#robotics Logs
Aug 11 2017
#robotics Calendar
12:00 AM rue_more: ?
12:03 AM rue_more: you cant turn it on and off at the driver?
12:38 AM MrCurious: anyone watched this series before? https://www.youtube.com/watch?v=U6vr3iNrwRA&list=PLgnQpQtFTOGQrZ4O5QzbIHgl3b1JHimN_
12:59 AM rue_more: nope
01:08 AM Anniepoo_: nope
01:56 AM Jak_o_Shadows: rue, jitters all over the place. annieonprinter anniepoo, blender game engine rigid body stuff just jitters like made, couldn't fix it
01:56 AM Jak_o_Shadows: I'd REALLY like to use it though
02:05 AM rue_more: are you sure of the units
02:05 AM rue_more: your not feeding in degrees when its expecting rads are ya?
02:05 AM rue_more: that causes jitter
02:05 AM rue_more: :)
02:48 AM Jak_o_Shadows: aah
02:48 AM Jak_o_Shadows: aha
02:48 AM Jak_o_Shadows: Nah, I'm not trying to drive it (yet)
02:48 AM Jak_o_Shadows: just boring old gravity
06:12 AM Jak_o_Shadows: Looking at gazebo now
06:12 AM Jak_o_Shadows: Flipping no windows binaries
09:07 AM rue_more: oooh
10:55 AM Miyu is now known as hackkitten
10:56 AM anniepoo: Jak - is there an adjustment for simulation speed? If so, turn it up (more simulating per frame)
10:56 AM anniepoo: if not, then it's kind of ugly, but you can compensate by turning gravity down, etc.
10:57 AM anniepoo: but bleah - ok, maybe need to be using ROS after all
10:57 AM anniepoo: sigh....
10:57 AM anniepoo: though a PR for 'high precision physics' might make sense, it's just a number of cycles per frame
01:08 PM synja: <Jak_o_Shadows> Flipping no windows binaries
01:08 PM synja: ahahaha
01:08 PM synja: serves you right
06:34 PM rue_more: zhanx, how can I make a heated chocolate conveyance pipe?
06:34 PM zhanx: like chocolate serving device?
06:35 PM rue_more: NO! for chocolate printing
06:35 PM zhanx: ah
06:35 PM rue_more: copper pipe with a lot of current thru it?
06:35 PM rue_more: but I want something flexible
06:35 PM zhanx: induction would do it
06:36 PM zhanx: mmm
06:36 PM SpeedEvil: the right plastic works just fine
06:36 PM SpeedEvil: or silicone
06:36 PM SpeedEvil: water-jacket is probably the best way
06:36 PM rue_more: maybe if I wrap a plastic pipe in .... ni.... a wire and just run current thru it
06:36 PM zhanx: toaster wire around a tube
06:36 PM rue_more: ooooh coaxial
06:36 PM zhanx: speed that is a good idea
06:37 PM rue_more: I LOVE a good opportunity to make soemthign coaxial
06:37 PM zhanx: warm up time for water will suck
06:37 PM zhanx: but it will be flexible
06:37 PM rue_more: low volume
06:38 PM rue_more: its not like its really high temp either
06:38 PM rue_more: 40-50c
06:38 PM rue_more: +- the temper on it
06:39 PM Purplex: Hello
06:39 PM SpeedEvil: overheating is bad alas.
06:39 PM rue_more: yea
06:39 PM SpeedEvil: Also, you can not heat it, and simply use high pressure
06:40 PM rue_more: tho good chocolate will take a lot of punishment
06:43 PM Purplex: I want an army of robots
06:44 PM Purplex: I want to build an army of robots and conquer all the world
06:44 PM anniepoo: rue, they make flexible heater elements to wrap around pipes in cold climates
06:44 PM Purplex: Enslave you all and be first Emperror of Robotic Republic
06:45 PM anniepoo: put one on a dimmer
06:47 PM anniepoo: I have a small rubber snail
07:03 PM Snert_: FERR, of course :)
07:04 PM rue_bed: rue_shop2, philips size to drive pot
07:09 PM synja: usually is
09:13 PM rue_bed: damn fan needs a regrease
09:26 PM Tom_itx: strange that it didn't blow the 1A fuse but it blew the trace across the room in a ball of fire
09:28 PM rue_shop3: and they were in series?
09:34 PM Tom_itx: yes
10:27 PM rue_shop3: zhanx, almost there, printer problems keeping me from checking
10:29 PM zhanx: k
10:40 PM rue_shop3: ITS GOOD!
10:41 PM rue_shop3: zhanx,
10:41 PM rue_shop3: ok
10:41 PM rue_shop3: make these from a moderatly hard wood,
10:41 PM zhanx: really?
10:41 PM rue_shop3: holes are for 5/16" threaded rod (use 5/16" if you like)
10:41 PM rue_shop3: its 1/2" thick
10:42 PM zhanx: i got 5/16 unthreaded
10:42 PM rue_shop3: no worries about that yet
10:42 PM zhanx: email it and i will build it
10:42 PM zhanx: the extruder is my issue
10:42 PM rue_shop3: so 1/2" thick, 5/16" holes, the dotted line is the sideways hole
10:42 PM zhanx: k
10:42 PM rue_shop3: I'll just throw it on the webserver
10:43 PM zhanx: k
10:44 PM rue_shop3: http://ruemohr.org/%7Eircjunk/projects/reprap2/i2foot.dxf
10:45 PM rue_shop3: its 1:1
10:45 PM rue_shop3: tell me if it dosn't open right
10:45 PM rue_shop3: you need to make 4 of them
10:45 PM zhanx: didn't open right. fixed it
10:45 PM rue_shop3: ok...
10:46 PM rue_shop3: the poly hole should be 5/16"
10:46 PM rue_shop3: the peg is .6" wide
10:46 PM zhanx: it is
10:46 PM rue_shop3: did it screw up mm/inches?
10:46 PM zhanx: no clue i switch to mm years ago remember
10:47 PM rue_shop3: yea, I do everything metric
10:47 PM rue_shop3: I like to know the quirks in the output of the cad program
10:47 PM rue_shop3: I know most of its operational quirks
10:47 PM Tom_itx: ok the motor is good which i really didn't doubt but it's part of the testing
10:47 PM Tom_itx: the isolation board fuse wasn't blown so it's probably ok
10:47 PM zhanx: oh the holes are not "round"
10:47 PM rue_shop3: no
10:48 PM rue_shop3: they were converted, they are poly's
10:48 PM rue_shop3: sorry
10:48 PM Tom_itx: did a 12v test on the 90v motor and it turns
10:48 PM rue_shop3: its supposed to be as easy as projecting a slice of the stl into a dxf, but the feature dosn't work
10:48 PM Tom_itx: so the only thing left to test is the driver board
10:48 PM rue_shop3: Tom_itx, isn't that a fancy driver board?
10:49 PM Tom_itx: no it's the stock sherline one
10:49 PM rue_shop3: !? rlly?
10:49 PM zhanx: ruw these / this is the errors http://imgur.com/a/65ueL
10:49 PM Tom_itx: but to use it requires an isolation board because gnd floats
10:49 PM zhanx: i can print it 4x on a paper and make them
10:49 PM Tom_itx: http://tom-itx.no-ip.biz:81/~webpage/cnc/PDF/
10:50 PM Tom_itx: KBIC and KBIS
10:50 PM rue_shop3: zhanx, if you make a middle and do a 5/16" dia hole your on
10:50 PM zhanx: yep
10:50 PM zhanx: just a convert errpr
10:50 PM rue_shop3: ok, say when finished, I'll chew on the next file
10:51 PM zhanx: chew on it now
10:51 PM Tom_itx: i'm trying to figure out what the ssr actually switched
10:51 PM Tom_itx: i wonder if one of those cheapo ardweenie relays stuck
10:52 PM zhanx: tom they do
10:52 PM Tom_itx: it shouldn't have because they switch at zero current
10:52 PM zhanx: from my experience
10:52 PM Tom_itx: during the driver inhibit
10:52 PM Tom_itx: maybe that's what went wrong all along then
10:52 PM Tom_itx: i'll test it tomorrow
10:52 PM zhanx: tom i got spares for spares on them fyi
10:53 PM Tom_itx: i don't think i have any extra
10:53 PM zhanx: i thought i posted something about my test on them, 1-2 weeks max
10:53 PM Tom_itx: i'll ohm them tomorrow so see if they switch
10:53 PM Tom_itx: to*
10:53 PM zhanx: granted i was doing on / off every half second
10:54 PM Tom_itx: i seldom read logs anymore
10:54 PM Tom_itx: just pick up where i sit down
10:54 PM Tom_itx: i purposely wired it the way i did to prevent contact stick
10:54 PM zhanx: i was pushing them to test them and and then math ordered 50 more at the price point it was cheaper
10:55 PM zhanx: and the
10:55 PM Tom_itx: send me a couple :D
10:55 PM Tom_itx: wtf are you gonna do with 50?
10:55 PM zhanx: replace them as they fail in the next 2 years
10:56 PM zhanx: tom sounds like my wife
10:56 PM Tom_itx: that would cause this problem if one relay stuck and the other didn't
10:56 PM Tom_itx: i should just get a DPDT instead of 2 SPDT like i have now
10:56 PM zhanx: yes you should
10:56 PM Tom_itx: may lessen the chance of this happening
10:57 PM Tom_itx: problem is finding one that will switch under FPGA current
10:57 PM Tom_itx: ~3v
10:57 PM zhanx: yea that is an issue
10:57 PM rue_shop3: 2n7000
10:57 PM zhanx: all mine are 5v
10:58 PM Tom_itx: if i had a spare IO i wouldn't need to use the extra 50pin header from the fpga
10:58 PM Tom_itx: but they're all used up on the first header
10:58 PM zhanx: multiplex it?
10:58 PM Tom_itx: i could double check to make sure but i'm almost positive they're all used
10:59 PM Tom_itx: i have plenty of io just not any on that header that have protection or 5v
10:59 PM zhanx: toss in a avr they are cheap now
10:59 PM Tom_itx: i have boards i could add to it but i didn't wanna clutter the box
10:59 PM Tom_itx: as it's nearly full as it is
10:59 PM zhanx: crap
10:59 PM Tom_itx: and to do that i have to reverse all the logic
11:00 PM zhanx: i forgot to fix this tv today
11:00 PM Tom_itx: i can't remember what all is on that header
11:00 PM Tom_itx: just a few io used
11:00 PM rue_shop3: zhanx, you needed that $20
11:00 PM zhanx: try 140 rue and i made more today than that
11:01 PM Tom_itx: so maybe nothing is bad but a stuck relay
11:01 PM rue_shop3: yay!
11:01 PM zhanx: tom test it, you need to
11:01 PM Tom_itx: zhanx, do you know off hand how much current they require to switch?
11:02 PM Tom_itx: i've been using them right off the FPGA pin
11:02 PM zhanx: current or 4.5 volts dc at 1.1 amps?
11:02 PM zhanx: for the 5v ones
11:02 PM Tom_itx: no, the coil
11:02 PM zhanx: oh no i didn't test that
11:03 PM zhanx: has to be close to 5v that is for sure on what i did test
11:15 PM rue_shop3: k, the piece after looks right, test later
11:17 PM zhanx: rue_shop3: idea on the final design?
11:18 PM rue_shop3: final design?
11:18 PM rue_shop3: its already designed, I'm just converting the peices to non-printed stuff for ya
11:19 PM Tom_itx: should i look for latching or non latching relay?
11:19 PM rue_shop3: good luck finding a latching one
11:19 PM Tom_itx: does that mean a spring holds it once it switches?
11:19 PM rue_shop3: two types
11:19 PM Tom_itx: mmm those are dual coil
11:19 PM rue_shop3: two coil, one for on, the other off
11:19 PM Tom_itx: i don't want that
11:19 PM rue_shop3: and a toggle coil
11:20 PM rue_shop3: I'd avoid them as they aren't standard equip
11:22 PM Tom_itx: i don't think i'll find one that the FPGA could drive without a driver transistor or something
11:23 PM Tom_itx: i think i'll try to find a spare IO off the daughter card
11:23 PM Tom_itx: those have alot stronger drive current
11:23 PM Tom_itx: and are 5v iirc
11:23 PM rue_shop3: uln2003
11:23 PM rue_shop3: 2n7000
11:23 PM rue_shop3: !?!?!?!?!?!
11:23 PM Tom_itx: yeah i saw that earlier
11:24 PM Tom_itx: or just do away with the reversing spindle
11:24 PM Tom_itx: i just wanted to see if i could do that
11:24 PM Tom_itx: practically it's not strong enough to tap much of anything
11:25 PM Tom_itx: that might solve alot of headaches
11:26 PM Tom_itx: i can still do the orient code just have it orient in one direction instead of the shortest path
11:28 PM zhanx: wait what?
11:28 PM zhanx: Tom that math is done and free
11:29 PM Tom_itx: what do you mean by that?
11:29 PM Tom_itx: i'm working on orient code
11:29 PM Tom_itx: it has more than one mode
11:29 PM Tom_itx: one mode it to rotate to the set angle using the shortes path
11:30 PM Tom_itx: or just do FWD only etc
11:30 PM Tom_itx: if it were a production machine i'd go for the shortest path
11:30 PM Tom_itx: and make it work
11:30 PM Tom_itx: this is just step 1 of a bigger picture
11:31 PM Tom_itx: getting tool change carousel code to work after this
11:31 PM zhanx: oh
11:31 PM Tom_itx: but the spindle has to be in the right orientation first
11:31 PM zhanx: ir sensors then\
11:31 PM Tom_itx: i'll never use it
11:31 PM Tom_itx: i'm just testing
11:31 PM Tom_itx: trying to help someone else that will never figure it out
11:32 PM zhanx: ok
11:32 PM zhanx: how is this
11:32 PM Tom_itx: they've even got a carousel comp to use for it
11:32 PM zhanx: a dimple on the tool. slowly spin the tool until it is seen
11:33 PM zhanx: then drop get new tool
11:33 PM Tom_itx: http://linuxcnc.org/docs/2.7/html/man/man9/orient.9.html
11:33 PM Tom_itx: http://linuxcnc.org/docs/2.7/html/man/man9/carousel.9.html
11:33 PM Tom_itx: don't need to, it's got a quad encoder with index on it
11:33 PM Tom_itx: it's just a matter of getting the code written right
11:34 PM zhanx: 3-4 weeks of fun
11:34 PM zhanx: atleast
11:34 PM Tom_itx: the carousel comes in and grabs the tool, the spindle goes up a bit to clear it, then the carousel turns and returns to the spindle, then the spindle comes back down to receive the tool
11:35 PM Tom_itx: then the carousel moves back out of the way
11:35 PM Tom_itx: there's already code for it for the most part
11:35 PM Jak_o_Shadows: anniepoo, yeah, there's a physics subsetps number. But increasing it doesn't seem to do much
11:36 PM Jak_o_Shadows: The physics timestep is locked to the frame rate as 1/fps
11:41 PM Jak_o_Shadows: Huh. I applied scale (CTRL+A) and it seems to be working now
11:44 PM Jak_o_Shadows: well, working better