail.
07:20 akem https://m.youtube.com/watch?v=t52edTD9RA0
07:21 akem 3d printed abs sir, other video with this bot but the details must be available as text since it is open source
07:22 akem the model he use i think hes alone on this projecg
07:23 deshipu could be perhaps used in a shopping window
07:25 akem http://www.linengineering.com/stepper-motors/
07:27 akem http://www.minebea.co.jp/english/news/press/2007/__icsFiles/afieldfile/2015/01/29/stepping_3.gif
07:27 akem interesting right
07:56 akem https://m.youtube.com/#/watch?v=alYKoIXIPoM
07:56 deshipu I have a whole bunch of those
07:56 akem hes not joking ladies.
07:57 deshipu http://www.aliexpress.com/item/NEW-50pcs-totally-5pcs-each-of-10-kinds-4-Wire-2-Phase-dc-micro-stepper-motor/1956059879.html
07:57 akem deshipu which size?
07:57 deshipu several different sizes
07:57 deshipu never actually got to do anything with them
07:58 deshipu I think those are mostly used for adjusting the lens in cameras
07:59 akem so cheap?? $17 ?
08:00 deshipu I think mine were even cheaper
08:00 deshipu that's CHina
08:00 akem yea thats what i ve read also
08:03 Snert deshipu: 4 wire, 2 phase. Sounds like encoding motors?
08:07 deshipu Snert: encoding motors?
08:07 deshipu Snert: with worm gear?
08:07 Snert from the link you posted
08:07 Snert funky chinese description maybe.
08:08 Snert 50pcs totally 5pcs each of 10 kinds 4 Wire 2 Phase dc micro stepper <--- this
08:08 Snert miniature steppers with encoders?
08:08 deshipu two coils, 4 wires
08:08 deshipu sounds about right?
08:09 Snert yea but I'm not sure.
08:09 Snert could be DC motors with an encoder.
08:09 deshipu they are not dc motors
08:10 Snert I've never encountered ac motors that small then
09:05 SpeedEvil They are stepper motors
09:05 SpeedEvil they are commonly used in cameras and shizzle at that sort of size
09:07 SpeedEvil Aliexpress items are a fun exercise in guesswork
09:07 SpeedEvil Or, if it's cooking gear, guesswok.
09:09 Snert I wasn't thinking of steppers as ac motors, but I guess if it alternates then it is.
09:10 Snert read between the lines with ali descriptions as you say
09:19 anonnumberanon haha
13:47 anonnumberanon got some 4017s in the mail today
13:53 akem about robotic hand, index finger, if you want to control precisely every hinges closing all of them or just the first one, keeping the finger straight
13:57 Snert_ muscle wire?
13:58 akem http://www.healthline.com/hlcmsresource/images/topic_centers/osteoarthritis/642x361-2-Finger_Bends.jpg
13:58 akem hm like tandons
13:59 akem only 1 motor for 1 finger
14:00 akem you stir the wire the finger close but thats all
14:00 akem not sure if inmoove can do better on that
14:01 Snert_ Although thin and lightweight, one of the most amazing things about Muscle Wire is that it can lift many times its weight and is able to do 100 times more work per cycle than the human muscle. This material is easy to use, small in size, operates silently, has a high strength-to-weight ratio, and is easily activated using AC or DC power. This technology is ideal where mechanics require minimization, such as electronic text
14:01 Snert_ iles projects, robotics or nano-applications.
14:02 Snert_ so I was thinking like 3 or 4 short strands at the finger first joint.
14:02 akem also only 1 axis rotation ,base of the finger on the pinky one in the picture rotate the whole finger on the Z axis（picture oriented）
14:03 Snert_ or anchored further back within the finger, but only moving the 1st joint.
14:07 Snert_ 6bucks a meter seems cheap enuff to play with.
14:14 akem Snert_ you used that already?
14:14 Snert_ no, but I'm interested in trying it.
14:15 akem http://m.youtube.com/#/watch?v=k9f-W6Xi_Wo
14:15 akem ok
14:15 Snert_ for lnly 6bucks...1 meter might make 1 joint with 4 strands or something.
14:15 Snert_ only
14:15 akem never heard about it before
14:15 Snert_ it seems to have million time repeatability.
14:16 Snert_ kinda take alot of current to heat it and make it contract, but not overly so that it's just too much current to be practile.
14:17 Snert_ flexinol
14:17 Snert_ good google
14:17 Snert_ made by dynalloy
14:19 Snert_ and it's been around for about 10 years.
14:20 Snert_ it's all heat activated.
14:21 Snert_ you can heat it with a lighter and it contracts too.
14:21 Snert_ or send a bunch of current through it to heat it.
14:22 akem if it has to be burning hot its impractical yes, but otherwise its pretty neat/incredible
14:22 Snert_ relatively speaking, it takes alot of current.....per it's size...but not so much current as to be impractical.
14:23 Snert_ and no, not that hot.
14:39 anonnumberanon i wonder the power requirements to make something usable such an entire finger that has the power of a human finger
14:40 anonnumberanon and by power i mean force and speed
14:41 anonnumberanon https://www.youtube.com/watch?v=zT1iDtDvx58
14:44 anonnumberanon https://www.youtube.com/watch?v=t52edTD9RA0
14:45 SpeedEvil Power is meaningless.
14:46 SpeedEvil Do you mean the pure isolated finger.
14:46 anonnumberanon yes of course
14:46 anonnumberanon i want to know how much length and how much current and voltage to use to replicate a finger, let's say, the strongest finger
14:46 SpeedEvil Because that's perhaps 200N*0.2s/5cm
14:46 SpeedEvil for 20kg force on the finger over 5cm distance
14:47 SpeedEvil 25cm/s, = 50 watts
14:47 anonnumberanon and im sure a single finger would need a bunch of wires together near each other
14:47 SpeedEvil Sounds reasonable
14:48 Snert_ ya know...6bucks for a yard of it. Easy to test out at the 5-10 strand level.
14:49 SpeedEvil https://en.wikipedia.org/wiki/Carpal_tunnel
14:49 SpeedEvil fucking amazing
14:49 SpeedEvil Please note - that above computation was for mechanical power of the human finger.
14:49 anonnumberanon SpeedEvil, yeah, also you'd need to find a way to strand the wires together in a way they all bend in the same direction, maybe around a couple hinges, and tubes
14:49 SpeedEvil https://upload.wikimedia.org/wikipedia/commons/f/f0/Carpal-Tunnel.svg
14:50 SpeedEvil I think the wire is at best a percent efficient, so you'd need several kilowatts per finger
14:50 anonnumberanon that $6 a meter would probably be unacceptable if one wanted to replicate a full hand
14:50 SpeedEvil perhaps 50kW/hand.
14:50 Snert_ just depends how much force ya really need.
14:51 Snert_ fully duplicating the strength of a human hand may not be required.
14:53 Snert_ I'm thinking a 2 finger lobster claw gripper maybe. In addition to the whole 5finger thing.
14:53 anonnumberanon yeah you can go for mediocre applications, or you can fully replicate the strength of a human hand
14:53 Snert_ do you need enough strength to do a pullup?
14:54 Snert_ or enough strength to crush an empty soda can.
14:54 Snert_ or pick one up full.
14:55 Snert_ prolly about halfway between the 2 for what might be called an average task.
14:55 anonnumberanon yeah id say make it weaker with enough force to hold a large coffee mug and enough speed to type on the keyboard at 60 wpm
14:55 anonnumberanon the only two things humans are good at in 2015
14:55 Snert_ beer lifts :)
14:56 anonnumberanon yeah, let's make that a pint of force
14:56 Snert_ if it could just pick up my empties and toss them in the trash.
14:57 anonnumberanon wait ill get my robot to step on them until they are crushed and then the Neato can sweep the remains
14:59 SpeedEvil Or just get a robot beer-brewer, and then get it to deliver you glasses of beer.
14:59 anonnumberanon SpeedEvil, by that efficiency you meant that if you send 1watt of energy into the wire it will only give us a hundredth of a watt of mechanical work?
14:59 SpeedEvil yes
15:00 anonnumberanon y u think it's thtat bad?
15:00 SpeedEvil Because you end up needing 5-10kW to mimic a hand.
15:00 anonnumberanon it's been done before?
15:00 anonnumberanon oh wait
15:01 SpeedEvil i haven't looked. In principle, you could do it without too much trouble, it just has very limited application
15:01 anonnumberanon by my question, i meant, why do you assume it has a 1% efficiency?
15:01 SpeedEvil because at that power you really need heavy water-cooling too
15:01 Snert_ he did the newton calculations or whatever.
15:01 SpeedEvil Because I memorised that as the approximate efficiency of muscle wire ages back
15:01 SpeedEvil In the best case.
15:01 anonnumberanon ah okay
15:01 SpeedEvil It is of course 0 efficiency if it's static at non-idle
15:02 anonnumberanon that is good enough reference for me
15:02 anonnumberanon "fellow engineer once read that"
15:03 SpeedEvil Actually computing the efficiency depends on lots of things
15:03 SpeedEvil For example, you can raise the efficiency statically if you insulate the wire, and cool with seperately switched air.
15:03 anonnumberanon hey remember oiur talks about current sensing? i just received 5 really fat resistors, 10Watts, 0.1 ohm for that purpose
15:04 SpeedEvil yay
15:04 anonnumberanon that'll let me measure all the servos of an entire leg
15:05 SpeedEvil I have just found CH/2391/2015 which is an awesome decision of the upper tribunal which proves the DM making the decision on my benefits was an idiot.
15:05 anonnumberanon ah im not aware of your legal debacles, what's going on?
15:06 SpeedEvil In short, terrible disability benefits admin.
15:06 SpeedEvil By people who are making decisions on the basis of what is fastest, not what is actually accurate.
15:06 anonnumberanon did you get rekt just like theBear?
15:06 SpeedEvil No, I happened to catch a benign virus at age 12.
15:07 eedEvil wishes there were light metals that were magnetically conduct
15:07 anonnumberanon what does it do to you, if anything?
15:07 robopal who said that just got some 4017 to play with
15:07 SpeedEvil In short, I'm exhausted all the time to the point I can't safely cook most of the time.
15:08 [cube] whats the virus called?
15:09 [cube] sounds like lyme
15:09 anonnumberanon slothinius
15:09 SpeedEvil Chronic fatigue syndrome.
15:09 robopal microweave meals and oven cooking is ok or still a risk?
15:09 [cube] thats an actual virus?
15:09 [cube] like little baddies swimming around your body
15:09 SpeedEvil Currently they have a test that can 100% tell between two groups of 100 people, one with CFS and not, which group has CFS
15:09 SpeedEvil there is no useful individual test.
15:10 [cube] im tired all the time too...
15:10 [cube] but i figure its just cause im old and sit at a computer all day
15:10 [cube] back problems probably have somethign to do with it
15:10 [cube] i noticed when i crack my back i feel less tired
15:11 robopal yeah, we need to be seen by the sun
15:11 anonnumberanon im currently doing tests but i think looking at the screeen for a very long time is tiring, and doing 1/3 screen time and 2/3 non screen work may be better
15:11 anonnumberanon like yesterday i really wanted to work on my bot, i had so many things to do to it, but no, i went to bed tired, at 4 pm
15:12 [cube] lol
15:12 [cube] i usually nap around 5
15:12 [cube] for 30 mins
15:12 [cube] keeps me going until midnight
15:12 SpeedEvil anonnumberanon: I have on good days 15-20 minutes a day I can concentrate on any novel task.
15:12 [cube] key is setting an alarm for the nap
15:12 SpeedEvil This is not 'being a bit tired'
15:12 anonnumberanon iv been considering nap cube, maybe ill try it right now
15:12 anonnumberanon how do you make it only 30 min?
15:12 SpeedEvil That task tires me as much as a couple of marathons might.
15:13 [cube] sounds psychological SpeedEvil?
15:13 SpeedEvil [cube]: Medical opinion disagrees with you.
15:13 [cube] i have some 'easy' tasks i need to do...but i reallly hate doing them...
15:13 [cube] so much so that it tires me out
15:14 [cube] and doing something cool/interesting i dont get that effect
15:14 SpeedEvil [cube]: I don't hate washing up - I find it mildly productive - yet I get so tired that I can be dangerous after 20 mins. Same happens playing an interesting game, learning an interesting thing, speaking to people
15:14 [cube] ah
15:14 [cube] i mean more in a work context
15:14 [cube] (web dev)
15:14 [cube] i have some interesting work
15:15 [cube] and some mundane boring-as-hell work
15:15 [cube] in the end its the same shit
15:15 [cube] but my perception of it affects my performance
15:17 Snert_ some intesting uses for musclewire.
15:17 anonnumberanon I'm gonna have to do something abou6t the non-linearity of these slide potentiometers.
15:17 Snert_ http://memry.com/sites/default/files/documents/Nitinol_Industrial_Applications_SMST00.pdf
15:17 anonnumberanon robopal, yeah dude i got a few 4017s
15:18 robopal SpeedEvil, I am curious, is there a task/occupation that keeps you energized/active for longer periods?
15:18 SpeedEvil robopal: Anything that involves being still, and not thinking too hard.
15:18 SpeedEvil I fucking hate sitting still and not thinking too hard.
15:19 robopal anonnumberanon, do you want to play with a 4-way-traffic-light circuit? (it also has a 555 timer)
15:19 SpeedEvil 555 logic is fun
15:20 anonnumberanon robopal, no I want to control a few dozen servos from a few microcontroller pins
15:21 robopal SpeedEvil, so lets say being here on IRC is easier for you but you hate it because it is not challenging, or did I misanderstood there
15:22 SpeedEvil It is 'least bad'. I can regurgitate facts i know with little physical or mental effort, and help others in many instances without actual thought.
15:22 SpeedEvil I want to be actually fixing my house, designing things, or ...
15:23 SpeedEvil I have fun ideas for a robo-garden tractor I want to get on with.
15:23 anonnumberanon can you pass out while standing?
15:24 SpeedEvil It's not really like that.
15:24 SpeedEvil Imagine every minute of activity is an hour.
15:24 SpeedEvil And how tired you'd be at the end of it if it was.
15:26 anonnumberanon like physically or mentally?
15:26 SpeedEvil both
15:35 SpeedEvil Does anyone know of a very cheap source of nylon gears?
15:35 Snert_ the 3d print guys?
15:35 SpeedEvil A vendor that has a selection of orderable parts, ideally, not ...
15:39 Snert_ SPD/SI - though like all things...better costs a little.
15:40 Snert_ http://www.sdp-si.com/
15:40 Snert_ there's also things like main frive gears for rc helis.
15:40 Snert_ about 2-3 inches in diameter.
15:41 robopal wow amazing site
15:42 Snert_ yea. the good stuff.
15:42 Snert_ but still not so expensive you couldn't afford it.
15:48 SpeedEvil bookmarked, thanks
16:05 SpeedEvil http://imgur.com/gallery/JdIqPUP - on the topic of beer, and german engineering.
16:09 akem open/close a trap or a locker inside a rail or whatever with just electromagnets push/pull
16:23 anonnumberanon im fighting to figure out how to hook up this 5 pin pot lol
16:23 anonnumberanon http://www.bitechnologies.com/pdfs/PSxxL_PSxxM_DS.pdf
16:23 anonnumberanon i got the led on at least
16:24 SpeedEvil measure between the other three pins.
16:24 SpeedEvil With the wiper set in the middle
16:25 SpeedEvil the maximum is the track ends.
16:25 SpeedEvil The other one is the wiper
16:25 SpeedEvil now, move the wiper to one end and...
16:25 anonnumberanon i seem to have checked every pin
16:25 anonnumberanon also i want to kill them for their datasheet
16:36 Snert_ have a beverage and do it again :)
16:36 Snert_ sometimes I'm half asleep and miss things.
16:37 deshipu this is an awesome talk from the MIT: https://www.youtube.com/watch?v=R0dgELUeBTU
16:39 robotustra do you understand formulas in this talk?
16:40 anonnumberanon called the company, they had no idea about that pot lol
16:40 anonnumberanon well they know, about it, but now why the datasheet dont have a pinout diagram for it
16:41 deshipu robotustra: not many formulas there
16:41 anonnumberanon only some generic stupid diagram with 3 pins
16:41 robotustra I think their model is too simplistic
16:41 deshipu robotustra: it works...
16:41 deshipu demonstrably
16:41 robotustra not so nice as in the nature
16:42 robotustra they should change the model of spine
16:42 deshipu it's better than anything we have so far
16:42 deshipu robotustra: they had spine initially
16:42 robotustra you are wrong
16:42 deshipu robotustra: but it turned out to be too hard to model
16:42 deshipu robotustra: in real time
16:42 robotustra there is better examples of walking
16:43 robotustra look at boston dynamics robots
16:43 deshipu robotustra: they can't jump over obstacles
16:43 robotustra and what?
16:43 deshipu mit cheetah can
16:43 deshipu also, uses 40 times less energy
16:44 robotustra did you see that? https://www.youtube.com/watch?v=wE3fmFTtP9g
16:44 deshipu yes, that's old
16:44 robotustra I think this one can jump as well
16:44 deshipu there is a reason why it has a gas engine
16:44 deshipu no it can't
16:44 robotustra because it's too heav
16:45 robotustra y
16:45 deshipu no, heavier animals can jump with less force
16:45 deshipu because it can't calculate the model for it
16:45 deshipu and because it uses rigid control
16:45 deshipu doesn't use its dynamics
16:46 robotustra may I ask you, where did you study physics?
16:46 deshipu UAM
16:46 robotustra what degree do you have
16:46 robotustra ?
16:46 robotustra master?
16:46 deshipu ad hominem much?
16:47 deshipu you don't want to talk, just don't talk
16:47 deshipu you want to discuss it, discuss it
16:47 robotustra I'm discussing
16:47 robotustra just curious how deep you are into physics
16:48 deshipu and the degree tells you that?
16:48 robotustra I'm doing the robot right now and I want to use the same technique
16:48 robotustra the level of your knowledge
16:48 deshipu the same as mit or the same as bd?
16:48 robotustra I'm PhD in physics
16:49 robotustra so I understand both modelsa quite good
16:49 deshipu so which one do you plan to use?
16:49 deshipu the trajectory optimization one, with stable cycles et all?
16:50 deshipu what algorithm are you using for solving the optimization?
16:50 robotustra I'm planning to do stable cycle but with elastic elements
16:50 deshipu are you planning to use*
16:51 robotustra and optimization of movement will be done programmatically
16:51 deshipu with a computer
16:51 robotustra yes
16:51 deshipu that's novel
16:51 robotustra with the code
16:51 deshipu who is going to write the code for you?
16:52 robotustra myself
16:52 deshipu so how do you plan to solve the optimization?
16:53 robotustra easy, I'll have current feedback from motors, by minimizing it I'll get optimal motion
16:54 deshipu you said you are going to do it with code, you don't have current feedback in code
16:54 robotustra in a 50-100 iterations I'll get an optimal trajectory
16:54 robotustra you are funny
16:54 robotustra robots have sensors
16:54 Snert_ current sensors on the motors then read the sensors.
16:54 robotustra to make loop control
16:54 deshipu by "iterations" you mean which of the optimization algorithms?
16:54 Snert_ ?
16:55 deshipu I mean, how do you decide which parameters to change and how bwteeen the iterations?
16:55 robotustra for instance, I have a hand, which has a lon of DOF and which moves from point A to point B
16:56 robotustra let say I have 5 motors moving together
16:56 robotustra and they are dependent
16:56 robotustra and I have a bunch of configuration of arm between these 2 point
16:57 robotustra it means I have some vector of forces applied to the hand
16:57 robotustra depending on time
16:57 robotustra Every moution from A to B I call iteration
16:58 deshipu well, 5 vectors, and that's a fully actuated system
16:58 deshipu a mobile robot is not fully actuated
16:58 robotustra each ineration I get the total energy spent
16:58 deshipu sorry, go on
16:59 robotustra it mean I do 1 pass, get the coefficients, and total energy, the second pass etc
16:59 deshipu how do you decide how the second pass should differ from the first one?
17:00 robotustra after that I use genetic algorythm to find optimal set of coefficients with respect to energy
17:00 robotustra do you know what is monte carlo approach?
17:01 deshipu yeah, I know monte carlo and I know evolutionary algorithms and I also know they are two different things
17:01 robotustra you change one of parameter randomly, do slight deviation
17:01 deshipu well, both are optimiaztion algorithms
17:01 deshipu randomly, I see
17:01 robotustra with some limits of cause
17:02 robotustra a number of iterations will give almost optimal motion
17:02 deshipu and the robot will run through 100 obstacles and jump over the 101st, provided they are all identical
17:03 robotustra obstacles doesn't matter they solve particular task, my task is different a bit
17:03 robotustra I want to optimize hand movements
17:03 deshipu sure, if you are programming factory robots that just do the same thing all the time, you don't even need that
17:04 SpeedEvil deshipu: https://www.youtube.com/watch?v=pgaEE27nsQw - are you familiar with?
17:04 robotustra my robot will be on wheels, because it will drive indoos only
17:04 deshipu SpeedEvil: yes, I saw that
17:04 SpeedEvil http://www.goatstream.com/research/papers/SA2013/
17:04 SpeedEvil Interesting
17:04 robotustra I will programm this hand http://i.imgur.com/c201ilj.png
17:04 deshipu SpeedEvil: it's great for generating stable gaits in consistent environment
17:06 deshipu SpeedEvil: proprioception is going to be one of the keys, I think
17:06 SpeedEvil The thing I found interesting about it was implementing spinal delays
17:07 SpeedEvil Also I guess you could use various specific environments, and cross-evolve the muscle placement and attachemnet without directly carrying across the gait control
17:08 deshipu animals like cocroaches actually walk mostly by reflex and the flexibility of thier legs and muscles
17:08 SpeedEvil yeah
17:08 deshipu and they walk super-efficiently
17:08 SpeedEvil humans do that to a large degree too
17:08 deshipu I love how those bipeds in the video have toes
17:08 SpeedEvil At least to the point that most of it is feed-forward
17:09 deshipu chikens will run with their head cut off
17:09 SpeedEvil The brain is fucking excellent in misleading you that it works in the 'now'
17:09 SpeedEvil not most of a second ago in most cases.
17:09 deshipu you know the chronostasis experiment?
17:09 SpeedEvil Nope
17:10 deshipu if you have a clock with a seconds hand
17:10 deshipu look away from it, and then quickly look at it
17:10 deshipu the seconds hand will appear to stay still for 2 seconds
17:10 SpeedEvil not seen that
17:11 deshipu and then move normally
17:11 SpeedEvil Eye saccades are _funky_
17:11 SpeedEvil Look in a mirror.
17:11 SpeedEvil Now look at one eye, then the other.
17:11 SpeedEvil Beautiful cinematic jump-cut
17:11 deshipu it's because you brain compensates for the time you didn't see anything due to a saccade, and then backfills the visual buffer with what you see now
17:12 SpeedEvil http://imgur.com/search?q=cucumber+cat - related
17:12 SpeedEvil push a cucumber silently behind a cat so it looks like it's sneaking up on it when it's not looking
17:13 deshipu cruel
17:13 deshipu that only shows that cats are actually aliens
17:13 deshipu from a planet of predator cucumbers
17:17 SpeedEvil The brain seems to store surprisingly little of the visual scene as a '3d awareness'
17:18 SpeedEvil I recommend 'the brain science podcast'
17:18 robotustra deshipu: where is your robot
17:20 deshipu robotustra: oh, I have a bunch, but they all use preprogrammed gaits
17:21 deshipu robotustra: nothing interesting
17:21 robotustra bipeds?
17:21 deshipu quadrupeds
17:21 robotustra ok
17:21 deshipu I am thinking about trying to reproduce a particular spring-based gait model
17:21 robotustra abd you plan to make optimisation of giats?
17:22 deshipu no
17:22 SpeedEvil https://whisperingdark.files.wordpress.com/2013/06/baba.jpg - I really want to do this for my shed
17:23 SpeedEvil It seems alas, dramatically impractical.
17:23 deshipu I don't think I can make my robots walk the hard way, I'm counting on finding a trick :)
17:23 SpeedEvil deshipu: hampster in an encounter suit.
17:24 robotustra baba yaga
17:24 deshipu robotustra: https://hackaday.io/project/3229-pengu-q1-quadruped <-- this is the project I'm thinking about mimicking, since the original author didn't have time for it
17:24 deshipu although this is direct control with no feedback
17:25 robotustra it's difficult to make good giat without feedback
17:25 SpeedEvil Interesting
17:25 deshipu I will settle for bad gait
17:26 SpeedEvil yeah - bad gait is fine if you can compromise on speed/power/...
17:26 deshipu since the robot is so small, it has abundance of power
17:26 deshipu well, it won't jump
17:27 SpeedEvil http://news.nationalgeographic.com/news/2013/09/130912-planthopper-gear-wheel-insect-legs-science/
17:29 deshipu I like how jumping spiders jump
17:29 deshipu they actually use hydraulics
17:30 SpeedEvil The synchronisation to avoid spinning madly is insane
17:30 deshipu spiders use thread for that
17:31 deshipu http://insects.about.com/od/spiders/f/How-Do-Jumping-Spiders-Jump.htm
17:37 Snert_ they do something that millions of dollars can't do
17:38 Snert_ so simple for them.
17:41 deshipu nevermind millions of years of evolution
17:41 deshipu you know how much that would have costed?
17:43 Snert_ I often wonder what enviornmental conditions prompted perfection such that an insect or plant decides to evolve one way or another.
17:43 Snert_ what made a venus flytrap learn to close it's jaws
17:44 deshipu well, I wouldn't call it perfection
17:45 deshipu considering the goal was to "survive", those organisms do pretty bad job in a very complicated way
17:45 deshipu single-cell organisms do it much better and more efficiently, and they still rule the planet
17:46 deshipu multi-cells are a dead end, really
17:46 deshipu freaks of nature
17:47 Snert_ so I'm a plant. Did I decide to try making my leaves all different colors before I found out that green made food for me?
17:47 deshipu you know they found some living single cells on satellites recovered from space?
17:48 deshipu that's survival skills
17:48 Snert_ it had to decide which color to try
17:48 deshipu Snert_: evolution doesn't "decide" anything
17:48 deshipu it just tries them all
17:49 deshipu well, not all colors, all changes to genes that may or may not result in a change of color
17:50 deshipu it's just a very parallelized optimization process
17:50 anonnumberanon man i just trash picked a flyback transformer that's 76$ new
17:50 deshipu nice find
17:51 anonnumberanon ill make electric arcs with it when i have time
17:51 anonnumberanon ive seen it done
17:51 deshipu bzzt
19:49 robotustra http://gif2.mycdn.me/image?t=45&bid=811929402257&id=811929402257&plc=WEB&tkn=*TiLjjr1UsilevOct7xPDS08Zb1o
19:50 robotustra chinese in the sea
19:51 Snert_ they got that thing where they all like to do that. It's weird.
19:51 Snert_ seadragon worship or something.
20:06 rantenki_lab rue_shop3, rue_more : Cutting 11 more.
20:50 rantenki_lab Also, that drill is amazing. Absolutely no measurable runout, not wobble at 30k rpm
21:57 rue_more rantenki_lab, armyofevilrobots!
21:57 rue_more rantenki_lab, should I modify the design for the rest? a bit longer?
21:58 rue_more I just got the messages about dave!
21:58 rue_more eek!
22:44 rue_more rantenki_lab, ?
22:44 rue_more armyofevilrobots,
22:44 rue_more ^^
22:55 armyofevilrobots @rue_more : Dave is fine. One of those scary things that turns out to be OK.
22:55 armyofevilrobots I just finished another 11. The drill worked perfect.
22:57 rue_more ok, what would make it better?
22:58 rue_more longer?
22:58 rue_more the size must have been right, 3/8
22:59 rue_more as the whole thing was meant to chuck up, I suspect it was about 5mm short
23:02 rue_more armyofevilrobots, ?
23:03 rue_more armyofevilrobots, have any favorite sizes? shall I just start with everything under 1/8?
23:58 anonnumberanon What's you guys' project?