#robotics Logs

Mar 16 2018

01:24 AM Jak_o_Ombroj is now known as Jak_o_Shadows
01:28 AM Jak_o_Shadows: ohhhh, a complex virtual joint
01:28 AM Jak_o_Shadows: Is how I think ofit
01:33 AM mrdata: tensegrity structures?
01:33 AM Jak_o_Shadows: oh, so it's like a virtual joint
01:33 AM Jak_o_Shadows: That'll have a somewhat complex feasible region just for that joint though right?
01:34 AM Jak_o_Shadows: ? I was replying to z64555's joint thoughts
01:44 AM mrdata: https://www.youtube.com/watch?v=_yCGe_kzo3g
01:51 AM mrdata: https://www.youtube.com/watch?v=0eC4A2PXM-U
02:02 AM rue_bed: so, maybe I should work on my editor
02:02 AM rue_bed: then maybe I can write code faster
02:03 AM rue_bed: and finish projects as fast as I come up with them
02:03 AM rue_bed: I think I'm 10 years behind me
02:03 AM rue_bed: arg
03:41 AM z64555: Jak_o_Shadows: I don't follow?
03:42 AM Jak_o_Shadows: Like, the spacethat the actuator can do isn't just an upper an lower limit
03:42 AM Jak_o_Shadows: it's a bunch more complicated?
03:43 AM z64555: Theres upper/lower limits, but for 3 axes
03:44 AM z64555: Not sure yet if this is like a "square" or "circle" limit, like what you'd see on some different joysticks
03:45 AM z64555: Rue was trying to get me to make the virtual joints, eh, less virtual
03:48 AM z64555: at which point I had replied that would involve simulating the actuators as a piston in a linkage system
03:49 AM * z64555 gets all the terms wrong
03:49 AM * z64555 is not a mechanical engineer
03:49 AM * z64555 sadface
03:50 AM Jak_o_Shadows: Do they specify this stuff in a mech eng degree? It seems kinda specific
03:50 AM Jak_o_Shadows: My gut is you're not going to have even a joystick type freedom
03:50 AM Jak_o_Shadows: Like, you'll be able to go a bit further in 3 directions than in between
03:51 AM z64555: There's a few graduate and high level undergraduate coarses concerned with industrial robotics, I'm sure they would cover something over this
03:51 AM Jak_o_Shadows: Not sure undergrad would
03:52 AM z64555: depends on the college, I guess
03:57 AM z64555: Erm, at any rate. I'm not focused on that just yet
03:58 AM z64555: Still trying to come to terms with how to put limits on euler vectors/quaternions without decomposing them into tait-bryan angles (pitch, yaw, roll)
04:00 AM Jak_o_Shadows: I think I call euler vectors axis angle?
04:01 AM Jak_o_Shadows: I used a 3-vector where the components are the axis to rotate around, and the magnitude is the amount to rotate
04:03 AM z64555: yep, that's a euler vector, also known as a rotation vector
04:03 AM * z64555 looks at wikipedia...
04:03 AM z64555: yeah.
04:04 AM z64555: tait-bryan angles are a variation common with aviation, so I've read. I've seen it used widely in flight simulators
04:04 AM Jak_o_Shadows: yeah
04:05 AM Jak_o_Shadows: through you might find that most aerospace engineers are dumbasses that don't know to properly specify how they're angles are
04:05 AM z64555: z+ = down. Lol
04:06 AM z64555: rotatations can really make people's heads spin. :D
04:06 AM z64555: *rotations
04:08 AM Jak_o_Shadows: yeah
04:09 AM Jak_o_Shadows: Worst thing is when people don't specify rotation order
04:09 AM Jak_o_Shadows: actually, a bit of software we paid to have written recently at work?
04:09 AM Jak_o_Shadows: They interpolated the euler angles in time
04:09 AM Jak_o_Shadows: Whicih DOEESN'T WORK AT ALL
04:11 AM z64555: hm
04:12 AM z64555: what they do wrong?
04:14 AM z64555: I'm guessing the euler angles were in the world frame, instead of the object's reference frame
04:15 AM z64555: which, yeah, would cause problems if you tried to interpolate that
04:30 AM Jak_o_Shadows: nah
04:30 AM Jak_o_Shadows: a) they didn't cover phase: 355 to 0 went through 180
04:30 AM Jak_o_Shadows: b) If you do the things indepdently, you don't move along the smooth arc between the two angles
04:30 AM Jak_o_Shadows: gotta be cleverer than that
04:35 AM veverak: hmm
04:36 AM Jak_o_Shadows: The best method is to convert to quaternions afaiik
04:36 AM veverak: it is
04:36 AM veverak: they have fluent linear interp
04:48 AM Jak_o_Shadows: Yeah.
04:49 AM Jak_o_Shadows: I'm quite annoyed really
07:18 AM z64555: hm, I should be writing notes on all of this
07:20 AM z64555: hm, I should get a binder
07:20 AM z64555: or a brad folder
07:32 AM Jak_o_Shadows: I recently got like a sleeve thing
07:32 AM Jak_o_Shadows: Not a folder - a portfolio thing with 8 sleeves or so. A bit thicker, and you can fit more than one bit of paper in
07:32 AM Jak_o_Shadows: ALso, yeah. I sometimes take it for granted how much background knowledge I have.
08:11 AM z64555: the notes are mainly for me to remember what I know :P
08:11 AM z64555: I don't have the joy of immediate recall
08:12 AM z64555: taking notes also helps form the memories into something more solid, and my note taking style tends to be something like a tutorial
08:13 AM z64555: Takes more effort, yeah, but I can throw it at somebody have have a non-zero confidence in they'll understand it :D
08:13 AM z64555: *and have a
08:33 AM z64556 is now known as z64555
12:00 PM Tom_itx is now known as Tom_L
05:09 PM -!- #robotics mode set to +o by niven.freenode.net
05:42 PM Jak_o_Shadows: Probably better than my spewing things at people
07:54 PM rue_: ironic sentence somehow
08:38 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Mill_Steel/Assembly/Final/SpeedHandle.jpg
08:38 PM Tom_L: for the vise
09:01 PM rue_bed: if I take a transformer, and do four diodes that compose two dc takeoffs, one set in each polarity, can I put them to two un-insolated dc lithium chargers that are on different cells in the same pack without them screwing with each other?
09:01 PM rue_bed: ... I think I can...
09:02 PM rue_bed: but if I take that transformer, and put an array of capacitor takeoffs before rectifiers, I can do as many as I want
09:03 PM rue_bed: so I can do a balanced charge of a large pack just fine on one transformer
09:03 PM rue_bed: and it can be square wave drive
09:03 PM rue_bed: hmm I think my brain is waking up
09:03 PM rue_bed: maybe I can ignore it
09:03 PM * rue_bed pulls the sheet back over his head
09:06 PM Tom_L: how's things?
09:06 PM rue_bed: I want to be asleep
09:06 PM rue_bed: and I want to go do things
09:06 PM Tom_L: no you don't
09:06 PM Tom_L: not really
09:07 PM rue_bed: so, i'v effectivly broken my operating point
09:07 PM rue_bed: I need to binge projects to completion
09:07 PM rue_bed: somehow
09:08 PM rue_bed: it means plastering thru anything resisting finishing a project immediatly
09:08 PM rue_bed: I need to work out how the brain works
09:08 PM rue_bed: I'm getting closer
09:09 PM rue_bed: but some critical info is missing
09:09 PM rue_bed: regarding general structure
09:11 PM Tom_L: is climb or conventional a better way to cut?
09:11 PM Tom_L: why can i never remember that...
09:11 PM rue_bed: climb or downmilling?
09:11 PM rue_bed: climb
09:12 PM Tom_L: which takes the most power?
09:12 PM rue_bed: iirc :)
09:12 PM rue_bed: there is a power difference?
09:12 PM Tom_L: yes
09:12 PM rue_bed: that I dont know
09:12 PM Tom_L: look at how they cut
09:12 PM Tom_L: one digs, the other scoops
09:12 PM rue_bed: same amount of sheer, I dunno
09:13 PM rue_bed: using my mental modeling
09:14 PM Tom_L: http://www.harveytool.com/secure/Content/Documents/Tech_ConventionalMillingVsClimbMilling.pdf
09:14 PM rue_bed: the surface stresses in the finish are different between the two
09:14 PM Tom_L: climb is better overall
09:15 PM rue_bed: and the torque is different
09:15 PM rue_bed: downmiling generates a spefific sheer plane
09:15 PM rue_bed: the sheer on the material is winder with climp
09:15 PM rue_bed: I think downmilling leaves more edge stresses in the material
09:16 PM rue_bed: in the climb milled material, the waste contains most of the sheer-stressed material
09:16 PM Tom_L: didn't you ask about making a broach once?
09:17 PM Tom_L: i need to make one..
09:17 PM rue_bed: maybe
09:17 PM rue_bed: I need square holes every so often
09:17 PM rue_bed: I think I'm generally waking up
09:18 PM Tom_L: you just needed a push
09:18 PM rue_bed: ... carefull its 5' down
09:18 PM rue_bed: thanks to thebear, its a bunk
09:18 PM rue_bed: stupid trailer with a low cieling
09:19 PM rue_bed: if I set up to go to sleep in my lazyboy, I could get up faster
09:21 PM Tom_L: i fixed my post to allow for fixture offsets couple days ago
09:22 PM Tom_L: works nice
09:22 PM Tom_L: and i cleaned up the rest of it while i was in there
09:22 PM rue_bed: nice
09:22 PM Tom_L: hopefully no extra code is output now
09:22 PM rue_bed: yes, a cnc, thats a good next main projet
09:22 PM rue_bed: or the mecha
09:22 PM rue_bed: arg
09:22 PM rue_bed: it dosn't matter thats what I'm learning
09:22 PM Tom_L: you can do more with a cnc
09:22 PM rue_bed: do it DO IT NOW
09:22 PM Tom_L: go that direction next
09:22 PM rue_bed: GET IT DONE FAST, NOTHING ELSE MATTERS
09:23 PM rue_bed: I'm like 10 years behind where I should be
09:23 PM rue_bed: which is, like, A MOUNTIAN
09:23 PM Tom_L: you're right where you are
09:23 PM rue_bed: I dont know if I can catch up
09:24 PM z64555: don't fret about it, lest you waste time doing so :P
09:24 PM rue_bed: I'm so far behind, I might come in first
09:24 PM rue_bed: "WTF YOU MEAN ANOTHER LAP!!!!!"
09:24 PM rue_bed: :/
09:25 PM rue_bed: I need everyone to stop bothering me for like a month
09:25 PM rue_bed: and to execute that month with like NO SLEEP
09:25 PM rue_bed: or eating... or going to the bathroom
09:25 PM rue_bed: no plan on how to do that
09:26 PM rue_bed: but I'm starting to understnad the design of the human brain a bit
09:26 PM rue_bed: not enough yet
09:26 PM rue_bed: oooooooh
09:26 PM rue_bed: its 7:30
09:26 PM rue_bed: need to think
09:26 PM rue_bed: I'll be lucky to get myself any time this weekend
09:26 PM rue_bed: everyone wants a peice of me this weekend
09:27 PM rue_bed: damnit, lost the remote for my bedroom lights again
09:27 PM rue_bed: should have put on a flipping override button
10:45 PM id4rk_ is now known as id4rk