#robotics Logs
Dec 23 2018
#robotics Calendar
02:06 AM veverak: :D
02:10 AM veverak: hmm
02:10 AM veverak: is there a spaceship at least assembled in 0G?
02:31 AM veverak: lol :)
02:32 AM veverak: I wonder when actuall hype will end
06:55 AM venthe6: Hello. Im planning to build a small indoor robot and I would like to use 2 cameras for SLAM but I have a few concerns. Do you know if stereo vision on its own is enough for obstacle avoidance and ROUGH mapping? What hardware specs in terms of ram and cpu do I need to run it assuming low-resolution images like 640x480. PS Im going to use ROS along with OpenCV for that.
08:47 AM SpeedEvil: venthe6: You can also cheat.
08:47 AM SpeedEvil: https://www.ebay.co.uk/itm/Black-Video-Smartphone-Mobile-Phone-Lens-Stereo-Vision-Camera-Portable-Mini-3D/173462094183
08:47 AM SpeedEvil: This does nice things for you like taking care of the whole exposure/shutter/frame sync issue.
08:48 AM SpeedEvil: Which helps enormously for reducing CPU usage
10:58 AM squirrel1 is now known as veverak
11:03 AM SpeedEvil is now known as Guest17085
11:32 AM rue_mohr: error: incompatible types when returning type 'fn_t {aka struct fn_s}' but 'int (*)(fn_t *) {aka int (*)(struct fn_s *)}' was expected
11:32 AM rue_mohr: ya gotta work hard for that kinda error
11:33 AM veverak: lol
11:38 AM rue_mohr: got it
11:38 AM rue_mohr: ok, I have a library for a stack of functions...
11:38 AM rue_mohr: you know
11:38 AM rue_mohr: saying that sounds funny
11:39 AM rue_mohr: cause, in assembler, you just push the addresses of the functions you want called on the stack, and then do a return
11:39 AM rue_mohr: huh
11:39 AM rue_mohr: C is bloatware!
11:40 AM rue_mohr: so wait, how else would I do this
11:40 AM rue_mohr: it would be trickey, cause all the params are on the stack in C
11:41 AM rue_mohr: < <
11:41 AM rue_mohr: > >
11:41 AM rue_mohr: ...
11:41 AM rue_mohr: int main(void) {
11:41 AM rue_mohr: fnStack_t stk;
11:41 AM rue_mohr: int(*f)(fn_t *);
11:41 AM rue_mohr:
11:41 AM rue_mohr: fnStackInit(&stk);
11:41 AM rue_mohr: fnStackPush(&stk, f1);
11:41 AM rue_mohr: fnStackPush(&stk, f2);
11:41 AM rue_mohr: fnStackPush(&stk, f3);
11:41 AM rue_mohr:
11:41 AM rue_mohr: while( (f = fnStackPop(&stk)) != NULL ) /*printf("%X\n", f); */ f(NULL);
11:41 AM rue_mohr: return 0;
11:41 AM rue_mohr: }
11:43 AM rue_mohr: iirc, the trick is to set it up so that each called fn can modify the stack
11:43 AM rue_mohr: and then you set it up with levels
11:43 AM rue_mohr: so a called fn can reset the stack to a particular level
11:44 AM rue_mohr: like when an error occurs and you need to abort the process and try something else
11:44 AM rue_mohr: but just this part of the process
11:44 AM squirrel1: bought new case for homeserver
11:44 AM squirrel1: the power button is somewhere else than it should be
11:44 AM squirrel1: I shall not put my leg there anymore
11:45 AM rue_mohr: oh no
11:45 AM rue_mohr: thats never good
11:45 AM squirrel1: which part?
11:46 AM rue_mohr: most of the time I used something like that, there would be processes and subprocesses
11:46 AM rue_mohr: the subprocesses could rewrite
11:46 AM rue_mohr: you would anticipate everything going well, and it if didn't you would rewrite the function stack
11:47 AM rue_mohr: er, well part of it
11:48 AM rue_mohr: I should go make sure the new 3d printer didn't catch fire
11:49 AM squirrel1: lol
11:50 AM rue_mohr: it doesn't have a heated bed, so the odds are low
01:18 PM rue_mohr: ok, well
01:19 PM rue_mohr: I should use 3 of the printers to give the first one some upgrades
02:00 PM BitEvil is now known as SpeedEvil
02:45 PM rue_mohr: paraffin melts at about 37c...
02:46 PM rue_mohr: chocolate is about 30c
02:46 PM rue_mohr: so, if I have an extruder that can do about 40c, I'm ok
02:46 PM rue_mohr: except sugar
02:46 PM rue_mohr: oh, right
02:47 PM rue_mohr: about 100c
02:47 PM rue_mohr: wait, no
02:49 PM rue_mohr: 167c
02:49 PM rue_mohr: hmm
02:49 PM rue_mohr: ok, so
02:50 PM rue_mohr: pla at 185c
02:50 PM rue_mohr: hmm
02:50 PM rue_mohr: well, the isomalt isn't going into a plastic syringe then
02:51 PM rue_mohr: so, I want to make the whole head swappable
02:52 PM rue_mohr: subframe and connectors
03:27 PM squirrel1 is now known as veverak
03:51 PM Tom_L: just put a temp sensor on the head and regulate it
04:01 PM rue_mohr: the sugar is a different leauge
04:01 PM rue_mohr: everyone else is 25-40c
06:04 PM rue_mohr: ... could linkedin or facebook be used to deliver packages?
06:04 PM rue_mohr: aka, I know a person who knows a person who knows a person who can give you this package?
06:05 PM veverak: hmm
06:05 PM veverak: maybe?
06:09 PM rue_mohr: routing: Jessica, please give to Tom, Tom, please give to Fred, Fred, please give to Pete, Pete please give to Angela, thanks.
06:12 PM rue_mohr: I suppose, if I'm gonna make a chololate printer, I'd be wise to put some in theis syringe and try it out by hand
06:15 PM rue_mohr: and can pla take 40 without going all flimsy
06:40 PM rue_mohr: I also, it seems, need to desgin a paste extruder
06:43 PM veverak: well
06:43 PM veverak: it compiles
06:43 PM veverak: :)
06:49 PM rue_mohr: heh, and immediatly segfaults?
06:49 PM rue_mohr: well, ok, at 40c, PLA has no appreciative weakening
06:49 PM veverak: nope, in process of finding the bug
06:55 PM flav0r: i found a bug today !
06:55 PM flav0r: a bug in my brain !
06:55 PM veverak: ok
06:55 PM veverak: it's a weird one
06:55 PM veverak: but I am on track finally
06:55 PM flav0r: get this
06:56 PM flav0r: i soldered up 3 identical boards
06:56 PM flav0r: however one of the ESP32 modules seemed to be a bit flaky
06:56 PM flav0r: was taking 3 seconds to register an interrupt and some other weirdness
06:56 PM flav0r: so i actually unsioldered and replaced the esp32 module
06:56 PM flav0r: same problem !
06:57 PM flav0r: then i went to trim down the code so i could do some testing … i read the first include which was for RTCLib.h
06:57 PM flav0r: that made my brain remember that one board didnt have a RTC chip (coz i ran out of them)
06:58 PM flav0r: turns out that ewas the problem !! the code inthat lib is doing something that causes the slow down due to missing chip
06:58 PM flav0r: arrrgh !
06:58 PM flav0r: haha
07:03 PM rue_shop3: its kinda looking like I should do an overhaul on my first 3d printer
07:04 PM rue_shop3: flav0r, haha
07:06 PM veverak: IT WORKS
07:06 PM veverak: :)
07:10 PM rue_shop1: zhanx, with a bit of io map tweeking, I think the extra heater drivers can be used for turning on the lights/fans...
07:13 PM veverak: http://squirrel.veverak.org/Screenshot%20from%202018-12-24%2002-09-11.png
07:13 PM veverak: \p/
07:13 PM veverak: the 'transparent gray boxes' marks 'area above walkable surface'
07:16 PM rue_shop1: ok
07:16 PM rue_shop1: is that player/stage?
07:16 PM veverak: no
07:17 PM veverak: just rviz (ROS) visualization of my motion planner
07:27 PM rue_mohr: by 65c, the pla turned to jello
08:05 PM rue_mohr: ok, that was a quick lesson on how NOT to load chocolate into a syringe
08:06 PM * rue_mohr licks his fingers
08:15 PM rue_mohr: ok well
08:15 PM rue_mohr: lots of force req'd
08:16 PM rue_mohr: so, it'll be trickey to make sure there isn't a spring problem in the extruder
08:16 PM Tom_L: you just wanted to eat some chocolate
08:17 PM rue_mohr: I got it hoter than I wanted to, 50c
08:17 PM rue_mohr: so its all soupie
08:18 PM Tom_L: screw type syringe extruder driven by a stepper
08:18 PM rue_mohr: but it points out that there is gonna need to be a LOT of force on that
08:18 PM Tom_L: i've got a cookie press that extrudes cookie dough to shapes the same way
08:19 PM rue_mohr: Id say I'm into about 20lbs
08:19 PM rue_mohr: and its too hot
08:19 PM rue_mohr: it'll be thicker..
08:19 PM rue_mohr: iirc I'm going to 36-40c
08:19 PM Tom_L: press is about 3" diameter
08:20 PM Tom_L: with 'dies' that fit on the end to press out the dough
08:20 PM rue_mohr: ~1:5 handle
08:20 PM rue_mohr: ?
08:21 PM Tom_L: just a screw
08:21 PM rue_mohr: ah
08:21 PM Tom_L: i remember it back as a kid
08:21 PM Tom_L: it's old
08:21 PM rue_mohr: I wonder what the silicone will be like
08:23 PM Tom_L: https://writes4food.com/2012/11/vintage-cookie-press/?doing_wp_cron=1545618055.1967430114746093750000
08:23 PM Tom_L: almost exact
08:26 PM Tom_L: https://i.pinimg.com/originals/21/2d/fa/212dfaf4eb799649f0cf181e194e3203.jpg
08:26 PM Tom_L: we did the xmas trees this year :)
08:27 PM rue_mohr: :) and now you can machine new stencils for it!
08:28 PM Tom_L: last one is more exacting to the one i have except it's all aluminum color
08:28 PM flav0r: just print them !
08:28 PM Tom_L: the black screw handle is the same
08:28 PM rue_mohr: tom didn't make a printer
08:28 PM Tom_L: old tech is hard to beat
08:29 PM rue_mohr: thats just cause it made it to being old
08:29 PM Tom_L: just trying to offer rue ideas on extruders for different materials
08:29 PM rue_mohr: thingiverse has a lot of paste extruders
08:30 PM rue_mohr: but, you know me, have to redesign it
08:30 PM rue_mohr: I dont like the elaborate loading requirements
08:30 PM rue_mohr: you kinda have to take most of them apart to load material
08:30 PM rue_mohr: I want 'slides in from front'
08:31 PM rue_mohr: and there are hints that the designs have problems
08:31 PM rue_mohr: and the people who say to use air, are paying like $600 for a crazy fancy control board that their hacking to use for air, instead of just using a flipflop
08:32 PM rue_mohr: that and dispite all the occurances on the internet "nobody has done it"
09:27 PM rue_mohr: I'm at 30c and its just as extrudable as it was at 50
09:56 PM rue_mohr: starting to get thicker at 28c
10:24 PM rue_mohr: google made a change to image search and I hate it...
10:24 PM rue_mohr: WHY AM I STILL USING GOOGLE
10:50 PM flav0r: because you secretly want to be part of the machine
10:50 PM flav0r: a spoke in the wheel
10:50 PM flav0r: cog in machine
10:50 PM flav0r: a SHEEP !
11:13 PM rue_mohr: BAAAAAAA
11:15 PM rue_mohr: so, the chineese k thermocouple temp meter
11:15 PM rue_mohr: uses a chopping amp, kinda
11:15 PM rue_mohr: they have an NTC for abmient temp comp.
11:15 PM rue_mohr: and they switch the amplifier between the + and - side of the T/C
11:16 PM rue_mohr: it looks like they run a current thru the t/c
11:17 PM rue_mohr: which seems all wrong, but, hey, china.
11:20 PM rue_mohr: ok, somewhere between 23 and 28c is the good print temp
11:20 PM rue_mohr: (at 23, the chocolate went solid)
11:37 PM rue_shop1: I should pull a zhanx on zhanx
11:38 PM rue_shop1: and build my 3d scanner tonight
11:38 PM rue_shop1: I got the whoooole night,and its only 9:30
11:40 PM * rue_shop1 puts the pi on the table
11:40 PM rue_shop1: hmmmmm
11:43 PM rue_shop1: to pull this off
11:43 PM rue_shop1: we need to teach you how to run a stepper motor
11:43 PM rue_shop1: we can make this easy tho
11:44 PM rue_shop1: cause over the last 14 years, things became available that make this easy
11:44 PM rue_shop1: you need to learn step and direction
11:44 PM rue_shop1: but, not realy
11:44 PM rue_shop1: becasue, my little friend, YOU do not need to concern yourself with direction
11:45 PM rue_shop1: but, you will neeed to turn *this* laser on, and off
11:45 PM rue_shop1: that might sound complicated, but its not that bad, you see, its not two things, its really just one
11:45 PM rue_shop1: because, if you dont tell the laser to be on, it will be off
11:45 PM rue_shop1: clever eh?
11:45 PM rue_shop1: ok, lets rework some hardware
11:46 PM * rue_shop1 blows 1/2" of dust off the rotary table
11:49 PM rue_shop1: hmmm, 5 wire stepper, this just got a little more complicated
11:49 PM rue_shop1: with a key ground into the shaft, well damn
11:51 PM rue_shop1: hmm, serves me, I just started taking something I built apart 14 years later, and, low and behold, I'v changed the mothods I use to do everything
11:53 PM rue_shop1: ok lets push the limit, do I have a 400ppr stepper that will fit this...
11:56 PM rue_shop1: ok, but its shaft is too small, looking again...
11:56 PM rue_shop1: same model...
11:59 PM rue_shop1: }:E and the bottom drops out of the crate of steppers.....