#robotics Logs
Dec 13 2017
#robotics Calendar
12:24 AM rue_mohr: I have the transition down to 2ms
12:27 AM rue_mohr: kd is 1
12:27 AM rue_mohr: kp is 1/8
12:27 AM rue_mohr: its hard when the numbers are so small
12:27 AM rue_mohr: and integers
12:27 AM rue_mohr: going to 8 to 9 is a huge % diff
12:27 AM rue_mohr: :/
12:42 AM rue_mohr: oh that would introduce some problems, the pid loop was going at a few x the adc reading rate
12:57 AM rue_mohr: yea I'm confused
01:26 AM Tholia2 is now known as Tholia
09:32 AM anonnumberanon: yeah cause i found brand new tablets for $30 at Micro Center, with a 7" screen, those would be good to buy just for the sake of ripping out the screen from em
09:32 AM veverak: lol
09:32 AM veverak: :)
09:33 AM anonnumberanon: i just wonder if the screen driver is generic, or if it's that super-hard-to-reverse-engineer driver, or screen, from mysterious manufacturing plant
09:33 AM * veverak wants e-paper screen
09:33 AM veverak: realized it might be the best for debugging the robots
09:34 AM anonnumberanon: why the best?
09:34 AM veverak: because it can leave the reason it died on the screen and than simply die
09:34 AM anonnumberanon: oh lol
09:34 AM anonnumberanon: absolutely
09:35 AM veverak: seriously, that still seems like a nice feature to me
09:35 AM veverak: and given that it is wifi-controlled robot
09:35 AM veverak: I would like to write actual way of connection on the screen
09:36 AM veverak: like "AP with an essid: ..." || "Connected to wifi network: ... with an ip: ... "
09:37 AM anonnumberanon: yeah that would be nice
09:38 AM anonnumberanon: if it's running Linux you could pull that info from some commands
09:38 AM anonnumberanon: iwconfig or something
09:38 AM anonnumberanon: it recently changed with stable debian, i dunno what to use now
09:38 AM anonnumberanon: iw ?
09:39 AM anonnumberanon: "not associated", "associated with ..."
09:39 AM veverak: yeah, that is the easy part
09:39 AM veverak: :)
09:39 AM anonnumberanon: That's when the wifi is connected to a router that it says that.
09:39 AM anonnumberanon: "Would you like to shut off all wifi router in your area?"
09:40 AM anonnumberanon: >no i don't hack
09:40 AM anonnumberanon: >yes please, aireplay is fun af
09:40 AM veverak: :D
09:42 AM anonnumberanon: ..sending disconnect packet to: -37 APs-
09:42 AM anonnumberanon: Refresh set to -2 seconds-.
09:42 AM veverak: https://www.aliexpress.com/item/Raspberry-Pi-2-7-inch-LCD-e-Paper-E-Ink-2-7inch-Display-less-power-than/32827746105.html
09:42 AM anonnumberanon: You start seeing people come out of their house. Babies crying. Cars crashing around the corner.
09:42 AM anonnumberanon: Gun shots.
09:48 AM z64555: its like nobody uses ethernet cables anymore
09:48 AM z64555: >says the guy using a 4G wireless router
09:51 AM veverak: z64555: ethernet cable and mobile robotics does not do well together
09:51 AM veverak: :/
09:51 AM z64555: just hang it from the ceiling
09:52 AM z64555: or have a long spool attached to the back of the bot. :D
11:41 AM anonnumberanon: soon to be 5G
12:10 PM rue_mohr: veverak, but its easy to know when you get to the limit of the range, with wifi, you never know where signal will dissapear
12:11 PM veverak: treu to that
12:11 PM veverak: *true
12:12 PM veverak: rue_mohr: one of the reasons why I think that eink display will be nice
01:02 PM Tom_L: rue_mohr, scored big at the scrap today
01:03 PM Tom_L: found a NEW piece of 1/4" plate instead of that old pitted rusty one i had and some nice 2" tube to make the Y axis screw fit now
01:03 PM rue_mohr: :)
01:03 PM Tom_L: the tube isn't quite as thick as what i had but still considered thickwall
01:04 PM Tom_L: should do just fine
01:04 PM Tom_L: and it was in NEW shape too
01:04 PM rue_mohr: sweet
01:04 PM Tom_L: better grab some more cutoff wheels...
01:04 PM Tom_L: and this way both Y base rails will be from the same metal
01:04 PM Tom_L: instead of trying to mis match pieces
01:05 PM Tom_L: lotsa ppl there hunting today too
01:06 PM Tom_L: one guy's making a gokart
01:06 PM Tom_L: he was impressed with my cnc endeavor
01:25 PM Tom_L: i'll send you all the scrap i have left for yours if you pay the postage
01:25 PM rue_mohr: hah
01:26 PM Tom_L: me and the gokart guy wanted the same piece but i had a tape measure so he let me have it
01:26 PM Tom_L: he wasn't sure what the thickness was
01:27 PM Tom_L: it had to have arrived today or yesterday
01:27 PM Tom_L: it was too clean
01:28 PM anonnumberanon: man at recycling around me the policy is you can't take stuff (i do anyway or some workers let me)
01:28 PM anonnumberanon: that is so maddening
01:28 PM anonnumberanon: the reason is, imo that recycling centers get money from the state according to the weight of stuff they recycle
01:46 PM Snert_: ppl tend to dump all sorts of crap that isn't necessarily recyclable.
01:47 PM Snert_: and they dump it on the recycler who has no way of disposing of certain things.
01:47 PM rue_mohr: yea
01:47 PM rue_mohr: and they wont just take money either
01:47 PM rue_mohr: we all need to start interception companies
01:48 PM Snert_: ppl donate us omputers. There's a reason they don't want it. And for the same reason we don't want them either.
01:48 PM Snert_: We don't want old slow crap that nobody else wants.
02:01 PM AchiestDragon: interesting point ,, though is crt monitors ,, did you know the power saving you get from using a 19" lcd over a 19" crt monitor given 8 hrs a day usage , you could buy a new 19" lcd (well actualy more like a 24" lcd ) monitor each year from the amount of extra electric you'd be paying to run a crt
02:02 PM Snert_: the cost of juice keeps rising. So it's not like you ever actually pay less to run the LCD monitor.
02:03 PM Snert_: 'ceptin' ... depending on your locale....your source of juice.
02:03 PM Snert_: we have hydro power and the cost stays about the same so one makes out but never notices.
02:04 PM AchiestDragon: but it was one of the other issues that you should mention when people try to dump old computer kit on you with hay look at this its got a 21" crt display
02:04 PM Snert_: I don't ejukate the sheeple. I just say no.
02:04 PM Snert_: no time for that.
02:21 PM veverak: hmm
02:21 PM veverak: "To put it differently"
02:21 PM veverak: is there better phras in english?
02:21 PM veverak: I described something, and now I would like to describe it in different way
02:21 PM veverak: "To put it differently" may not be best way to start the second description
02:23 PM durrf: you want to put "to put it differently" differently?
02:25 PM veverak: hmm
02:26 PM veverak: so maybe it is correct english phrase? :)
02:26 PM veverak: Instead of that, the author prepared a graph-based solution. Vertex of the graph represents a state of the leg. Edge contains price required to change the state of the leg.
02:26 PM veverak: The task of getting the movement path for the leg is to find the cheapest path between vertices.
02:26 PM veverak: To put it differently, the movement path is found in the graph structure, not generated.
02:26 PM AchiestDragon: "alternatively "
02:26 PM rue_mohr: hmm naw
02:27 PM rue_mohr: veverak, the movement path is generated using an algorythm that finds the cheapest path between the verticies
02:28 PM rue_mohr: otherwise integrated
02:28 PM veverak: eh
02:29 PM rue_mohr: well your second sentence dosn't make sense
02:29 PM rue_mohr: your trying to say its realtime generated, right
02:29 PM veverak: rue_mohr: that paragraph describes second version of the algorithm, the first one generated the path based on sophisticated formulas.
02:29 PM rue_mohr: you have to fix the spelling issues in what I suggest
02:29 PM veverak: the last sentence refers to the "generating the path"
02:29 PM veverak: I mean, it makes sense in the context it is used
02:30 PM rue_mohr: for the last sentence then, how about
02:31 PM rue_mohr: Therefore, the path is found graphically, not
02:31 PM rue_mohr: see generated isn't a good word
02:31 PM rue_mohr: did any of those images work out?
02:31 PM veverak: got it in todo, will come back to it later
02:31 PM veverak: it is 21:31
02:32 PM rue_mohr: your generating them graphically, not ... algebraically?
02:32 PM veverak: 13:00 tomorrow is deadline
02:32 PM veverak: gotta prioritize
02:32 PM veverak: rue_mohr: that may be better terminology, thanks
02:32 PM rue_mohr: ok
02:33 PM anonnumberanon: veverak, "in other words,..."
02:33 PM anonnumberanon: maybe?
02:33 PM veverak: wow, yeah that sounds better
02:33 PM rue_mohr: so, you left all the work till the last moment and then got your international team to help you with it?
02:34 PM rue_mohr: :)
02:34 PM veverak: not all the work
02:34 PM * anonnumberanon chortles
02:34 PM veverak: I started two months ago
02:35 PM veverak: of course the hours spent on this approached 24h as the deadline is closer
02:35 PM veverak: found out I am not exactly good at writing
02:37 PM anonnumberanon: I corrected my cousin's PHD thesis because he had to do it in English and he is French.
02:38 PM anonnumberanon: 40 pages
02:39 PM anonnumberanon: He never was good at spelling or grammar so I had quite a bit of work.
02:39 PM veverak: 40 pages ?
02:39 PM veverak: like pdf pages?
02:39 PM anonnumberanon: microsoft word
02:40 PM anonnumberanon: with some margin layout, some diagrams
02:40 PM anonnumberanon: veverak gonna be PHD YO!!!
02:41 PM anonnumberanon: :)
02:42 PM veverak: not really
02:42 PM veverak: that is bachelors
02:43 PM anonnumberanon: oh lol
02:43 PM anonnumberanon: ok
02:43 PM anonnumberanon: kinda like the final project report we did with our quadcopter, it was ~40 pages but it was 5 of us
02:44 PM anonnumberanon: at the end of my bachelor
02:44 PM anonnumberanon: s
02:44 PM veverak: I've got pdf with 123 pages
02:44 PM veverak: which is are 60 normalized page
02:44 PM veverak: (one normalized page = 1800 characters )
02:46 PM anonnumberanon: they want 20 right?
02:46 PM veverak: 30
02:47 PM anonnumberanon: i kind of want this http://www.digibarn.com/collections/systems/zenith-laptop/DSC06154.JPG
02:48 PM veverak: ok
02:48 PM veverak: let's continue
02:48 PM veverak: I define "clearance" as inverted distance from the environment
02:49 PM veverak: and want to find a path with minimal clearance
02:49 PM veverak: "the path with lowest clearance"
02:49 PM anonnumberanon: range of motion?
02:50 PM anonnumberanon: lowest clearance is good imo
02:50 PM veverak: "best clearance" sounds better
02:51 PM anonnumberanon: and you mean the path for which the leg will soonest hit something in the environment?
02:52 PM veverak: nope, the oposite :)
02:52 PM veverak: clearance of a vertex in graph is: clearance = 1/distance
02:52 PM veverak: distance is the distance from the environment
02:52 PM veverak: the smaller, the higher the distance, the better
02:56 PM veverak: reciprocal!
03:05 PM anonnumberanon: yes so best clearance is right
03:05 PM anonnumberanon: or, "widest range of motion"
03:06 PM anonnumberanon: "the path which offers widest range of motion for this particular actuator"
03:15 PM veverak: meantime
03:15 PM veverak: my graphviz-fu is increasing
03:16 PM veverak: latex dot2tex package is awesome
03:16 PM veverak: I can write the graphviz files inside the tex <#
03:16 PM veverak: <3
03:20 PM anonnumberanon: nice
03:24 PM veverak: http://squirrel.veverak.org/Screenshot%20from%202017-12-13%2022-23-36.png
03:46 PM anonnumberanon: rue_mohr, look in the back of this picture
03:47 PM anonnumberanon: 43 minute ride, probably $5 in tolls
03:50 PM rue_mohr: dont want that
03:51 PM rue_mohr: one stepper motor and its part of the floppy drives frame
03:57 PM anonnumberanon: https://newyork.craigslist.org/mnh/zip/d/free-office-printer/6423089255.html
03:57 PM anonnumberanon: posted yesterday
04:28 PM rue_mohr: yea!
04:28 PM rue_mohr: two AWESOME stepper in that top unit for cnc
04:29 PM rue_mohr: the trays might have steppers but often their just dc motor
04:30 PM rue_mohr: while your there, ask for their old fax machine
04:30 PM anonnumberanon: but for this i need to rent a trailer for $25 :)
04:31 PM anonnumberanon: the canon one you're talking about?
04:38 PM anonnumberanon: wtf
04:38 PM anonnumberanon: where is the first printer i posted
04:38 PM anonnumberanon: thought i posted it
04:38 PM anonnumberanon: and then you answered about ti
04:38 PM anonnumberanon: about it
04:38 PM anonnumberanon: but you didn't see it
04:38 PM anonnumberanon: what' sgoing on
04:39 PM * anonnumberanon gives up
04:48 PM veverak: ok
04:48 PM veverak: now I need to draw three tables
04:48 PM veverak: and lines between them
04:50 PM anonnumberanon: rue_mohr, this is the one that i meant to post first https://jerseyshore.craigslist.org/zip/d/free-office-furniture-and/6416402774.html
04:50 PM anonnumberanon: that was 45 minutes away
05:01 PM rue_mohr: yea that got a top sheet feeder too
05:02 PM rue_mohr: ya dont need a trailer, you jsut need ascrewdriver, their usually modular
05:21 PM veverak: one last question
05:22 PM veverak: name of the section: "Pattern generator based algorithm"
05:22 PM veverak: (the reader is aware of what "pattern generator is"
05:23 PM rue_mohr: yea
05:24 PM Tom_L: more pics...
05:24 PM anonnumberanon: Pattern-generator-based algorithm
05:24 PM anonnumberanon: spelled like this ^
05:24 PM anonnumberanon: or maybe like this: patter generator-based algorithm
05:25 PM z64555: nah, not chinglish enough
05:25 PM rue_mohr: its probably relivent to say that 90% of the people reading it aren't smart enough to understand a darn bit of it, and that you could say whatever you want, but somehow hide in it that if someone wants the real details to contact you
05:26 PM rue_mohr: (maybe make it verticle letters that lineup on a page somewhere or soemthing, dumb people are really dumb and will miss ANYTHING STARRING THEM IN THE FACE)
05:28 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Mill_Steel/Welding/Base1.jpg
05:28 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Mill_Steel/Welding/Base2.jpg
05:28 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Mill_Steel/Welding/Base3.jpg
05:28 PM rue_mohr: yea
05:28 PM anonnumberanon: veverak, also this is the kind of construct that makes long sentences that usually kill my creativity during writing, i used to write like that then english classes tought me to do more like this: algorithm based on a patter generator
05:29 PM rue_mohr: those look like good tacs
05:29 PM veverak: yeah
05:29 PM veverak: much better, thanks
05:30 PM rue_mohr: Tom_L, still, make/weld end caps on those Z risers
05:30 PM rue_mohr: for lat.
05:31 PM veverak: anonnumberanon: I've got this feeling that I will get a good idea about how to write thesis
05:31 PM veverak: tomorrow at 13:01
05:32 PM veverak: deadline is 13:00
05:32 PM anonnumberanon: lol
05:32 PM anonnumberanon: wow how much do you still have to look over?
05:32 PM anonnumberanon: pages
05:33 PM SolG is now known as SolGriffin
05:33 PM veverak: I am at 94/125
05:33 PM veverak: (pdf pages)
05:33 PM veverak: I think that text ends in 10-15 pages
05:33 PM veverak: rest is biliography/glossary/acronyms
05:55 PM z64555: I need to crack open this silly car charger and see what aweful circuit they're using
05:55 PM z64555: it gets way too hot during use
05:56 PM anonnumberanon: not switch mode ? :S
05:56 PM z64555: for a $1, I suspect it might not even be a decent voltage regulator
06:35 PM Tom_L: rue_mohr, yeah i can do that
06:36 PM Tom_L: that was the plate i got today btw
06:41 PM Tom_L: rue_mohr, i think i'll add some 'legs' out the sides too for stability
06:42 PM Tom_L: that smaller tube
06:42 PM rue_mohr: the cad makes your design lots more complex than mine :)
06:42 PM Tom_L: i'll probably use those 4 small pieces to the left in this pic: http://tom-itx.no-ip.biz:81/~webpage/cnc/Mill_Steel/Welding/Base1.jpg
06:42 PM rue_mohr: but I like the way its folded
06:43 PM Tom_L: complex?
06:43 PM rue_mohr: lots of pieces
06:43 PM Tom_L: oh
06:43 PM rue_mohr: its hard to imagine it finished casue of the lack of table
06:43 PM rue_mohr: its gonna be huge
06:43 PM Tom_L: i'm rather proud of that plate
06:43 PM Tom_L: oh, i'm using the orignial base tubing instead of the one i got today afterall
06:44 PM Tom_L: it was gonna end up being to long
06:44 PM rue_mohr: everyt5hing looks good
06:44 PM Tom_L: i'll just have to cut the ballscrew
06:44 PM rue_mohr: you have the ... Y sliders yes?
06:44 PM Tom_L: yes
06:44 PM Tom_L: i just have to cut them
06:44 PM Tom_L: since they'll be the 1000mm rod i'm using
06:45 PM Tom_L: the X will be 700
06:45 PM Tom_L: it fits ok on it
06:45 PM Tom_L: so 400 x 500 x 700
06:45 PM Tom_L: i may go finish welding it
06:45 PM rue_mohr: .. that bench isn't deep enough is it
06:45 PM Tom_L: i had to run errands
06:46 PM Tom_L: you haven't seen my steel table
06:46 PM Tom_L: it's to the right of the wood one
06:46 PM rue_mohr: or were the errands running you...
06:46 PM rue_mohr: k
06:46 PM Tom_L: 1/2" plate
06:46 PM rue_mohr: so I'm guessing that thing already weighs 200lbs?
06:46 PM Tom_L: got that when i was a kid
06:46 PM Tom_L: from my neighbor at a garage sale
06:46 PM rue_mohr: :) good score
06:46 PM Tom_L: along with an old old buffalo drillpress i wore out
06:47 PM rue_mohr: buffalo?
06:47 PM Tom_L: it's not big but it's heavy
06:47 PM Tom_L: antique probably
06:47 PM Tom_L: he was a welder by trade
06:48 PM rue_mohr: how you metering twist on that bottom frame?
06:50 PM rue_mohr: I just came up with an idea ath involves, ground down screws and nichrome wire
06:50 PM Tom_L: i'm not
06:50 PM rue_mohr: k
06:50 PM Tom_L: tacking all sides
06:50 PM Tom_L: welding in a pattern to avoid as much as possible
06:50 PM rue_mohr: then you need to check tho
06:50 PM rue_mohr: the plate seems really flat?
06:50 PM Tom_L: i ran a straighedge over the top
06:50 PM Tom_L: it is
06:51 PM rue_mohr: cross wise I mean
06:51 PM Tom_L: i laid the ruler on it too
06:51 PM rue_mohr: twist
06:51 PM Tom_L: it's pretty flat
06:51 PM rue_mohr: ___
06:51 PM rue_mohr: __-
06:51 PM Tom_L: if it doesn't look it, it must be optics
06:51 PM rue_mohr: k
06:51 PM rue_mohr: I wouldn't be abel to tell
06:51 PM rue_mohr: it was a problem on the first machine I did with a welded frame
06:51 PM rue_mohr: I really thought I had it
06:52 PM rue_mohr: I didn't
06:52 PM Tom_L: i wouldn't run a laser beam across it but it's flat
06:52 PM Tom_L: i gotta be careful not to weld certain things though
06:53 PM rue_mohr: I remembered to look at that peice of channel I have today
06:53 PM Tom_L: i think i'm gonna flip it over and tack the bottom before i start
06:53 PM rue_mohr: yup, that end peice will snap off if given the chance
06:54 PM rue_mohr: ya see it rolled a bit on ya too eh?
06:54 PM rue_mohr: good to stick a bit of a spacer in ones like that so there is room to roll it away from the shrinkage
06:55 PM rue_mohr: so can you find some polyurethane rubber to fill the column with?
06:56 PM Tom_L: i can probably find something
06:57 PM rue_mohr: I dont know if there would be an advantage to antivibration in the base
06:57 PM Tom_L: i'm gonna tack the bottom then cut my square plates for the standoff and the new Z base to replace the rusty one
06:57 PM rue_mohr: iirc sand scored pretty good as antivibration too
06:57 PM Tom_L: so i can drill the mounting holes for Z before i weld it on
06:57 PM rue_mohr: yup
06:58 PM rue_mohr: pilot holes in while its mocked up?
06:58 PM Tom_L: probably
06:58 PM rue_mohr: I want to play too :(
06:58 PM Tom_L: i'll drill the plate on the press then align it on the base and drill those
06:58 PM Tom_L: i may drill it all on the press i'll see how it goes
06:58 PM rue_mohr: hey
06:59 PM rue_mohr: cant you bring the column forward a bit, and cantalever the z sliders the extra 1.5 lower to get more span betweent he bearings?
07:00 PM Tom_L: why would i need the column forward?
07:00 PM rue_mohr: so the sliders can hang down to the lower frame
07:00 PM rue_mohr: which means you can drop the bottom bearing blocks for the z carrige a bit and get more base to their grip
07:01 PM rue_mohr: talking even 1/16"
07:02 PM rue_mohr: do you know what I mean?
07:03 PM Tom_L: they don't go that far anyway
07:03 PM Tom_L: but they don't need to
07:03 PM rue_mohr: the more the better
07:03 PM Tom_L: the base sliders and the XY plate mount etc make it taller
07:04 PM rue_mohr: your bit play is the bearing play times, whats now, 1.5
07:04 PM Tom_L: the mill will alredy go below the table as it is
07:04 PM rue_mohr: if you can get a 9" base between the zxz bearings, its bearing play *1
07:05 PM rue_mohr: (the z bearings, claws slipped)
07:05 PM Tom_L: i guess i'm not following you
07:06 PM rue_mohr: k, so, lets say thre is 10 thou play on the bearing block in the direction trying to pull it directly off the track
07:06 PM rue_mohr: just for argument sake
07:06 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Mill_Steel/Z_Slide1.jpg
07:06 PM Tom_L: that's Z as it will be
07:06 PM rue_mohr: your block spacing is 6" and your out 9"
07:06 PM Tom_L: yeah
07:07 PM rue_mohr: lever math, says that the 10 thou will be 1.5x at the bit
07:07 PM rue_mohr: so 15 thou vert play at the bit
07:07 PM Tom_L: it'll be better than the sherline
07:07 PM rue_mohr: if you can space the blocks 9" its just 10 thou
07:07 PM Tom_L: it's pretty tight with all 4 slides in place
07:08 PM rue_mohr: ... maybe its moot now :)
07:08 PM rue_mohr: was just a thought
07:08 PM Tom_L: the plates are drilled and waiting
07:08 PM * rue_mohr nods
07:08 PM Tom_L: no design changes there
07:08 PM Tom_L: except i need to trim that one to size
07:09 PM Tom_L: that last pic, see how much space between the bottom of the rails and the base?
07:09 PM Tom_L: it'll still go below the table
07:09 PM Tom_L: and all that is Z travel
07:10 PM Tom_L: i considered maybe trying to find a place to normalize it after welding then let my bud mill it flat
07:10 PM rue_mohr: funny how that overhead adds up eh?
07:10 PM rue_mohr: bonfire
07:10 PM Tom_L: it's in the cad
07:10 PM Tom_L: it all needs to turn cherry red
07:10 PM Tom_L: i don't think i can do that
07:11 PM Tom_L: i may just let him mill it without
07:11 PM Tom_L: depends how it bends
07:11 PM Tom_L: the Z column is fine
07:12 PM Tom_L: you're gonna have to come over and help me move it when it's done though
07:12 PM Tom_L: i don't think i'll be able to lift it
07:12 PM rue_mohr: I have that problem with the pipe bender right now
07:12 PM rue_mohr: I cant move it
07:12 PM rue_mohr: not since the heavy motor met the heavy bender
07:13 PM Tom_L: i could take that tube i got today and make a stand for it
07:13 PM Tom_L: it would be kinda short
07:13 PM Tom_L: ok, enough wasted time...
07:14 PM Tom_L: may get a few more pics tonight..
07:15 PM rue_mohr: where is our build log page
07:15 PM rue_mohr: want my cheat
07:15 PM rue_mohr: ?
07:15 PM rue_mohr: http://ruemohr.org/~ircjunk/robots/arm10/slide.htm get a copy of that
07:16 PM rue_mohr: http://ruemohr.org/~ircjunk/robots/arm10/imglist.txt make a file called that, with a carrige-retured list of all the filenames of the iamges in it
07:16 PM rue_mohr: done.
07:19 PM rue_mohr: I shoudl fix that one day
07:19 PM rue_mohr: meh
07:20 PM Tom_L: maybe later
07:20 PM Tom_L: we'll see if it's a success or not first
07:20 PM rue_mohr: hah
07:20 PM Tom_L: out to the garage
07:20 PM Tom_L: later..
07:21 PM rue_mohr: I'm going to go look at my spare livingroom till I get too dizzy to keep standing
07:23 PM rue_mohr: I have 3 full size plotters
07:23 PM rue_mohr: could all those with 3 full size plotters please raise your hands
07:23 PM rue_mohr: right, ok
07:23 PM rue_mohr: see, I'm doubting I need 3 full size plotters
07:24 PM rue_mohr: its just a crazy thought
07:26 PM rue_mohr: "He saves what others throw away, claims that he'll be rich some day; ..."
07:26 PM rue_mohr: theBear, I need a fog-box for setting up lens arrays
07:28 PM orlock: fog-box?
07:44 PM rue_mohr: so ya can see the lights beams
07:45 PM orlock: ahhh, gotit
07:45 PM orlock: those should be reasonably easy to make now
07:45 PM orlock: seeing as handheld smoke machines and the chemicals for them are everywhere
07:46 PM orlock: we have like, 6 at home :-\
07:46 PM rue_mohr: heh
07:46 PM rue_mohr: how much heat to they put to the fluid?
07:46 PM rue_mohr: 80c?
07:46 PM rue_mohr: less?
07:46 PM orlock: depends
07:46 PM orlock: it's a whole thing
07:47 PM orlock: depends on the power supply, coil ohms, etc
07:48 PM orlock: 150 - 212?
07:48 PM orlock: Farenheight
07:56 PM anonnumberanon: is it cold outside?
07:57 PM orlock: there's no kind of atmosphere!
07:57 PM anonnumberanon: it says it should snow right about NOW! but there ain't no snow
07:57 PM orlock: I'm all anlong, more or less.
08:09 PM rue_mohr: how about pulse frequency modulated analog?
08:36 PM -!- #robotics mode set to +o by card.freenode.net
09:06 PM Tom_itx: ok
09:09 PM Tom_itx is now known as Tom_L
09:09 PM Tom_L: that's better
09:10 PM Tom_L: rue_mohr,
09:10 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Mill_Steel/Welding/Base_Column1.jpg
09:10 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Mill_Steel/Welding/Base_Column2.jpg
09:11 PM Tom_L: the base is pinned in location
09:12 PM veverak: fuuuu
09:12 PM veverak: bibliography si broken
09:12 PM veverak: fuck
09:12 PM Tom_L: did you procrastinate too long?
09:19 PM rue_mohr: nice
09:19 PM veverak: not exactly
09:19 PM veverak: it just won't break lines in some cases
09:24 PM rue_mohr: Tom_L, what does shielding gas cost out there?
09:25 PM Tom_L: i'm not sure, i haven't bought any for a long time
09:25 PM Tom_L: and i've still got a spare full tank
09:25 PM Tom_L: this one still has about 1k lb in it
09:25 PM rue_mohr: your not leasing the tasnks?
09:25 PM Tom_L: no, they're too small
09:25 PM Tom_L: that's why i use em
09:25 PM Tom_L: just trade
09:25 PM rue_mohr: huh, tanks are $100/year here
09:26 PM Tom_L: i know, bigger ones are here too
09:26 PM Tom_L: but i don't weld that much so it's a win win
09:26 PM rue_mohr: hmm, that might be the key
09:26 PM rue_mohr: I never has gas cause I ahve no tank
09:27 PM Tom_L: ask about those smaller ones
09:27 PM Tom_L: maybe 18" tall or so
09:27 PM rue_mohr: ok, is that grounds to do a brian powercycle?
09:27 PM Tom_L: a who?
09:27 PM rue_mohr: that statement
09:27 PM rue_mohr: it should ahve said I never to any tig/mig cause I have no tank
09:28 PM Tom_L: oh well the interweb translated for me :)
09:29 PM Tom_itx: https://www.google.com/search?q=shield+gas+bottle+sizes&ie=utf-8&oe=utf-8&client=firefox-b-1
09:29 PM Tom_L: mine are probably like the green one
09:30 PM Tom_L: only i don't have to pay for the tank
09:31 PM Tom_L: just the gas
09:31 PM Tom_L: i may have when i got the welder, i really can't remember but they just keep trading them
09:31 PM rue_mohr: yup
09:31 PM Tom_L: shower..
09:32 PM rue_mohr: and I have to go shopping
09:32 PM rue_mohr: I really dont want to
10:02 PM Tom_L: tonight?!?!?
10:14 PM rue_mohr: I dont know
10:14 PM Tom_L: i got quite a bit done tonight
10:15 PM Tom_L: apple doesn't let you access their devices like a drive do they?
10:15 PM rue_mohr: I got in and put on my seatbelt, wanter to wipe the window, so I stepped on the brake, that wasn't it so I utned the key, that wasn't it so I turned on the headlights, that wasn't it so I released the parking brake, thought for a minute about where the wiper control was before leaving
10:16 PM Tom_L: i can plug my galaxy into the usb and it's just a drivee
10:16 PM rue_mohr: only kinda with itunes
10:16 PM Tom_L: this thing pops up some stupid app
10:16 PM rue_mohr: correct
10:16 PM rue_mohr: itunes
10:16 PM Tom_L: what if i wanna put some stuff on it?
10:16 PM rue_mohr: itunes was a flop of a program, so they made it the hub of all their mobile junk and everything
10:16 PM Tom_L: just files etc
10:17 PM Tom_L: i believe it
10:17 PM rue_mohr: its an apple, you put it on apple cloud
10:17 PM Tom_L: it does work nice for the program i got it for
10:17 PM Tom_L: that's all i care about really
10:17 PM rue_mohr: I can tell its not for playing musci
10:17 PM rue_mohr: using it as a level?
10:17 PM Tom_L: but i was just gonna see if i could view files on it
10:17 PM Tom_L: no
10:19 PM rue_mohr: you know that 'files' are 'yesterday' ? eveything is about content now, not files
10:19 PM rue_mohr: if you think about it, the notion of files has been removed from evreything social media
10:19 PM Tom_L: i was gonna put some mp3 on it or wmf or whatever they are
10:20 PM rue_mohr: no an apple product wont let you do that
10:20 PM rue_mohr: drm
10:20 PM rue_mohr: if you want to listen to music you have to buy it in the itunes store and put it in your onlien colelction
10:20 PM rue_mohr: on your knees and say "yes big brother"
10:20 PM Tom_L: i always knew there was a reason i hated apple
10:21 PM Tom_L: it will serve it's purpose
10:21 PM rue_mohr: <its a level>
10:21 PM Tom_L: job related software
10:22 PM rue_mohr: dosnt' even make a good map anymore since they ditched google maps and started using their own erronious database thats full of ads
10:22 PM Tom_L: in fact, i shut everything off on it i could
10:22 PM anonnumberanon: lol
10:22 PM Tom_L: anonnumberanon see todays updates?
10:22 PM anonnumberanon: yeah
10:23 PM anonnumberanon: it's weird your z axis, before, I thought that was your bottom frame...
10:23 PM rue_mohr: anonnumberanon, did you get the steppers?
10:23 PM rue_mohr: he has it on the bench for mock assembly
10:24 PM rue_mohr: geez, I got a pack of cut fruit, I dont even know what some of this stuff is, its definitly plant...
10:24 PM anonnumberanon: No it would take me 2 hours to get them and ~$30 in tolls and fuel.
10:24 PM rue_mohr: anyone wanna see my overshoot?
10:24 PM rue_mohr: anonnumberanon, ugh, why ... why did you ask then!?
10:24 PM anonnumberanon: Tom_L, I'm just patiently waiting until you have to start aligning things and see how you'll deal with it.
10:24 PM Tom_L: it wouldn't make much sense to me but i know you're working on PID
10:25 PM rue_mohr: just pd so far
10:25 PM anonnumberanon: rue_mohr, because I'm traveling through there Friday
10:25 PM rue_mohr: ah
10:26 PM anonnumberanon: I received stepper drivers too that I can play with this weekend.
10:26 PM Tom_L: leadshine?
10:26 PM rue_mohr: ?
10:28 PM anonnumberanon: Not exactly, the ones that use the A4988 chip. Few bucks a piece kind of deal, to connect to avr .
10:28 PM Tom_L: so low current drivers
10:28 PM Tom_L: like 2.5A or so
10:28 PM Tom_L: tops
10:30 PM rue_mohr: can take a pretty big motor and he has to learn basics
10:30 PM anonnumberanon: yeah something like that
10:31 PM anonnumberanon: it's to drive my little nema 17 I got, just to experiment
10:31 PM anonnumberanon: I'll be using TB6600 for a real machine
10:32 PM rue_mohr: the a4988 are good for 3d printers, you need 4
10:32 PM anonnumberanon: what is the 4th one for? the hot nozzle?
10:33 PM rue_mohr: xyze
10:33 PM rue_mohr: 4 dimational machine
10:33 PM rue_mohr: srsly?
10:33 PM rue_mohr: sime.///
10:33 PM rue_mohr: dimentional
10:34 PM rue_mohr: you should just stick with 4d
10:34 PM anonnumberanon: Tom_L, btw i found a tool that's kinda nice for alignment of the spindle with the table (assuming the table is straight itself...
10:34 PM anonnumberanon: gotta find it
10:34 PM rue_mohr: yes, yes, I should have just gone with 4d
10:36 PM anonnumberanon: Tom_L, https://smithy.com/sites/default/files/iadmin/pictures/products/19-006-withoutLR.jpg
10:37 PM rue_mohr: http://ruemohr.org/~ircjunk/images/p1090543.jpg
10:38 PM anonnumberanon: you chuck this to your spindle, orient it at a 45 degree angle away from the axis of the spindle, make sure it makes contact with your board, then turn the spindle by hand and the measurement should be the same, that means your spindle is straight related to your board
10:38 PM rue_mohr: tom has a better one
10:38 PM anonnumberanon: oh yeah i remember who used it ill find one
10:38 PM rue_mohr: contact sensor
10:38 PM anonnumberanon: ill find it
10:38 PM anonnumberanon: oh yeah?
10:38 PM rue_mohr: touchoff
10:39 PM rue_mohr: its good to the arcing distance of 5V
10:40 PM Tom_L: anonnumberanon i have one of those. it's called a last word indicator
10:40 PM rue_mohr: 1/200 mm or so
10:43 PM Tom_L: anonnumberanon, http://tom-itx.no-ip.biz:81/~webpage/cnc/G0704%20Tramming%20Part%201.mp4
10:43 PM Tom_L: there's 2 others there too
10:44 PM Tom_L: part2 & part3
10:46 PM rue_mohr: proportionate: error = error/16; with minor overshoot
10:46 PM rue_mohr: PD
10:46 PM rue_mohr: error = error/8;
10:46 PM rue_mohr: error -= deriv/4;
10:46 PM anonnumberanon: Tom_L, that's your mill?
10:47 PM Tom_L: no
10:47 PM Tom_L: it shows how to set one up
10:47 PM anonnumberanon: k
10:47 PM Tom_L: but i already know how
10:47 PM rue_mohr: slightly underdamped
10:47 PM Tom_L: fairly much anyway
10:47 PM Tom_L: rue_mohr, undershoot a bit and let I fix it
10:48 PM rue_mohr: I usually makes everything worse
10:48 PM Tom_L: P gets you close, D gets you close faster, I gets you home
10:50 PM Tom_L: linuxcnc adds a few more terms to the mix too
10:50 PM rue_mohr: an example I saw uses an error buffer
10:50 PM anonnumberanon: oh i get it now
10:50 PM rue_mohr: so that ki kd are only based on the last 3 samples
10:51 PM anonnumberanon: i saw that tool is either 30 bucks or 150 bucks, what gives?
10:51 PM Tom_L: quality
10:51 PM rue_mohr: http://paste.debian.net/1000643/
10:51 PM rue_mohr: so far the loop is simple
10:52 PM anonnumberanon: quality is in the mind of the user though
10:52 PM Tom_L: quality is in the product
10:52 PM z64555: "last three samples" It doesn't engaged I, D, within a % or range to target?
10:52 PM z64555: *doesn't engage
10:53 PM anonnumberanon: there are no gains rue_mohr
10:53 PM anonnumberanon: nothing? no pid?
10:53 PM rue_mohr: thats PD
10:54 PM rue_mohr: http://ruemohr.org/~ircjunk/images/p1090544.jpg
10:54 PM rue_mohr: the /8 is Kp and the /4 is the Kd
10:54 PM z64555: ?
10:54 PM z64555: derivative is on the input signal, not the error?
10:54 PM rue_mohr: yup
10:55 PM rue_mohr: your trying to limit the systems rate of change
10:55 PM rue_mohr: not the rate your changing it as, so symmantically
10:55 PM rue_mohr: ...
10:55 PM z64555: limit the systems rate of change by, well, taking into accoutn what it's doing
10:56 PM z64555: derivative on just the input signal something from a feedforward configuration
10:56 PM rue_mohr: there are a number of impementations
10:56 PM rue_mohr: I'm playing, its just an rc network
10:57 PM rue_mohr: the open loop is much worse than the proportionate, and thats worse than the PD
10:57 PM rue_mohr: as you can see, I didn't throw in I yet
10:57 PM rue_mohr: but I'm haivn some interesting results, or lack there of
10:57 PM z64555: ok, so you're experimenting with different configurations
10:57 PM rue_mohr: yup
10:57 PM z64555: ok
10:57 PM z64555: carry on
10:57 PM rue_mohr: have a thicken I can throw in the coup?
10:57 PM rue_mohr: chicken
10:58 PM rue_mohr: I blew it
10:58 PM rue_mohr: you did
10:58 PM z64555: no, I already deep fried it
10:58 PM anonnumberanon: rue_mohr, would be neat if you could show the waveform for raw signal, and then PID corrected
10:58 PM rue_mohr: have you done pid? sounds like it
10:58 PM rue_mohr: well if I change the comments around like...
10:58 PM z64555: I know enough about PID to get into trouble. :D
10:58 PM anonnumberanon: even better but that might be too challenging for you, the raw signal before RC, after RC, then after RC PID corrected
11:00 PM rue_mohr: yea, its doing that again, the raw dosn't look much different again
11:00 PM rue_mohr: huh
11:00 PM anonnumberanon: could even use those pictures in some new guide if your method ends up being a serious hack, but I'm advancing myself here :)
11:02 PM z64555: There was a PI Difference type, which used a slightly different method for the D term. I can't remember the specifics, sadly
11:02 PM rue_mohr: http://ruemohr.org/~ircjunk/images/p1090545.jpg
11:02 PM rue_mohr: :/
11:03 PM rue_mohr: there is a difference, its just on the tail end
11:04 PM rue_mohr: if I dial it to Kp of 7 (1/7) I get undershoot, fast responce
11:05 PM anonnumberanon: hook up a pot, connect the pot to the gain, adjust the gain by hand while watching the oscilloscope
11:05 PM rue_mohr: http://ruemohr.org/~ircjunk/images/p1090546.jpg
11:07 PM rue_mohr: the challange is doing it all fixed point without saturating any of the values
11:08 PM anonnumberanon: this is where having a board with 5 sliding potentiometers like i made would get handy
11:08 PM rue_mohr: yesno, the fixed value division is much faster on the avr
11:08 PM anonnumberanon: use 3 for kp ki kd, quickly be able to find out the best setting by looking at the scope
11:08 PM z64555: so use push buttons instead
11:08 PM z64555: inc/dec
11:08 PM rue_mohr: same diff
11:09 PM z64555: lol
11:09 PM rue_mohr: fixed value math is faster
11:09 PM rue_mohr: and I didn't check my time margins so I dont know how close I am
11:09 PM z64555: faster than a live test?
11:09 PM anonnumberanon: lol
11:09 PM anonnumberanon: wtf is going on here
11:10 PM rue_mohr: were performing the impossable
11:10 PM rue_mohr: commitee PID
11:10 PM rue_mohr: tuning
11:11 PM rue_mohr: I'm thinking about that other alg
11:11 PM rue_mohr: its amazing how little REAL microcontroller pid code is out there, MOST of the code is the academic / impracticle stuff
11:12 PM rue_mohr: I'm losing my ability to type on this thing...
11:14 PM z64555: its like they're patented or unpublished or something
11:15 PM anonnumberanon: ill give a hint
11:15 PM anonnumberanon: find something that uses a PID, download code for that thing, look for PID part in that code
11:16 PM rue_mohr: hah
11:18 PM rue_mohr: ok, this test paid off
11:19 PM rue_mohr: I turned off the pwm to see what the raw step was, and its as fast as I'm getting anyhow
11:20 PM rue_mohr: I cant do better than the step response of the whole kitten-maru
11:22 PM rue_mohr: so the benifits I can get here, are
11:22 PM rue_mohr: I can get 10 bit dac values from an 8 bit pwm system
11:22 PM rue_mohr: I can have error compensation
11:29 PM rue_mohr: I wonder what happens if I just make a digital comparitor
11:32 PM rue_mohr: yup, tahts pretty noisy
11:43 PM anonnumberanon: bed