#robotics Logs

May 06 2014

00:13 rue_more its probably good now, i'm gonna leave it another day
00:13 rue_more apparently I need 10% bentonite
00:13 rue_more I put about a liter in, so I should get 11L of casting sand :)
00:13 rue_more wait a sec... thats not nearly enough...
00:15 rue_more if only I could pummel aluminum into dust and sonically fuse it into the shapes I want
00:18 rue_more if I had all the time in the universe would anything not be worth doing?
00:23 rue_more I cant come up with a good gui idea for controling the robot arm
03:30 GuShH_ rue_more: there's the original but also a fork of the original... not sure which one to go with
03:30 GuShH_ this is the fork http://www.machineconcepts.co.uk/engines/nelson/nelson.htm
03:30 GuShH_ it mentions the original (boll aero) which is imperial
03:31 GuShH_ I think he complicated it by adding the bearings instead of going with the bushings
03:31 GuShH_ and some other extra parts
03:32 GuShH_ hmm nobody sells B12 MT2 arbors here
03:32 GuShH_ non standard :(
03:35 GuShH_ check out his radius milling tool for the lathe, specific for the conrod http://www.machineconcepts.co.uk/engines/nelson/conrod_4.jpg
03:35 GuShH_ heh
03:36 GuShH_ also notice the adapter for the flat rod http://www.machineconcepts.co.uk/engines/nelson/conrod_2.jpg
03:36 GuShH_ to avoid using the 4 jaw chuck
03:42 GuShH_ he did increase the bore size, so that means slightly bigger iron stock... luckly I think mine is big enough
03:47 GuShH_ damn he uses two different bearings... 618-7 and 628-8 which are rather small and I don't know if I can source them..
03:51 GuShH_ "meehanite cast iron" hrm
03:55 GuShH_ nice, uses an M16x40 cap screw for turning the crankshaft
04:11 GuShH_ holy shit magnesium castings http://www.machineconcepts.co.uk/engines/atom/atom.htm
08:33 rue_bed its only tuesday!? I object!
08:34 rue_bed oh I have to move a server room today
08:35 rue_bed hum
09:13 rue_more adjustable displacement, interesting
09:13 rue_more er, compression ratio
09:27 rue_more I'm not used to seeing drawings that integrate the machining stages
09:32 rue_more http://www.machineconcepts.co.uk/engines/nelson/conrod_3.jpg
09:32 rue_more thats an important image
09:32 rue_more he just used a sharpened screw, clever
14:51 f00bar80 I have this code that drives a robot arm / gripper http://pastebin.com/WPsJ542W , but actually the gripper is for some time continues opening/closing , anything wrong in the code ? the gripper's servo is which pluged to pin 9
14:51 Hyratel pin 9 of a what
14:51 f00bar80 Hyratel, Arduino UNO
14:55 f00bar80 am in the wrong channel ?
14:57 Hyratel no, i'm just afk
14:57 Hyratel try in #sparkfun too
14:58 SquirrelCZECH f00bar80: hmm
14:58 SquirrelCZECH first cycle 0->180 servo2
14:59 SquirrelCZECH second cycle 180->0 servo2
14:59 SquirrelCZECH third cycle 0->180 servo
14:59 SquirrelCZECH fourth cycle 0->180 servo2
14:59 f00bar80 SquirrelCZECH, ....?? what do you mean
15:00 SquirrelCZECH f00bar80: people usually want to move two different servers from 0 to 180 to 0
15:00 SquirrelCZECH you did something completely different :D
15:00 SquirrelCZECH and I actually do not really understand what is wrong
15:02 f00bar80 first the gripper opens and close, then arm moves up , after that the gripper keeps opening and closing for like 5 times, then arm goes down. this is the arm here https://www.dropbox.com/s/p7py5jjq1933aja/2014-04-03%2023.02.14.mp4
15:03 SquirrelCZECH seeems like not only software problem
15:04 SquirrelCZECH f00bar80: change it into version, when in first part you got 0->180-> on servo2
15:04 SquirrelCZECH and in second part go 0 to 180 to 0 with servo
15:04 SquirrelCZECH recheck your wiring
15:04 SquirrelCZECH try to switch servo pins
15:05 f00bar80 SquirrelCZECH, but what do you suggest for the coding?
15:05 SquirrelCZECH like I said, change it :D
15:05 SquirrelCZECH the way I wrote....
15:05 SquirrelCZECH f00bar80: you use "myservo2" in first, second and fourth cycle, and "myservo" in third
15:06 SquirrelCZECH f00bar80: and you go 0 to 180 in first, third and fourth cycle... and 180 to 0 in second cycle
15:06 SquirrelCZECH so, change it to 0 to 180 in first/third cycle, 180 to 0 in second/fourth cycle
15:07 SquirrelCZECH and also "myservo" in first/second cycle and "myservo2" in third/frouth cycle
15:07 SquirrelCZECH and now I have to go afk :-/
16:51 TELunus Hyratel, is there not a specifically arduino channel?
17:17 Hyratel TELunus, why would you need one?
17:19 TELunus Well, I guess you wouldn't, but I'm just a little surprised that all the arduino people haven't created one.
17:21 GargantuaSauce there definitely is an arduino channel.
17:40 _Odin_ Hi, what’s the best way to obtain an accurate heading of a robot (that only moves on the ground)?
17:41 _Odin_ Would it be to just combine magnetometer / accelerometer / gyro readings?
17:51 Hyratel _Odin_, yes, though you'll want to determine how much error the structure of the robot incurrs on the magnetometer
17:51 _Odin_ How would I figure that out? Also, ist here any literature that you could point me to in terms of figuring out how to combine magnetometer / accelerometer / gyro readings?
17:52 Hyratel for magnetometer, what I'd suggest would be a "lookup table" with an average for between them
17:52 Hyratel bench the magnetometer in a space without magneic interference, in the 4 cardinal directions
17:53 Hyratel then bench it attached to the robot. compare the difference at the cardinal and 45*
17:53 Hyratel if there's not enough deflection to notice, you can omit any offsets
17:54 _Odin_ Ok, thanks
17:54 _Odin_ Anything about combining values?
17:54 Hyratel from that you can get your headings
17:55 Hyratel if your compass is accurate enough for ground heading, i don't think you'll need to combine a Rotoaccelerometer reading with it
17:56 _Odin_ I don’t think my compass would be accurate enough
17:56 Hyratel if you're wanting to back up the compass heading reading with a secondary though, if you know you'll be working in magnetically-noisy spaces, that's beyond my skill level
17:56 _Odin_ Yeah
17:56 _Odin_ Alright, thanks
17:56 _Odin_ Do you at least know what I would look into to deal with that?
17:56 Hyratel hmmmm
17:56 Hyratel not off the top of my head but #sparkfun crowd might know
19:10 rue_shop2 kitty litter or not, that dosn't get ground much finer
19:11 GargantuaSauce _Odin_: kalman filtering is the technique usually used for that
19:11 _Odin_ Ok. Are there any good tutorials / papers for that sort of thing?
19:12 GargantuaSauce if i can find it in my sea of browser tab
19:12 GargantuaSauce s
19:14 _Odin_ Do you think that an arduino could handle a kalman filter?
19:14 GargantuaSauce probably at a pretty low rate
19:14 GargantuaSauce actually take a look at http://www.scribd.com/doc/29754518/A-Efficient-Orientation-Filter-for-IMUs-and-MARG-Sensor-Arrays
19:15 GargantuaSauce i hope your math is up to snuff.
19:18 GargantuaSauce i would be very wary of the correctness of any arduino-based tutorials or libraries for sensor fusion
19:19 _Odin_ Ok, thanks
19:20 GargantuaSauce also precisely timing your samples is pretty vital
19:20 _Odin_ Sweet, that looks exactly like what I am looking for
19:20 _Odin_ What do you mean?
19:21 GargantuaSauce i mean don't just stick a delay() between each cycle of querying the sensors and doing your calculations
19:22 GargantuaSauce the time between each sample needs to be constant and fairly precisely measured
19:28 _Odin_ What other way could I do it?
19:31 GargantuaSauce the sensors could have provisions for automatically taking samples at a fixed rate, or you might need to use the microcontroller's timers
19:31 GargantuaSauce or both
19:34 _Odin_ Oh right
19:34 _Odin_ Thanks
23:30 GuShH_ rue_shop2: ever tried to make a custom threading die?
23:30 GuShH_ my first attempt was a complete failure...
23:30 rue_more hell no
23:30 GuShH_ rue_more: some sizes are non standard, hence not available everywhere
23:31 GuShH_ and for soft metals, carbon steel works fine
23:31 GuShH_ http://chestofbooks.com/crafts/machinery/Shop-Practice-V1/Tool-Making-Part-II-Standard-Tools-Thread-Cutting-Dies.html#.U2my3vldUg0
23:31 GuShH_ looking at that
23:31 GuShH_ my previous layout was wrong, got me some odd cutting angles...
23:33 rue_more most of the dies I ahve, have no rake
23:33 GuShH_ they recommend negative brake for brass and such
23:33 GuShH_ rake
23:33 rue_more !?
23:33 ShH_ should read what he t
23:33 rue_more wouldn tthat make it stick worse?
23:34 GuShH_ "For brass castings, the cutting edges should have a slight negative rake, as shown in Fig. 129"
23:34 GuShH_ I wasn't sure if I had to thread first, cut the holes later... maybe with an endmill
23:34 GuShH_ or do it the other way around
23:34 GuShH_ interrupted cuts like that suck.
23:35 GuShH_ so I'll try threading after the 4 holes are made
23:35 GuShH_ thing is I don't know if I have an endmill with the right size, and I certainly don't want to use a tapered reamer for an interrupted cut like that
23:36 GuShH_ there's no way a twist drill will cut that center hole without something going wrong
23:37 GuShH_ ah they use a sacrificial packing to make the clearance holes after threading
23:38 GuShH_ that's a shitload of work.
23:39 GuShH_ lol they call for counterbores... mine are huge.
23:40 Hyratel is there a multiple DIO+ AI/PWMO to I2C bridge (that isn't just an attiny with the requisite code in it)?
23:47 rue_more AI?
23:47 rue_more hobby pwm capture?
23:47 Hyratel whut
23:47 Hyratel AIO
23:48 rue_more digital io, analog in? pwm out?
23:48 Hyratel analog input/pwm out
23:48 Hyratel yeah
23:48 rue_more I have mega32 code, buts its plain serial
23:48 Hyratel I mean i could just roll an ATtiny for it
23:48 rue_more it will do io polling and notification for you too
23:49 Hyratel eh
23:49 rue_more tell you when a bit changes, if you ask it to
23:49 rue_more its not i2c tho
23:49 rue_more its 9600 baud serial
23:49 Hyratel another joystick i have has an 8in i2c input buffer for the butotns on the handle
23:50 Hyratel 3 triggers, a missile button and a hat
23:50 rue_more 8 inch!
23:50 Hyratel 8 inputs
23:50 rue_more ah
23:50 Hyratel 8DI
23:50 rue_more 8dpi is pretty low for a hat switch
23:50 Hyratel Gravis Blackhawk Gameport digital
23:50 Hyratel it's a 4 way hat
23:52 rue_more is it one of the PCF chips?
23:53 Hyratel uhhh i haven't had it open in a year or mor, I couldn't say offhand
23:53 Hyratel you're thinking it might be a pcf8575?
23:54 rue_more been a long time since I laid out some i2c chips and tried to work out what to use them for
23:54 Hyratel how frequently do you think I would have to poll to get reliable keypresses? 4kHz?
23:54 rue_more hmm
23:55 rue_more buttons are evil
23:55 Hyratel that'd be often enough to reliably get 'down' and 'up' events
23:55 rue_more the problem is noise and bounce
23:56 Hyratel hmm
23:56 Hyratel i was thinking "you'd have debounce in the i2c slave code" but that's only possible if it's a programmable device