#robotics Logs
May 29 2020
#robotics Calendar
04:26 AM goppo: anyone know if the housing of those UR arms is plastic or metal?
04:32 AM goppo: https://www.youtube.com/watch?v=I1MOxx0Q1D0
04:59 AM mun: looks like plastic
09:10 AM rue_mohr: heh, are people stripping those things down already?
09:10 AM rue_mohr: seems like they just came out yesterday
09:10 AM rue_mohr: what was it? 2005?
09:12 AM rue_mohr: oh I see
04:14 PM Samian: anyone know of any kits that'll let you control on/off of a circuit via software?
04:14 PM Samian: something that comes with a c++ sdk ?
04:19 PM polprog: the paralell or rs232 port in your computer
04:20 PM polprog: Samian: ^
04:20 PM deshipu: few computers have that anymore
04:21 PM deshipu: but you can use pretty much any microcontroller or programmer board
04:22 PM polprog: pretty borad question
04:22 PM polprog: broad*
04:23 PM Samian: cool thanks
04:23 PM Samian: I needed just some words to google, and you gave me that
04:24 PM deshipu: I suppose the key word is "gpio"
04:24 PM polprog: or discrete output
04:25 PM polprog: there are interface cards, which are costly but they are out of the box and probably come with some form of an SDK, if you dont feel like making something up from micros
04:25 PM polprog: work out of the box*
04:28 PM Samian: another question, I want to make a high speed micro linear actuator that's like this: https://www.youtube.com/watch?v=08zM0pswP14
04:29 PM Samian: a rod that moves 2 inches in/out
04:29 PM Samian: that's not super difficult is it?
04:29 PM Samian: I want to control the linear actuator via software. That's my end goal
04:30 PM Tom_L: still working on the french fry sorter/
04:30 PM Tom_L: ?
04:30 PM Samian: Tom_L yes
04:31 PM Tom_L: they use high speed cameras with air blasts to remove bad ones
04:31 PM Samian: Tom_L I know.. but we want to compete. We already have contracts with the same companies
04:31 PM Tom_L: physical actuator would get in the way
04:31 PM Tom_L: they move too fast
04:32 PM Samian: Tom_L so far we just do analysis from sampling off the line
04:32 PM Samian: https://youtu.be/O9fOvlizvrM?t=103
04:32 PM Samian: that's the software I wrote.
04:32 PM Tom_L: you must have missed the yt i posted the other day
04:32 PM Samian: using machine learning / neural networks
04:33 PM Tom_L: at that speed you will go broke
04:33 PM Samian: Tom_L I know how it's done now for knocking bad fries off the main line
04:34 PM Tom_L: if you plan to compete, this is a more realistic goal: https://www.youtube.com/watch?v=Xf7jaxwnyso
04:34 PM Samian: Tom_L oh the conveyor runs much faster now.
04:34 PM Samian: this is also for sampling off the line
04:34 PM Tom_L: skip to the part they pass over the air gap with the cameras
04:34 PM Samian: the fry companies don't have any other way to measure / count fries
04:35 PM Samian: Tom_L I've already seen it
04:35 PM Samian: Tom_L the fries fly through the air and air blasts at the bad fries knocking them into the crack below
04:36 PM Samian: Tom_L maybe I need to google for an array of straws that blasts air?
04:36 PM Tom_L: the blast of air is there and gone, with a linear actuator it would take 2x as long because it still has to retract to get out of the way for the next one
04:37 PM Samian: Tom_L right, the actuators would need to be extremely high speed
04:37 PM Tom_L: at those speeds they would just punch a hole in the fry
04:37 PM Tom_L: like a bullet
04:38 PM Tom_L: things to consider
04:38 PM Tom_L: there are reasons they do it the way they do
04:42 PM Samian: Tom_L an actuator had a rectangular plane attached to it, and the actuator was fast enough, it would push the entire fry down
04:42 PM Samian: it wouldn't have the bullet effect anymore
04:42 PM Samian: but more like the hit-by-a-bus effect
04:43 PM Samian: you disagree?
04:43 PM Tom_L: test it. i stand by what i said
04:43 PM Samian: cool
04:43 PM Tom_L: for that to work you would need to know the orientation of every fry every time
04:44 PM Samian: Tom_L sorting would be another project. The idea was to use actuators to act like gates that push fries to a different part of the conveyor, not to actually hit the fries.
04:44 PM Samian: see this? https://www.dropbox.com/s/fslxp3dcmdtuer2/WhatsApp%20Video%202019-11-09%20at%2011.35.25%20PM.mp4?dl=0
04:45 PM Samian: some of the fries overlap. I wrote software that can count the fries anyway, but it's better to get rid of overlapping fries all together
04:46 PM Samian: the idea I have is using actuators that shoot down so the fry on top hits the actuator and slides off the fry under, so no fries overlap