#robotics Logs
Nov 09 2014
#robotics Calendar
00:16 Loshki Walkerdine: you have a job and go to school. I expect you're busy.
00:18 Walkerdine Oh yeah, I work around 30 hours a week while taking 6 classes
00:19 Walkerdine but its still only 13 credit hours
00:22 Loshki Walkerdine: still, it's the kind of thing I like to see on resumes. It usually means good time-management skills.
00:23 Walkerdine I'm also on the Formula One Hybrid team
00:24 Loshki Walkerdine: Also, hustle and "fire in the belly". Can't teach those, people either have it or they don't.
00:25 Walkerdine Oh I definitely feel it. I would have to say I'm almost too impatient with trying to get where I wanna go
00:26 Walkerdine I've had an issue because of my broad range of interests
00:26 Walkerdine I can't seem to fall in love with anything specific
00:33 anonnumberanon If anyone know how I can acquire the computer sounds in this video clip I would be very grateful http://a.pomf.se/elcxxh.mp4
00:36 Loshki Walkerdine: how old are you? Is this a BSc or an MSc? And who's paying :-)
00:55 Hyratel Walkerdine, it's tough being a generalist
00:56 Hyratel most places want specialists, even though they need generalists
00:56 Walkerdine Im going for bs right now in compe
00:56 Walkerdine and im 20
00:56 Walkerdine gotta lotta time
00:59 Hyratel you're far ahead of me, then. all I have is 4 years of FRC as a Member, then 5 as a Mentor
01:05 vbk_ce join #math
01:05 vbk_ce whops
01:05 vbk_ce xD
01:24 vbk_ce Hello every one!
01:24 vbk_ce Is there a control theory channel?
01:24 vbk_ce I've google it but could not find it...
01:29 Hyratel why not jsut ask here?
01:30 vbk_ce Hyratel, ok. Thanks in advance.
01:31 vbk_ce I would like to land a lander... the control variables are the angle of the lander and the thrust power
01:32 Hyratel you need to get the vertical and horizontal motion components
01:32 Hyratel is this a virtual vehicle?
01:32 vbk_ce yes.
01:32 Loshki Walkerdine: With a bit of luck, your working life may span 40 years or more. You can't possibly anticipate the things you'll see. Concentrate on basic skills, problem solving, how to research, and as Hyratel says, read, read, read.
01:33 Hyratel how virtual (abstract) is your sensor suite?
01:33 Hyratel Walkerdine, ^ yeah.
01:33 Hyratel play with everything that crosses your desk
01:33 Hyratel take widgets apart
01:33 Hyratel try and figure out how you would have approached a task differently
01:34 vbk_ce well, it gives me the vertical and horizontal speeds each second and
01:34 Hyratel figure out where a design went wrong (or can go wrong) and braisntorm how to avoid or fix it
01:34 Hyratel each second? son, how fast is it moving?
01:34 Hyratel because 1 Hz is a dredfully low sample rate
01:35 Hyratel and half a second can be the difference between a pancake and a safe landing
01:35 vbk_ce whell ... there is a limit on the speed 40m/s h and 20m/s v
01:36 vbk_ce it is a land simulation
01:37 Hyratel are you "hopping" or dropped in from above?
01:37 Hyratel also what kind of altitude does it start from
01:38 Hyratel Walkerdine, beep
01:39 vbk_ce it starts at 2500m, and it is supose to land at about 100m
01:40 Hyratel 20m/s Vertical is .... not much
01:40 Hyratel how much thrust do you have?
01:40 Hyratel m/s^2
01:40 vbk_ce max 4m/s2
01:40 Hyratel dV limited?
01:40 Walkerdine I need to read more books yes
01:40 ratel is a Spess
01:41 Walkerdine It seems I'm missing out on alot of resources by trying to find everything on the internet
01:41 Hyratel Walkerdine, oh, it's all there online
01:41 Hyratel you just need to know what to look for
01:41 Hyratel focus on one thing (at a time)
01:41 Walkerdine Right now my focus is to figure out coding more
01:42 Hyratel beeeeeep vague
01:42 Hyratel find a Task
01:42 Hyratel distill it to its simplest component Tasks
01:42 vbk_ce whell, I found some good articles about coding at quora...
01:43 vbk_ce (sorry to bother...)
01:44 vbk_ce dV is limited to -+1
01:44 Hyratel what
01:44 Walkerdine Okay I have an end goal
01:44 Hyratel deltaVelocity
01:45 Walkerdine Build a "Robot" I guess, that plays Jenga
01:45 Hyratel which is defined by the fuel mass you carry and the Isp factor of the engine
01:48 vbk_ce no, no... my bad Only thrust and angle variations are restricted
01:49 Hyratel thrust force and attitude rot rate
01:50 Hyratel acceleration
01:50 Hyratel IRL, vehicle mass varies dramatically as fuel is burned, so the same thrust force can give a very different acceleration
01:51 Hyratel is the vehicle mass fixed?
01:52 vbk_ce this is something more complex. I'm utopicaly desconsidering mass changes ....
01:52 Hyratel simplifies your calcs though
01:52 vbk_ce just as the fuel mass are very small compared to the mass of the lander
01:53 vbk_ce just like *
01:53 Hyratel yeah - the amount of fuel you're burning on final approach is small compared to progressive braking manuvers
01:54 vbk_ce so, any guidelines?
01:54 Hyratel and the thrust-> acceleration calcs should already be done in the computer anyway
01:54 Hyratel for Apollo, for example, they had ti set so the "throttle" would, for landing, affect vertical speed
01:54 Hyratel and not actual thrust
01:55 vbk_ce nice, didnt know that.
01:55 Hyratel actual thrust was computed for throttle setting, VertVel, and fuel mass
01:56 Hyratel took a lot of load off the pilot without taking away any control
01:57 Hyratel are you working in a preexisting program?
01:57 Hyratel ...
01:57 Hyratel or are you using LunarLander and writing a landing autopilot?
01:58 Hyratel do you play KSP?
01:59 vbk_ce no.... I"trying to solve a problem at codingame actulay
02:00 vbk_ce actualy
02:00 anonnumberanon Walkerdine, what have you coded in C?
02:00 Hyratel condingame?
02:00 Hyratel codingame?
02:00 Hyratel is the operational envelope preset?
02:00 vbk_ce yes. xD
02:01 Hyratel why is the Hvel threshold set so high?
02:01 Hyratel also, what kind of terrain is there? just a flat surface?
02:01 Hyratel do you have a destruction speed threshold?
02:01 vbk_ce yes
02:02 Hyratel witha complete sentence, please XD too many Qs for yes/no
02:02 vbk_ce and no, there is a 1000m planar surface region
02:03 Hyratel it's better if you kill Hvel first
02:03 Walkerdine I'm not exactly sure how much different C is than C++ but either of them its still not much
02:03 Hyratel killing Vvel first means you have to then do horizontal braking
02:03 Hyratel which is a /pain/
02:04 anonnumberanon well in my understanding C++ is object oriented and has a lot of libraries but I'm sure I'm wrong
02:04 anonnumberanon about the libraries
02:04 Hyratel Walkerdine, do you write C inside OS or C for uCs?
02:04 anonnumberanon Walkerdine, you must become very good at C as soon as possible, your power level will increase greatly as a result in your field
02:05 vbk_ce I solved the first quest, but I'm not proud of it. Thats why I would like to model the problem first
02:05 Walkerdine I feel like going to community college for 2 years was a mistake
02:05 anonnumberanon I assume he is currently taking his first programming class. Is it Java? C++?
02:05 Walkerdine Yup
02:05 Walkerdine C++
02:05 anonnumberanon Nice
02:05 Hyratel Walkerdine, get yourself an arduino and the trimmings from sparkfun
02:06 Walkerdine I think thats my next task exactly
02:06 anonnumberanon Do you know how fast you can output Hello World in the terminal per second on your computer Walkerdine ?
02:06 anonnumberanon That's always a neat thing to measure
02:06 anonnumberanon while(1), increment a counter and display counter value
02:06 Hyratel https://www.sparkfun.com/products/12789
02:07 anonnumberanon my laptop goes at thirty thousand times a second :)
02:07 anonnumberanon I should try it on my desktop.
02:07 Hyratel https://www.sparkfun.com/products/12060
02:07 Hyratel spendy but worth it
02:07 Hyratel vbk_ce, ok
02:07 Hyratel so you've got a Hvel and Vvel
02:08 Hyratel are they explicitly available?
02:08 Hyratel abstracted available
02:08 Hyratel or do you know Provel heading and magnitude at each sample time
02:08 Walkerdine I got a station set up to do this its just hard to find the time
02:09 Walkerdine I have almost all the basic components I need I think
02:09 Walkerdine like capacitors and resistors and what not
02:09 Hyratel you need sensors
02:09 Hyratel they're how you interact with the world
02:09 vbk_ce Hyratel, they are sensor variables.
02:09 Hyratel Hvel, Vvel?
02:10 vbk_ce ... data
02:10 vbk_ce yes
02:10 Hyratel so you don't have to manually decompose the vector?
02:10 vbk_ce but to compute the next speed I think I need to
02:10 Hyratel the net speed?
02:10 Hyratel the provel magnitude
02:10 Walkerdine yeah, I'm going to try to find one without all those extra parts cause I'm short on cash
02:10 Hyratel and heading
02:11 Walkerdine but I do need sensors
02:11 Hyratel browse around sparkfun
02:11 anonnumberanon How do you get nice 3D mapping with a non-$1000 laser sensor Hyratel ?
02:11 Hyratel anonnumberanon, quick or slow?
02:11 anonnumberanon Hyratel, here is the deal
02:12 anonnumberanon as a senior design project, I am trying to emulate the function of this sensor but with very little money to get the parts https://secure.robotshop.com/en/hokuyo-urg-04lx-ug01-scanning-laser-rangefinder.html
02:12 vbk_ce (what is provel?)
02:12 Hyratel at full speed?
02:13 Hyratel prefixes pro- retro
02:13 anonnumberanon well I can use any speed increase I can get
02:13 Hyratel pro- forward
02:13 vbk_ce ah
02:13 vbk_ce sorry
02:13 Loshki anonnumberanon: cant you use the lidar from a Neato?
02:13 anonnumberanon the one offered by this specific sensor (100msec/scan) is more than enough for the goal
02:13 Hyratel provel : provelocity heading
02:13 anonnumberanon Loshki, what is that?
02:13 vbk_ce thanks
02:13 Walkerdine My school has a PCB etcher or whatever
02:13 Hyratel anonnumberanon, that's way above my paygrade
02:13 Walkerdine and I can use it anytime
02:14 Hyratel vbk_ce, but really, you need the Hvel and Vvel components more than you need the vector heading and magnitude
02:15 Hyratel first, you cancel out Hvel
02:15 Hyratel so you're going straight down
02:15 Loshki anonnumberanon: a domestic robotic vacuum, the competition to Roomba. The neato uses a lidar, a you can pick up beaters on ebay for < $100. www.neatorobotics.com
02:15 vbk_ce but I need to go above the plane surfacer first
02:15 Hyratel vbk_ce, ehwhat?
02:15 Hyratel explain more clearly please
02:16 anonnumberanon oh!
02:16 vbk_ce because at the begining of the "game"
02:16 vbk_ce ...
02:16 Loshki anonnumberanon: oh good, or oh crap?
02:16 vbk_ce they give a map of the surface, and there is a plane region
02:17 Hyratel oh
02:17 vbk_ce and that is where I need to go first, I think.
02:17 Hyratel so you need to Fly To the LZ first
02:17 anonnumberanon Loshki, oh interesting! but also, oh crap, now I have to figure out how fast it samples and its range, and how to make it spin in order to get a whide angle range of measurements
02:17 Hyratel so yes, it /is/ like Lunar Lander
02:18 Hyratel anonnumberanon, mirror on a fanmotor
02:18 anonnumberanon Hyratel, I have seen that project.
02:18 anonnumberanon But the author seems to have gotten stuck and the whole apparaturs is quite big.
02:19 anonnumberanon Plus poor documentation.
02:19 vbk_ce so maybe there is a solution that I can study
02:19 Hyratel vbk_ce, it's getting way too freaking late for me to be thinking this hard.
02:19 vbk_ce ok, no problem
02:19 Hyratel I wish you luck though, and I hope my input has given you a direction
02:20 vbk_ce yes. Thank you very much
02:20 Loshki anonnumberanon: already done, by cleverer people than me. See e.g. http://xv11hacking.wikispaces.com/LIDAR+Sensor
02:20 vbk_ce Good rest to you.
02:21 anonnumberanon Loshki, thank you so much
02:21 anonnumberanon now let's get to reading
02:22 Loshki anonnumberanon: let me know if you want more neato mod urls...
02:23 anonnumberanon ok, yeah I'm trying to get the specs right now
02:25 anonnumberanon I-am-very impressed by this machine https://www.youtube.com/watch?v=4EN01xdgbK4
02:25 Loshki anonnumberanon: http://profmason.com/?p=13246 and this http://blog.tkjelectronics.dk/2014/08/handheld-xv-11-lidar-with-stm32f429-and-matlab/
02:28 Walkerdine To bed I go
02:28 Walkerdine Nice talking with all of yall
02:28 Walkerdine I'll be back tomorrow
02:29 anonnumberanon k bye dude good night
02:29 anonnumberanon dream in code
02:29 anonnumberanon Oh my god! "With a low-cost CMOS imager and a DSP for subpixel interpolation, we get good range resolution out to 6 m with a 5 cm baseline, at a 4 KHz rate."
02:30 anonnumberanon this may just be my ticket if I can make this happen
02:46 anonnumberanon Loshki, I can't thank you enough.
02:47 anonnumberanon Our project proposal presentation is wednesday and this was the weak link of the whole project.
02:48 Loshki anonnumberanon: hope it goes well, let me know if I can help.
02:49 anonnumberanon Yeah I'm fixing to buy a few Neatos off of ebay. This guy's solution is amazing look: https://www.youtube.com/watch?v=WkW55b-WQx4
02:52 Loshki anonnumberanon: mine is even better. It does the vacuuming as well :-)
02:53 anonnumberanon hahaha
02:53 anonnumberanon wow actually this guy wants to mount this on top of a quadcopter, exactly my project
02:56 Loshki anonnumberanon: it's pushing it to expect all the above and crashproof too?
02:58 anonnumberanon well I did find some pretty much unbreakable quadcopter frames... used for quadcopter combat games, the propellers would be the weak points, unless craddled by guard rails
02:59 anonnumberanon and of course all the extra strenghening would add to the weight
03:00 anonnumberanon akin to armored hummers in the military which lose performance due to the weight
03:00 anonnumberanon not too big of an issue I would guess
03:03 anonnumberanon there is an extra capability I haven't talked about yet :)
03:03 shki plays a fanfare then a drumroll. Take it away, anonnumber
03:04 anonnumberanon real time video stream from the quadcopter sent to a remote machine for remote viewing
03:05 anonnumberanon (That's the easy part heh)
03:05 Loshki anonnumberanon: remote viewing, or what they call fly-by-wire (or something?). The latter is notoriously difficult I understand.
03:07 anonnumberanon the craft is autonomous so no pilot, just all kinds of neat feedbacks
03:08 anonnumberanon although I don't see adding a few keyboard keys to take control any harder than when the brain will send to the flight controller as far as pilot commands go
03:09 anonnumberanon what I'm trying to say is that there will be no RC remote control
03:10 Loshki anonnumberanon: very cool. Got any autonomous navigation software? How will that bit work?
03:13 anonnumberanon ATmega328 or similar for flight control, coding it to keep the craft stable seems to be the most challenging part so far. It will hover at an altitude predetermined and this will be checked by a common distance sensor pointed at the ground. Side to side strafing stability will be done thanks to the 3d mapping.
03:13 anonnumberanon (no GPS sensor)
03:14 anonnumberanon atmega328 for the time-critical flight control and raspberry pi for the 3d mapping and all the other capabilities
03:19 anonnumberanon the question that remains is whether or not I can implement getting more points than just the outline of the walls because that is 2D, the entire room in 3D would be a lot nicer
03:21 Loshki anonnumberanon: how long do you have for all this?
03:22 anonnumberanon May 2015, 5 other people, I am team leader
03:24 Loshki anonnumberanon: how's your team? strengths/weaknesses etc?
03:26 anonnumberanon 4 chinese guys fresh off the boat, half of them struggle with english
03:26 anonnumberanon 1 american kid who is clueless about hardware and coding
03:26 anonnumberanon I get results by overmanaging them and playing cool.
03:27 anonnumberanon (believe me I don't feel cool sometimes lol)
03:27 anonnumberanon When I ask one guy to make a block diagram of and research an electronic speed controller and no work is done
03:28 Loshki anonnumberanon: been there, don't miss it.
03:29 anonnumberanon I got a guy however who seems to do a lot of things well.
03:29 anonnumberanon I have him and myself the task of doing the 3D mapping.
03:31 anonnumberanon I got them all to install Linux today it was a good day.
03:39 Loshki anonnumberanon: it's 1:20am here, I'm off to bed. Best of luck.
03:39 anonnumberanon thanks I'll see you around I take it you're in Cali?
03:41 Loshki anonnumberanon: yes :-). If you say my name, I'll see it eventually, even if it's not till the next day :-)
03:42 anonnumberanon yeah thanks again for the help
03:42 anonnumberanon I'll let you know how the presentation went
03:45 Red_Onyx I haven't gotten started with the Raspberry yet
07:05 roxlu Hi
07:05 roxlu Someone from #electronics pointed me to this channel, great to see there is a robotics channels
07:05 roxlu Someone around who has programmed a ABB robot maybe? I’m wondering what the best way is to create a networked RAPID application.
07:07 xlu
07:07 R0b0t1 <roxlu> R0b0t1: thanks. I know how to create networked applications in C/C++/.. but when working this out in RAPID Im getting different results then I expected. e.g. synchronising two tasks seems to be extremelly slow
07:08 R0b0t1 Can you explain ABB and RAPID and what you mean by synchronizing?
07:08 roxlu Hehe ok
07:08 R0b0t1 i.e. there's probably not a 5min answer, but there still might be one
07:09 roxlu ABB is the robot manufacturer, I’m programming for this one: http://www.mach.ro/images/detailed/1/irb140_gfh5-s5.jpg
07:09 roxlu One can program such a robot using a language called RAPID
07:09 roxlu I want to create an application where I send commands (like move to this position, toggle I/O etc.) over a socket
07:10 roxlu With RAPID it seems you can do some sort of multi threading, e.g. creating what they call tasks
07:10 R0b0t1 Socket? As in POSIX socket on a computer which represents a TCP/IP connection to the robot?
07:10 roxlu Yes, posix socket
07:11 R0b0t1 Ok
07:11 R0b0t1 what are the threading primitives and what does the language look like
07:11 R0b0t1 how have you defined your tasks and how are you measuring their response and how is it slow
07:11 roxlu I’m not sure what kind of threading primitives they support but it seems they simulate condiation variables/muteces in one kind of type and they have support for interrupts
07:13 roxlu R0b0t1: Those are actually exactly the things I’m wondering about .. how can I measure the performance correctly and can I use e.g. printing to the console (which is possible with RAPID) as a somewhat real-time indicator of what’s happening in my code.
07:13 roxlu For now I’m just sending one command every second, just to see if the code I wrote in RAPID does what I think it should do
07:13 roxlu (I’m a new with RAPID btw)
07:14 roxlu For example, I’m not so sure that writing to the console works how I would think it work, i.e. it outputs directly when I write to it
07:14 roxlu I seems to buffer some output first but this is not mentioned anywhere in the manual / reerence
07:15 roxlu s/I/it
07:15 roxlu So I have a couple of these questions which can probably easily answered by someone who has been working with RAPID / ABB longer then a couple of weeks
07:15 R0b0t1 roxlu: Well you said it was slow
07:15 R0b0t1 on what basis did you say that
07:16 R0b0t1 And yeah I've got no idea about RAPID
07:16 R0b0t1 can you provide a testcase of what you're doing
07:17 roxlu it’s not something you can test easily w/o any knowledge of abb/robotstudio
07:17 R0b0t1 roxlu: So basically the skills you're looking for you're really only going to find if you offer to pay someone for it
07:18 roxlu R0b0t1: yes I guess so
07:18 R0b0t1 Anyhow, most people (myself included) are likely to have enough experience with programming and shitty proprietary languages that they can figure out what is going on
07:24 roxlu R0b0t1: are you willing to dive into this a bit?
07:25 R0b0t1 sure but I don't think it'll get that far
07:25 R0b0t1 it's sunday
07:25 roxlu yeah exactly :)
07:25 R0b0t1 you've got me until noon
07:25 roxlu thanks anyway
07:25 roxlu I’ll wait in here and maybe someone comes along who programs ABBs
07:26 R0b0t1 heh
07:26 R0b0t1 that's not going to happen
07:27 R0b0t1 your problem isn't as specific as you think it is
07:27 R0b0t1 I mean if you keep insisting I guess I will drop it but I'm sure I can solve your problem in like 30mins, or tell you where to look for an answer, *if* you give me what I asked for
07:28 roxlu Lol please prove me wrong but you’re not going to solve this in 30min from scratch :)
07:28 R0b0t1 uhuh, sure
07:28 R0b0t1 sample code please
07:28 roxlu I’ll rephrase what i’m after
07:30 R0b0t1 Yes?
07:30 roxlu I want to send ‘Commands’ over a socket connection to a ABB controller. A Command is like a basic task-based protocol command. E.g. turn on I/O port, move to this point etc.. I want to collect commands in a ‘buffer’ and upon receiving a special command like “execute” I want to hand over all received commands to the main task and let the robot do whatever is instructured by these commands (for now just moving on a 2D plane).
07:31 R0b0t1 Okay, and
07:32 roxlu And I need this to be done in RAPID which can be programmed using RobotStudio which you can download here: http://new.abb.com/products/robotics/robotstudio/downloads It has superb manuals with plenty of code examples.
07:33 R0b0t1 VERSION HexChat 2.11.0 / Linux 3.2.0-70-generic [x86_64/1.20GHz/SMP]
07:33 R0b0t1 I'm not sure I can run that.
07:33 R0b0t1 (as I expected)
07:34 R0b0t1 Additionally, it seems like you're trying to con me into doing your homework.
07:34 R0b0t1 You didn't describe a problem... just a solution. Which isn't yet in the form it needs to be.
07:34 R0b0t1 Is the manual accessible online?
07:35 R0b0t1 If I know the primitives of the language I'll give you a sketch of an implementation
07:35 roxlu R0b0t1: if you’re running linux you’re not going to get this workign in 30 min :0
07:36 R0b0t1 false
07:36 R0b0t1 well okay I said I could solve your problem in 30mins
07:37 b0t1 opens ma
07:38 R0b0t1 Ooo these are nice machines
07:38 R0b0t1 Okay so what I was asking for is on page 26
07:38 roxlu Of what manual?
07:40 R0b0t1 http://www05.abb.com/global/scot/scot352.nsf/veritydisplay/6a22220e833bff4bc1257cc3006fdc05/$file/3HAC032104-en.pdf
07:40 roxlu ok
07:41 R0b0t1 ffs this is organized by menu location not functionality
07:42 roxlu there are also other manuals
07:42 roxlu and a reference
07:43 R0b0t1 where?
07:43 roxlu in the help of the application, not sure if they are also online
07:43 roxlu (they use htmlhelp)
07:44 R0b0t1 http://www05.abb.com/global/scot/scot241.nsf/veritydisplay/b755e615bec21b52c1257c32004fa70f/$file/PR10296EN_R3.pdf
07:45 R0b0t1 oh look some sample code
07:45 roxlu I googled this: http://rab.ict.pwr.wroc.pl/irb1400/datasys_rev1.pdf
07:45 roxlu looks a bit old though
07:45 R0b0t1 alright so ABB makes nice stuff but I already hate them
07:45 R0b0t1 they did the thing
07:46 roxlu http://www05.abb.com/global/scot/scot241.nsf/veritydisplay/688894b98123f87bc1257cc50044e809/$file/Technical%20reference%20manual_RAPID_3HAC16581-1_revJ_en.pdf
07:46 R0b0t1 they structured their product in such a way as to breed confusion
07:46 roxlu that’s a newer reference
07:46 roxlu :)
07:46 R0b0t1 I see why you thought it would be necessary to have previous experience
07:47 roxlu R0b0t1: yes I was trying to explain that, this needs experience. Also to work around how things work in their robot studio
07:48 R0b0t1 no, it doesn't
07:48 R0b0t1 it needs the ability to cut through copious amounts of bullshit and bad design choices
07:48 R0b0t1 psuedocode in a minute or two
07:54 R0b0t1 actually this is too simple for pseudocode unless you can point to a more specific problem you are having after my explanation
07:54 R0b0t1 you're looking to implement a remote procedure call interface
07:54 R0b0t1 https://en.wikipedia.org/wiki/Remote_procedure_call
07:55 R0b0t1 so, how do you do this in the context of robot studio
07:55 R0b0t1 you have the data types and control structures to accept remote calls from a socket, batch them, and then execute them
07:55 R0b0t1 you might be surprised to know this is rather similar to how SQL works
07:55 roxlu Yep exactly super simple
07:56 R0b0t1 anyway, the tricky part is interpreting the remote procedure calls. I don't see any metaprogramming facilities in ABB's language so you'll basically have to wrap every function you want to call
07:56 roxlu if it only worked like that. One issue for example: I connect to the remove socket and directly send some commands but these commands never seem to arrive
07:56 roxlu When I add a delay of 100ms after connecting and before sending the first command it works
07:56 R0b0t1 ..
07:56 R0b0t1 So wait 100ms
07:57 R0b0t1 Or, more properly, figure out how to check the status of your connections
07:57 roxlu Right…. I can also move the robot manually using a joystick, but that’s not what I want
07:57 R0b0t1 I never said you would need to do so
07:57 roxlu I -am- connected :)
07:57 R0b0t1 so what is your problem
07:57 roxlu exactly :)
07:58 R0b0t1 ?
07:58 R0b0t1 are you trolling?
07:58 R0b0t1 you've just told me you are capable of solving the problem you originally presented
08:00 R0b0t1 hmm
08:01 R0b0t1 perhaps I see the problem
08:01 R0b0t1 "The integrated RAPID editor is useful for editing all robot tasks other than robot motion."
08:01 R0b0t1 what you are wanting to do may be impossible if you wished to execute arbitrary movement over your rpc interface
08:02 roxlu you can interpret that in many ways
08:02 R0b0t1 and it seems it is impossible by design
08:02 roxlu It’s less usefull because it’s not so easy to position the robot using code. You can also use a 3D model to position it
08:02 R0b0t1 wow are you serious
08:02 roxlu again: experience :)
08:03 R0b0t1 this has nothing to do with experience and more to do with you not knowing what you want
08:03 R0b0t1 http://www05.abb.com/global/scot/scot352.nsf/veritydisplay/6a22220e833bff4bc1257cc3006fdc05/$file/3HAC032104-en.pdf#newid-425
08:03 R0b0t1 pg 441
08:03 R0b0t1 literally what you want to do but with the IO module instead of sockets
08:04 R0b0t1 or wait nm
08:05 R0b0t1 roxlu: *You* don't find it easy to position the robot using code.
08:05 R0b0t1 I find it perfectly easy
08:06 roxlu great, please tell me how you handle singularity issues in your code?
08:10 R0b0t1 I don't, because they don't exist.
08:10 roxlu They surely do.
08:11 roxlu But I want to thanks you for willing to help me but this is not getting us any further. I’m going to continue working on this.
08:13 R0b0t1 roxlu: If you insist, but you seem to be incapable of solving your problem.
08:13 roxlu false
08:13 R0b0t1 Also, no. It is not true that a machine path must go through a singularity.
08:14 R0b0t1 It might be likely, in which case... deal with them like you have been dealing with them by limiting the acceleration of your axes, which you should have been doing anyway
08:15 R0b0t1 You've got to be trolling because this is the first time someone has managed to piss me off in a long while
08:15 R0b0t1 Familiarity with some arbitrary, proprietary software does not imply ability in a field
08:16 roxlu Probably because someone actually says that you’re talking nonsense by saying you can get this working in 30min (which was 30+ min ago).
08:16 R0b0t1 You kept changing your problem!
08:16 R0b0t1 YOU DON'T KNOW WHAT YOU WANT
08:17 R0b0t1 I solved like 3-4 different problems
08:17 R0b0t1 Is the problem "I don't have this program already finished"?
08:17 R0b0t1 if so go fuck yourself
08:17 roxlu Ah we clearly have a different view of what ‘solving” means :)
08:18 R0b0t1 says the guy who came onto irc asking for free engineering services
08:18 roxlu Add 100ms delay is not a production quality way to solve the something :)
08:19 R0b0t1 I noted as much
08:19 R0b0t1 <R0b0t1> So wait 100ms
08:19 R0b0t1 <R0b0t1> Or, more properly, figure out how to check the status of your connections
08:19 R0b0t1 Oh you mean you're connected but it takes 100ms to respond
08:19 R0b0t1 see
08:19 R0b0t1 I asked you for information related to *that* way at the beginning
08:20 R0b0t1 what does your code look like so I know what is causing the hang
08:20 R0b0t1 skafjlskjdfa
08:20 R0b0t1 but no you wouldn't give it to me
14:19 rue_house http://learnfx.biz/wp-content/uploads/2013/08/ANIMATRONICS-AT-LEARNFX.BIZ_.jpg
14:19 rue_house hmmm
14:21 Hyratel those look like screen door springs or (pushwire sleeves - term?)
14:21 Hyratel throttle cable sleeves
14:23 GuShH meh just say bowden :p
14:23 LeelooMinai Any tip as to what I should look for in terms of torque for, I guess, a servo that would have to rotate say 50kg/100lbs on a pole/mast remotely (using brake cables) - I just need some general sense idea:)
14:23 Hyratel that's the term I couldn't remember
14:24 GuShH depends on leverage and other factors, LeelooMinai
14:24 GuShH is that a static or dynamic load, etc
14:24 GuShH also expected service life
14:24 GuShH and budget
14:24 GuShH meh I'm out
14:24 Hyratel GuShH, just ignore LeelooMinai
14:24 GuShH why
14:24 Hyratel I did, and haven't regretted it
14:24 GuShH lol
14:25 ShH is cleaning up the T slots on the milling machine t
14:26 GuShH so much build up of dirt, oil, flakes of paint, brass swarf... you couldn't even slide a T nut before!
14:26 LeelooMinai It's a mast with two solar panels. The leverage I will be able to adjust as I will drive the rotation from the bottom, using two break cables that will run into the house where the actual servo/motor will be.
14:26 GuShH now I'm going to passivate the slots, then paint 'em blue
14:26 GuShH LeelooMinai: avoid servos, unless they are industrial grade... which, are not cheap
14:27 LeelooMinai I could use some stepper, but not sure if it could handle the torque (after some gearing).
14:27 GuShH you do need positional feedback unless you want to go open loop
14:27 GuShH speed is not an issue
14:27 LeelooMinai Yes, I will handle feedback/electronics
14:27 GuShH so you can step way down
14:27 GuShH to gain torque
14:28 LeelooMinai Yes, speed can be snail-like
14:28 GuShH there ought to be a lot of solar tracking projects out there
14:28 GuShH cheapest stuff is with bike chains and pinions
14:28 LeelooMinai Yes, but they are mostly based on linear actuators not a mast like I will have.
14:29 GuShH because gearboxes in general are not cheap, and not easy to make.
14:29 GuShH you want to mimic a ship sail?
14:29 GuShH ok rephrase: they are relatively easy to make, if you own a machine shop and have the time for it
14:30 LeelooMinai I imagine I could use a setup where I have a big circle with gears on the edge and drive it with stepper - that would be simple and should give me the ratio
14:30 GuShH look into planetary gearboxes :p
14:30 LeelooMinai Or maybe use belt-driven setup?
14:31 LeelooMinai Planetary gearbox sounds expensive
14:31 GuShH if this is going indoors, sure
14:31 GuShH you can get an ac motor with a reduction gearbox
14:31 GuShH that already spins slow and has a good torque
14:31 GuShH then step it way down with belt and pulley setup
14:31 GuShH it is going to get relatively big, in size
14:31 LeelooMinai Yes, indoors - that's why I want to use those break cables, to avoid having this driving motor exposed to -30 deg C...
14:32 GuShH otherwise, find an old gearbox/motor from a traffic arm, like the ones they use on parking lots
14:32 GuShH they are way common, lots of buildings have them
14:32 GuShH also you can gang the common reduction gearboxes to step it way down
14:33 LeelooMinai Ok, great, some ideas - thx
14:33 GuShH another idea... an old motorcycle transmission, can be had cheap in bulk and make your own transmission
14:34 GuShH or just get a big stepper and get the ratio down with a GT pulley
14:34 GuShH tracking doesn't have to be smooth, but it would need a fast mode if you wanted to catch up in a relatively low amount of time
14:35 GuShH a 3 phase ac motor with a reducer and a vfd would be the easiest way, but also most likely the least efficient and not cheap either :(
14:35 LeelooMinai Yes, right, no, I need to make it rather cheaply - not much spare money for that.
14:36 GuShH who has "much spare money" for anything really
14:36 GuShH unless you are disgustingly rich in which case chances are you wouldn't be on IRC
14:36 LeelooMinai Rich people:)
14:36 GuShH yeah well they hardly ever do anything productive do they
14:37 GuShH because they can afford to just sit on their ass all day long doing nothing
14:37 GuShH so why bother.
14:37 LeelooMinai SOme do - like this guy, what is his name? Musk? The one from Tesla and rockets.
14:37 GuShH no idea on names...
14:37 GuShH he just points his finger and has stuff made
14:38 LeelooMinai He has engineering knowledge and has real input into those projects afaik
14:38 GuShH like the koenigsegg guy
14:38 GuShH input, but he would never break a sweat
14:39 GuShH huge difference between a bunch of rich guys throwing money around and actually doing something for the greater good
14:39 ShH doesn't buy into the electric cars
14:39 GuShH likewise, solar doesn't make sense where I live
14:39 GuShH cost of batteries, etc. you never break even
14:40 LeelooMinai Right, that was my point - he seems not to be one of those: he has ideas and then gets them done. They rockets are quite successful.
14:40 LeelooMinai eir*
14:41 LeelooMinai My solar project will not even use batteries - it's just two 200 watt panels that will offset the minimum house power consumption hopefully - it never goes less than 300watt
14:42 LeelooMinai So justs inverter directly into mains.
15:33 GuShH LeelooMinai: but 400W worth of solar panels means you maybe get 200W usable in average
15:33 GuShH you might peak at 400W depending on your electronics
15:33 GuShH and we are assuming you are tracking the sun, but efficiency goes to hell anyway
15:34 GuShH say your inverter for instance it may be efficient at a certain load, you still got losses there.
15:34 GuShH then dust and contamination in general to the panels will lower your efficiency as well
15:35 LeelooMinai Yes, but I will be happy if it offsets the bill a bit and gets payed back in long run - say 5 years even.
15:35 GuShH in 5 years your panels may be barely working if you got hail storms or hipsters with rocks in town
15:36 GuShH I'm considering solar heating for water, but nothing else.
15:40 GuShH LeelooMinai: it helps a lot if you don't leave wallwarts plugged in, lights on when you aren't using them...
15:41 GuShH switch to LED if you can
15:41 GuShH or at the very least cfl / fluoros
15:41 LeelooMinai Yes, doing both
16:12 Tom_itx http://i.imgur.com/PSaJyLu.gif
16:16 rue_house see? thats cool
16:17 rue_house https://www.youtube.com/watch?v=cR-YlZ9NdIA
16:18 rue_house the gif is better...
16:21 rue_house the plastic boxes I use for screws have been disintegrating for a while now, today I finally managed to get around to making a few wooden boxes
16:23 Tom_itx don't let termites get to them
16:25 Tom_itx i wish i had more of those stamped aluminum boxes i have
16:25 Hyratel GuShH, would adding mirrored perpendicular louvers on the solar panels help in a setting where you can't suntrack?
16:39 rue_house Hyratel, carefull, you can easily spend more on trying to get that little extra bit of light than it'll give you back
16:39 Hyratel oh i know
16:40 Hyratel just musing on hardware-efficient ways to
16:40 rue_house http://learnfx.biz/wp-content/uploads/2013/08/ANIMATRONICS-AT-LEARNFX.BIZ_.jpg <-- give it tenticles!
16:41 Hyratel hard part of replicating that - finding bowden sleeves soft enough to be effective
16:41 Hyratel level 2 design: add another pair of servos and a second independent actuation segment
16:41 Hyratel mastery level: go for broke with 5 segments of actuation
19:20 rue_shop2 the ones on arm9 worked out well
19:21 rue_shop2 well, I just built about 1/6 of the 'other' retainer wall, I'm ready for a break
19:21 rue_shop2 one thing I learned is not to sheath the cable if you dont have to
19:22 rue_shop2 free-air it where you can
19:26 rue_house this back breaking stuff is back breaking, I need to make a robot to do it
19:26 rue_house oh wait, I am
19:26 rue_house I need to work on it
19:27 rue_house if you saw a 12' robot, how do you think its elbow would be contructed?
19:39 rue_house http://s531.photobucket.com/user/innerbreed/media/underside-1.jpg.html
19:39 rue_house I'm surprised to see someone just using plain tube
19:45 Tom_itx this ain't gonna be easy
19:46 rue_house https://www.youtube.com/watch?v=CiFWZ8MC2cE
19:46 rue_house ahahhaah
19:51 Tom_itx rue_house
19:51 Tom_itx what do you know about motor winding?
19:52 Tom_itx do the windings go to ajacent commutator bars or 180 deg opposite?
19:52 Tom_itx this says they are ajacent
19:53 rue_house I know that its not worth it to try to rewind any motor
19:54 Tom_itx i know
19:55 Tom_itx i just was messin with it since i can't really hurt it any more than it already is
19:55 Tom_itx i'm not gonna rewind it, rather just hook the wires back up where they were pulled loose
19:58 rue_house your lucky if they all broke off at the commutator side, most of the time they will break off from within the stator
19:59 rue_house the offsets between the windings and the commutator will be the same all the way around
19:59 Tom_itx they all pulled from the loops at the commutator
19:59 Tom_itx i can see it pretty plainly
19:59 Tom_itx they were just crimped in place
19:59 Tom_itx and the crimps pulled
20:00 Tom_itx near as i can tell there are 2 wires per coil
20:00 rue_house the alignment between the commutator and the windings is important as its part of the timing
20:00 Tom_itx i know
20:00 rue_house usually they are crimped away from the windings, interesting
20:01 Tom_itx it's odd because it looks like one pair is ajacent but the other of the pair may be opposite
20:02 Tom_itx it looks like one wire wraps around the shaft to the other side
20:03 Tom_itx would that make sense at all?
20:03 Tom_itx one ajacent winding and one opposite?
20:03 rue_house yup
20:04 rue_house motors do not activate one winding at a time,
20:04 rue_house they release one winding at a time from operation
20:04 rue_house so all but one winding generates the field at any given time
20:05 Tom_itx looks like one coil goes thru the iron and another coil goes thru the same slot in the iron a different direction
20:05 rue_house yup
20:05 Tom_itx that may explain why part of the wires wrap around the shaft
20:05 rue_house they meet up an adjacent commutors and carry on to the next winding
20:06 Tom_itx this could be difficult to figure out
20:06 rue_house if the number of poles is .. odd (?) then they may have a dummy winding on the bottom for balance
20:06 Tom_itx they are even
20:06 Tom_itx 16
20:07 Tom_itx and 16 core slots
20:07 rue_house if you look up the offset between the commutator and winding for another motor with the same number of poles, it should be the same
20:07 rue_house hmm
20:08 rue_house can you change the angle of the brushes<->field magnet?
20:08 rue_house sometimes its adjustable
20:08 Tom_itx doesn't appear so
20:08 Tom_itx it didn't twist on the shaft though so it should be aligned
20:09 Tom_itx it slid on it
20:09 rue_house what is the approx angle between the brushes and the middle of the field magnets?
20:09 rue_house (are there 2 or 4 field magnets)
20:10 Tom_itx 2
20:10 Tom_itx i'd have to go check the angle. that part is in the shop
20:11 Tom_itx i just brought the armature in
20:12 rue_shop2 this one is 12..
20:12 rue_shop2 10 on that one...
20:12 Tom_itx there are probably 2 wires on each commutator bar
20:13 Tom_itx near as i can tell
20:13 Tom_itx one should be ajacent? and one opposite?
20:13 rue_shop2 22 on that one...
20:13 rue_shop2 yea
20:13 rue_shop2 they loop thru
20:13 rue_shop2 -
20:13 rue_shop2 n
20:14 Tom_L http://www.groschopp.com/how-to-check-a-motor-armature/
20:14 Tom_L looking at that
20:15 rue_shop2 http://www.torridheatstudios.com/ftp/share/pictures/Tascam%2058/Dirty%20Reel%20Motor%20Commutator.JPG
20:15 rue_shop2 the 3rd wire is a capacitor leg, ignore
20:15 Tom_L yeah
20:15 Tom_L mine is quite similar
20:15 rue_shop2 that lower diagram is the good one
20:16 rue_shop2 (your page)
20:16 Tom_L 180deg opposite?
20:16 rue_shop2 depends on the brush angle
20:16 rue_shop2 thats how the time the windings
20:16 Tom_L i know that much
20:16 orlock got webcam working with my baby router
20:17 Tom_L if i can find the first one the rest may not be quite as bad
20:17 orlock now i just need to stick the webcam outside
20:17 rue_shop2 the offset is matched with the angle between the field magnets and the brushes
20:17 orlock point it at one of the weird holes in the ground
20:17 Tom_L the top diagram in my link makes me think part of them are ajacent
20:17 Tom_L ?
20:17 rue_shop2 orlock, how big wierd holes?
20:17 orlock figure out what the fucks living in them, - maybe i dont want to know
20:17 orlock 1"
20:17 orlock could not fit a golf ball down there
20:18 orlock probably a frog
20:18 rue_shop2 orlock, your supposed to use metric in australlia
20:18 Tom_L in Au you probably don't want to know
20:18 orlock i hope its a frog
20:18 orlock Tom_L: if i was in NSW/Sydney, i'd be nope'ing away and filling it with concrete
20:19 orlock http://en.wikipedia.org/wiki/Sydney_funnel-web_spider
20:20 orlock rue_shop2: i know! (re: metric vs yank)
20:21 orlock i got my SPI working, but i cheated
20:21 orlock i just plugged it into an arduino
20:22 rue_shop2 heh
20:22 rue_shop2 spi to what?
20:23 orlock Tom_L: my parents house and the last place i lived, both pretty urban had redback spiders in the backyard - our place is out bush, no redbacks (havent spotted a snake either) but have seen echidnas and scorpions
20:23 orlock rue_shop2: nokia LCD
20:23 orlock rue_shop2: i'
20:23 Walkerdine I want to buy something from sparkfun but I need some more sparkfunds
20:23 orlock 'll probably just wire than and a serial LCD up to an arduino for displaying "stuff"
20:24 Tom_L rue_house
20:25 Tom_L would it be bad to solder the wires to the commutator if i get them figured out?
20:25 Tom_L (this is just something to do to pass the time)
20:26 rue_shop2 no, most repair stuff solders them
20:26 rue_shop2 sometimes the crimps disconnect and you have to solder them anyhow
20:28 Tom_L ok
20:41 Tom_L 1.6 ohms seems to be the ajacent ones
21:37 rue_house yes, each coil would jump 1 commutator set
21:38 rue_house sets 180 out via winding will be 180 out on commutator
21:39 Tom_L 6 out of 16 fixed
22:10 Walkerdine I'm guessing everyone here is mostly a hobbyist
22:17 rue_shop2 sure
22:18 Tom_L 3 poles left...
22:28 Loshki Walkerdine: I would think so. IRC is kind of a dusty corner of the internet.
22:29 Walkerdine I've always had no luck talking to people on forums for anything
22:29 Walkerdine I used to
22:39 Tom_L ok rue_house
22:39 Tom_L i think i got it
22:42 Tom_L i get about 3.2 ohms across ajacent poles now and about 5.2 across opposites on all poles.
22:42 Tom_L no conductivity between poles and shaft or poles and core
22:45 Tom_L not an exercise for the faint of heart
23:50 rue_shop2 indeed not
23:51 rue_shop2 I just finished new bearings for the heater blower
23:51 rue_shop2 some idiot used bushings