#robotics Logs

Nov 07 2017

12:08 AM rue_mohr: ok I went thru another few boxes of pcb's got rid of a bunch more
12:11 AM rue_mohr: I still think holesaw + pipe section = better
12:11 AM rue_mohr: jig to cut a notch out of the side
12:11 AM rue_mohr: you ever gonna finish that drink?
12:29 AM rue_mohr: I paid a wack of money for a replacement controller chip for my soldering station
12:29 AM rue_mohr: special AC temperature regulator
12:30 AM rue_mohr: so, today I found one on a pcb
12:30 AM rue_mohr: a heat regulator for a laser printer
12:30 AM rue_mohr: kinda makes sense
12:30 AM rue_mohr: NaRf
12:49 AM rue_mohr: https://www.youtube.com/watch?v=tXaimUkCVU0
12:49 AM rue_mohr: wtf
12:49 AM rue_mohr: why did the ink only stick the the unetched metal?
01:20 AM rue_mohr: can I photoetch with UV nailpolish as a resist?
02:22 AM z64555: er, try placing a drop of the stuff in etching acid and see if it'll disolve
02:22 AM * z64555 is unhelpful
02:23 AM z64555: how's annipoo these days? haven't seen her lately
02:52 AM * Jak_o_Shadows was wondering that recently
05:12 AM edmoore: so cnc peppers
05:12 AM edmoore: bedroom mill
05:12 AM edmoore: just get taig?
05:32 AM z64556 is now known as z64555
06:33 AM SpeedEvil: https://www.haascnc.com/machines/vertical-mills/mini-mills/models/minimill.html#gsc.tab=0 ?
06:41 AM Tom_L: edmoore just get a bridgeport
06:41 AM edmoore: in my spare room?
06:42 AM Tom_L: you don't need to sleep anyway
06:42 AM edmoore: i don't think it would get up the stairs
06:42 AM edmoore: it would probably just fall through the floor back downstairs
06:42 AM Tom_L: so start there. move everything else up
06:44 AM SpeedEvil: Mills are a thing I wish scaled better.
11:22 AM veverak: I did not cloated over my last design!
11:22 AM veverak: *gloated
11:22 AM veverak: https://www.youtube.com/watch?v=yyiMKBUQLkM
11:22 AM veverak: prototype for paddle wheeled boat
11:22 AM veverak: as you can see I have plenty of power
11:22 AM veverak: :)
01:40 PM i-make-robots: *gloat
01:45 PM weyland|yutani: is this a gloat boat?
02:03 PM veverak: not yet
03:00 PM Snert__ is now known as Snert_
07:33 PM z64555: so uh. Anybody have experiance dealing with nonlinear systems
07:34 PM z64555: one of the methods I'm aware of to working with them is to set limits on such equations where they are close enough to linear to allow approximation
07:35 PM rue_mohr: yea
07:35 PM rue_mohr: I had to make a lokup table
07:36 PM rue_mohr: for the airmuscle
07:36 PM rue_mohr: the error goes against a table to work out the pulse time
07:36 PM z64555: hm, not sure if I can make a lookup for this one
07:37 PM rue_mohr: er, wait
07:37 PM rue_mohr: I used 4 tables
07:37 PM rue_mohr: the error determines which table is used
07:37 PM z64555: ? each table working on a specific range of errors?
07:38 PM rue_mohr: iirc
07:38 PM z64555: hm
07:38 PM rue_mohr: the responce of a digital water valve when used with air it REALLY nonlinear
07:39 PM rue_mohr: for doing propotionate air control
07:39 PM z64555: who would've thought. :D
07:39 PM z64555: sadly, I'm doing something more involved. Trying to make a simulation that can handle a system of force fields, where each field may have a different function
07:40 PM z64555: spring, fluid friction, inertia, gravity, bouyancy, that sort of thing
07:43 PM z64555: they seem to devolve to something parametric, trying to find the position of a particle at time t when several of the forces depend on the position, velocity, acceration of the particle
07:46 PM rue_mohr: do it on a graphics card
07:47 PM rue_mohr: (gpu)
07:47 PM z64555: still need to know what function to use :P
07:49 PM z64555: I was wondering if doing a partial-differential approach would be valid. That is to say, iterate the function through some time frame keeping all but one acceleration/force source constant
07:49 PM z64555: Then lerp the results
08:08 PM SpeedEvil: z64555: this sounds a bit like a discussion in another channel, where someone is moaning about there being no closed form solutions to plaentary motion for >2 bodies
08:18 PM rue_mohr: z64555,
08:27 PM rue_mohr: damnit I keep getting distracted
08:29 PM rue_mohr: https://www.youtube.com/watch?v=z2xWgEdiTwc
08:29 PM anonnumberanon: z64555, would be nice if you posted the data
08:29 PM rue_mohr: watch the screen, table lookup
08:29 PM rue_mohr: oh its just a lookup table
08:29 PM rue_mohr: 10 entry
09:32 PM z64555: anonnumberanon: what data. lol
09:32 PM z64555: SpeedEvil: Yeah possibly. Although it does fall under satellite robotics :D
09:32 PM z64555: I'm not expecting a closed form, at any rate
09:37 PM anonnumberanon: z64555, your function curves
09:37 PM anonnumberanon: how the fuck does it look? lol
09:38 PM z64555: standard kinematics equation, but acceleration is due to sum of different forces
09:39 PM z64555: x(t) = 1/2 * a(t) * t^2 + v(t) * t + x(0)
09:39 PM z64555: where v(t) = a(t) * t + v(0)
09:40 PM z64555: and a(t) = sum(F) / m + a(0)
09:40 PM z64555: That last equation is not 100% accurate though, because kinematics is based on constant acceleration
09:41 PM z64555: rue_mohr: neat
10:33 PM anonnumberanon: uhm
10:33 PM anonnumberanon: what?
10:36 PM z64555: its a system of forces
10:48 PM anonnumberanon: alright
10:58 PM z64555: could use the ol difference solution
10:58 PM z64555: assume forces are costant between t_n-1 and t_n
10:59 PM z64555: don't like that tho. computationally expensive