Back
[00:09:12] <rue_house> "Yes, tech support?, I'd like random numbers that are more random please?"
[00:24:28] <rue_house> ok, rotate a point, I know how to do this
[00:24:36] <rue_house> cos.....
[00:24:42] <rue_house> x=....
[00:24:46] <rue_house> theta....
[00:24:56] <rue_house> something about radius...
[00:30:01] <rue_house> #include "2d.h"
[00:30:30] <rue_house> result = rotatez(pt(x, y), angle);
[00:30:33] <rue_house> DONE
[00:40:23] <rue_house> CrazyEddy, "SALE!"
[00:40:36] <rue_house> *sigh*
[00:42:58] <rue_house> oooh, wrong keyword
[00:43:00] <rue_house> duh
[03:17:08] <rue_house> hmm, openscad cant handle a few hundred spheres
[03:20:30] <jrowe47> openscad can't handle your balls, eh?
[03:23:36] <rue_house> povray could have done this in less than 4 seconds
[04:20:37] <rue_house> ok, its 126 spheres, openscad has been trying to render it for a half hour now
[04:22:40] <veverak> lol
[04:22:50] <veverak> rue_house: $fn = 1024 is maybe too much don't you think?
[04:22:52] <veverak> :)
[04:25:06] <rue_house> its set to 35 :/
[04:25:26] <rue_house> I really need to write a polygon library
[04:26:16] <rue_house> RAYTRACING is supposed to be the slow method
[04:26:27] <rue_house> THIS is like 3000x slower
[09:02:53] <Tom_itx> rue_house, use solidworks
[11:37:19] <rue_bed> can I script csg in solidworks?
[12:30:00] <rue_house> trying to render a tree I can 3d print
[12:31:14] <rue_house> it took openscad 7 hours to render 126 spheres
[12:32:18] <rue_house> well, 50/50 with some cones
[12:47:49] <z64555> sounds like its rendering engine is too accurate
[12:48:11] <z64555> should be some settings to reduce the polycount
[12:48:49] <rue_shop3> its low
[12:48:57] <rue_shop3> slic3r dosnt like the model either
[12:49:33] <rue_shop3> 10cm high branched tree
[12:49:42] <rue_shop3> that dosn't break the printing rules ;)
[12:56:26] <rue_shop3> slic3r took down the computer!
[12:56:32] <rue_shop3> :) bad hardware
[13:08:50] <rue_shop3> about a day after I changed the motherboard caps, there was an explosion in one of the computers...
[13:09:42] <z64555> generally Not a Good Thing (TM)
[17:43:31] <rue_shop3> hmmm the drivers in my grbl box aren't ideal...
[17:47:11] * DJuego says bye!
[17:47:13] <anniepoo> I'm designing the rest of the hardware for the robot
[17:47:32] <anniepoo> I've got some places where servos don't use their full travel
[17:47:58] <anniepoo> So I'm wondering how to deal with 'wild servo' action at startup
[17:51:40] <z64555> don't, or can't?
[17:51:57] <anniepoo> good question
[17:52:08] <anniepoo> 'it's convenenient'
[17:52:14] <anniepoo> convenient
[17:53:10] <z64555> don't -> servo's are unable to achieve their full travel due to input limited, or mechanical failure
[17:53:57] <z64555> can't -> servos are unable to achieve their full travel, due to the mechanism they're acting on has physical stops, inentional or not. Or otherwise cannot travel in some areas because it would damage the system
[17:54:05] <anniepoo> I mean I need to, for example, wag the tail, and the simplest way to do it is to attach a rod to the servo horn
[17:54:13] <anniepoo> latter case
[17:54:47] <anniepoo> I read your question as 'can you redesign to use full travel' which is a valid question sometimes
[17:54:52] <anniepoo> but not always
[17:55:03] <anniepoo> the tail shouldn't wag 270 deg.
[17:55:25] <anniepoo> and will come against physical stops
[17:55:31] <z64555> alright, then put mechanical stops in the servo and reduce the servo's range centered around 0
[17:55:51] <z64555> or and do +/-135 degrees
[17:55:56] <anniepoo> how reduce the servos range
[17:55:57] <z64555> *and then
[17:56:17] <anniepoo> ok, physically, I can put some stops strong enough the servo can't move them
[17:56:26] <anniepoo> but then won't the servo strip it's gears?
[17:56:41] <z64555> yeah, which is why you need to adjust the controller in it
[17:56:46] <anniepoo> software solution won't help - we're talking about in the middle
[17:56:51] <anniepoo> you can adjust the controller?
[17:57:06] <z64555> or hack in some resistors. :)
[17:57:12] <anniepoo> oh! clever - ok
[17:57:30] <anniepoo> but wait
[17:57:34] <anniepoo> that's the feedback
[17:57:54] <anniepoo> if I put fixed resistors in it reduces the dynamic range of the pot
[17:58:10] <anniepoo> which means if I command full travel it's going to slam over even harder
[17:58:32] <z64555> digital servo?
[17:58:33] <anniepoo> kinda suprised this isn't something with a well known solution?
[17:58:36] <anniepoo> yes, actually
[17:58:41] <anniepoo> well, some of them are
[17:59:36] <z64555> ok, then work on the input signal
[18:00:21] <anniepoo> http://www.modelairplanenews.com/blog/2012/02/09/total-control-the-right-way-to-set-up-servos/
[18:00:29] <z64555> rue should know more about this, as he so likes to let everyone know
[18:01:07] <anniepoo> for example, this RC plane page shows setting up a control surface for 120 degrees of servo motion
[18:01:38] <z64555> Yeah, it's working on mechanics
[18:02:04] <z64555> can do that too
[18:02:35] <anniepoo> ok, but when the system comes up, what keeps it from slewing to an out of bounds position?
[18:03:41] <anniepoo> ah, some of these servos don't slew 270
[18:04:35] <z64555> yup
[18:04:43] <anniepoo> apparently you can send some digital control signals to program a digital servo?
[18:05:17] <anniepoo> I asked on here the other day if I needed a different controller for a digital servo, was assured nope, it's PWM like analog
[18:06:58] <anniepoo> https://www.youtube.com/watch?v=uWLpEZJutZ4
[18:07:00] <z64555> First time I've heard of being able to program a servo
[18:07:11] <z64555> It's doable
[18:07:12] <anniepoo> using a little box
[18:07:26] <z64555> Depends on the servo you're using
[18:07:30] <z64555> um.
[18:07:41] <rue_shop3> the digital servos just remember the last control pulse and do more corection pulses in between
[18:07:42] <z64555> I'd try the resistor hack on one of them as a test
[18:07:57] <z64555> on the wiper
[18:08:04] <rue_shop3> what are you trying to do?
[18:08:13] <z64555> reduce the range a bit
[18:08:14] <anniepoo> https://www.youtube.com/watch?v=ymW0Lwmw14Y
[18:08:20] <rue_shop3> if the positions are out from a avr library , you have to adjust the numbers, not the servo
[18:08:27] <rue_shop3> cant do youtube here
[18:08:30] <rue_shop3> just tell me!
[18:08:47] <z64555> she's trying to prevent the servo from going out of range
[18:09:05] <anniepoo> rue, scroll back, see original question
[18:09:10] <rue_shop3> dont mod the servo, change the number range in the software
[18:09:25] <anniepoo> ok, but how's that help me at startup time?
[18:10:08] <anniepoo> or is the servo magically going to come up in some 'off' mode?
[18:10:15] <rue_shop3> you dont send servo pulses at startup
[18:10:19] <z64555> You'd set the min/max pulse width, essentially. That'll achieve the same results with less manual work
[18:10:22] <rue_shop3> no pulses = off
[18:10:33] <anniepoo> ah, ok
[18:10:42] <z64555> aka less fun
[18:10:43] <anniepoo> I'm starting with this OSEPP kit
[18:10:45] <z64555> :(
[18:10:53] <rue_shop3> regular servos only do a correction pulse when they get a control pulse
[18:11:00] <rue_shop3> digital servos have a timeout
[18:11:07] <anniepoo> ok - that works
[18:11:29] <rue_shop3> http://ruemohr.org/~ircjunk/tutorials/elex/hobbyservo/servo101.html
[18:13:06] <anniepoo> ok, so I should disconnect the push rod or whatever, determine the software min/max rotation settings, and only then connect the pushrod and be good
[18:13:08] <anniepoo> 8cD
[18:13:15] <rue_shop3> yup
[18:13:46] <rue_shop3> my 3d printer is doing a good job on this tree
[18:14:18] <anniepoo> 8cD a tree?
[18:19:09] <anniepoo> the little 9g servos I've got are 180 degrees, suprisingly
[18:20:54] <rue_house> yes, I spent some time last night writing some code to generate openscad 3d trees
[18:21:08] <rue_house> they are only still simple, but take openscad about 7 hours to render
[18:21:28] <anniepoo> ah
[18:22:48] <rue_house> anniepoo, how many servos?
[18:26:08] * z64555 reads over rue's page on servos, for the 5th time
[18:26:33] <rue_house> do suggest things I could rephrase if you like
[18:26:39] <anniepoo> 7
[18:26:52] <rue_house> 7 times? yea, somehing should be rephrased
[18:26:55] <rue_house> er
[18:27:06] <rue_house> anniepoo, what controller?
[18:27:19] <anniepoo> pi servo HAT
[18:27:44] <rue_house> ah glorious, someone found a name other than shield
[18:27:56] <rue_house> its not using that stupid chip is it?
[18:28:02] <anniepoo> 8cD the small ones use 'bonnet'
[18:28:26] <anniepoo> it probably is indeed using that stupid chip. I dunno, I send it I2C and it moves the servos
[18:29:11] <rue_house> just trying to dig up the part number
[18:29:54] <rue_house> not PCA9685 is it?
[18:30:04] <anniepoo> hang on
[18:30:16] <rue_house> they claim 12 bits, but by the time its controlling a servo, you get less than 8 bits
[18:30:32] <anniepoo> yup
[18:30:35] <rue_house> ugh
[18:30:38] <anniepoo> it's a pca9685
[18:30:48] <rue_house> you would be MUCH better off with a mega328 running my software
[18:31:00] <rue_house> I suppose give it a go
[18:31:29] <anniepoo> at the moment it's a low engineering time solution
[18:31:44] <rue_house> if you have a spare arduino pro mini, do try my software
[18:31:56] <anniepoo> ok, that's interesting
[18:32:38] <anniepoo> for right now it's not where I want to put my effort. I have a simple solution for which I have a python library. 8cD
[18:32:51] <rue_house> yep
[18:33:08] <anniepoo> if I don't spend time fiddling with some 'better' solution, I have time to work on the more important stuff
[18:34:15] <rue_house> the only part that might be fiddly with mine for you is that a) its serial, not i2c b) its a bit-packed 2-byte packet, which I can give you C code for, but I have no python for
[18:34:58] <rue_house> but it allows for things like synchronized updates, which can be nice
[18:35:39] <anniepoo> I'll start with what I've got
[18:35:57] <anniepoo> if I have issues, I know I can deal with them
[18:36:00] <anniepoo> 8cD
[18:36:25] <anniepoo> 8 servos, actually - forgot the die cutter for the printer
[18:36:43] <anniepoo> 8cD snaily can 'poop' bits of receipt printer output
[18:36:57] <rue_house> https://www.youtube.com/watch?v=j4uz_yha9uI
[18:37:06] <rue_house> that ones using the m328 code
[18:37:36] <anniepoo> that your robot?
[18:37:49] <rue_house> one of the 4 'brothers'
[18:37:56] <rue_house> er 3...
[18:38:01] <rue_house> 3 or 4...
[18:38:26] <rue_house> you can see one of the other in the background
[18:38:42] <rue_house> er, 2 of them
[18:39:49] <anniepoo> in your 'robot dance' video, how'd you create the animation?
[18:40:14] <rue_house> I ahve an input arm
[18:40:16] <rue_house> well, 2
[18:40:19] <anniepoo> ah
[18:40:35] <rue_house> I can record motions with them
[18:40:47] <rue_house> the marbelator is a set of like 5 recorded motions
[18:41:03] <rue_house> that are program triggered depending on where the marble ended up
[18:41:11] <rue_house> -was a set-
[18:41:23] <anniepoo> I was going to use a blender model, animate it, transfer the bvh file to robot
[18:41:36] <rue_house> Love bvh files
[18:41:54] <rue_house> https://www.youtube.com/watch?v=qZF9WJ1Ueew
[18:41:55] <rue_house> :)
[18:42:27] <anniepoo> when I worked for UH I did a bunch of hacking on them
[18:42:35] <anniepoo> do you know the program bvhacker?
[18:42:45] <anniepoo> it's sweet, lets you do lots of stuff with bvh files
[18:42:53] <anniepoo> and the format's simple
[18:43:02] <rue_house> no, but I wrote a bvh rendering program cause I couldn't find ANYTHING that would play a generic bvh
[18:43:10] <anniepoo> it's a skeleton definition, then a bunch of rows of joint positions
[18:43:19] <rue_house> I know, I wrote a player
[18:43:24] <anniepoo> neat
[18:43:43] <rue_house> https://www.youtube.com/watch?v=TrR8moC3qbA
[18:44:08] <rue_house> the bvh was just downloaded, but I can play a bvh with ANY skelaton
[18:44:22] <rue_house> I been wanting to mod it to make bvh files
[18:44:42] <anniepoo> we made a trainer for kinesiology students
[18:44:45] <rue_house> but bvh's have that little twist in the recursion thats defeated my brain so far
[18:44:57] <anniepoo> a guy, in the virtual world, with a pair of legs
[18:45:18] <anniepoo> by moving controls around on angle-angle diagrams you could alter his gait
[18:45:21] <z64555> So the guy is holding a pair of legs?
[18:45:23] <z64555> ok.
[18:45:30] <anniepoo> no, he's just a body and legs
[18:45:40] <z64555> I was being silly :P
[18:45:40] <rue_house> I was just doing the line rendering, beacuse my goal was to use bvh to operate robots
[18:46:02] * z64555 still thinks about servo hacks
[18:46:20] <anniepoo> then you could push a button and be given a bvh file in world to use with yoru own avatar
[18:46:54] <z64555> pull-down shifts the center, pull up scales up, and the wiper scales down?
[18:47:00] * z64555 thinks a bit
[18:47:01] <rue_house> https://www.youtube.com/watch?v=P75b7vqQWRQ
[18:47:04] <rue_house> thats the other one
[18:47:25] <rue_house> in my program you can zoom and rotate the model
[18:50:48] <z64555> hm, no, the wiper does the same thing as the pull up.
[18:52:18] <z64555> possibly more power efficient, assuming little current flows through the wiper
[18:56:03] <anniepoo> brb, time for another coat of lacquer
[19:02:10] <z64555> pulldown from wiper does nothing
[19:02:15] <z64555> pullup wiper, though...
[19:07:27] <z64555> that might scale down, but it has "wierdness" with it
[19:07:38] <z64555> I'd have to write it down on paper to look at
[20:19:16] <anniepoo> I'm using the guts of a receipt printer to print slips of paper the robot can leave behind
[20:19:57] <anniepoo> the printer doesn't have a cutter built in. Anybody have a clever idea for how to make the cutter?
[20:24:44] <Jak_o_Shadows1> how wide is the receipt?
[20:24:51] <anniepoo> 2.25 inch
[20:25:48] <z64555> servo-actuated scissors :D
[20:26:17] <anniepoo> yah, was thinking of that. Wondered if a reasonable sized servo would have the oomph
[20:27:14] <anniepoo> thought about a rotary cutter blade attached to servo horn, roll along and move across
[20:29:08] <anniepoo> hang on, will find a reciept, scissors, and scale
[20:32:07] <anniepoo> takes about 600 grams on the end of the blade
[20:32:45] <anniepoo> for a medium pair of scissors, receipt halfway up
[20:33:42] <anniepoo> maybe I can drive it with the motor
[20:33:49] <anniepoo> (the tread motors)
[20:35:25] <anniepoo> drop a 'foot' of some sort and then drive forward
[20:41:46] <Jak_o_Shadows1> What about moving a razor blade?
[20:41:51] <Jak_o_Shadows1> easier to replace as they blunt
[20:45:58] <anniepoo> might work
[20:46:48] <anniepoo> coming straight down as chopper means all the force applied at once
[20:47:11] <anniepoo> rotary cutter blades are easily replaced
[20:48:09] <anniepoo> if I pre-perforated the tape I could 'step on' the printed part
[20:48:14] <anniepoo> and drive forward
[20:48:27] <anniepoo> I have a laser, doing the perfs is easy
[20:49:02] <anniepoo> could I make a small gap and use a flyback transformer and make sparks?
[20:50:15] <anniepoo> I already have a vacuum pump aboard. An additional air valve would let me use a vacuum diaphram. Thatd generate a large force
[20:51:15] <z64555> err.
[20:51:31] <z64555> I think Jak_o_Shadows1 was referring to taping a razor blade onto a servo horn
[20:51:37] <anniepoo> I don't need much motion if I use a 'chopper' action. Nitinol wire or a servo drving an eccentric roler?
[20:52:05] <anniepoo> just spinning with servo might work!
[20:52:07] <anniepoo> 8cD
[20:52:56] <anniepoo> I can reprofile the blade so it's cutting against one side, not in center, and have a hardened steel bar
[20:52:56] <z64555> you'll need some sort of anvil for the blade to shear against
[20:53:06] <anniepoo> yes.
[20:53:20] <z64555> maybe use another dulled razor blade
[20:53:33] <anniepoo> I can make that. We have machine shop
[20:53:55] <anniepoo> I think Jak's method's best of these - sounds simple
[20:54:04] <Jak_o_Shadows1> Probably don't need an anvil if you have the blade at an angle, and just glide it
[20:54:06] <Jak_o_Shadows1> maybe
[20:56:26] <anniepoo> I can use a tissue slicing blade -
[20:59:16] <anniepoo> have 4 strips, in pairs above and below paper on each side of cut, swing blade through
[20:59:33] <anniepoo> might need 55g servo, but should wokr
[20:59:56] <rue_house> how about a push-stamp thats wide enough to cut it
[21:00:22] <rue_house> or a blade-against-rubber-pad guillotine
[21:00:23] <anniepoo> that's the same as 'chopper'?
[21:00:40] <rue_house> you know the shape stamps for craft stuff?
[21:00:44] <anniepoo> yes, but then the cutting occurs at same time all along
[21:00:50] <anniepoo> yes, and thought about those
[21:00:55] <rue_house> k
[21:01:14] <anniepoo> it'd be nice to have the paper cut to oval or something
[21:02:18] <anniepoo> crafters using die cut dies (what you're talking about) use a press that just has a roller
[21:02:31] <anniepoo> Oh, hm...
[21:03:04] <anniepoo> leave blade stationary, roll a roller across the top of it
[21:03:12] <anniepoo> hard rubber or wood or something
[21:03:43] <Jak_o_Shadows1> That's a fair bit of force
[21:04:09] <rue_house> https://www.aliexpress.com/item/free-shipping-3-inch-about-7-5cm-Circle-shape-of-eva-punch-craft-punch-DIY-paper/32780748934.html
[21:04:16] <rue_house> sorry I cant find a good one
[21:04:28] <anniepoo> oh, hmm
[21:05:13] <anniepoo> well, scissors are a somewhat similar thing
[21:05:17] <anniepoo> they're a sort of 'die'
[21:05:25] <rue_house> yea
[21:05:33] <rue_house> with a wide enough one...
[21:05:39] <anniepoo> ok, this is slightly crazy
[21:05:52] <anniepoo> make a roller with a spiral ridge
[21:06:00] <anniepoo> turned by servo
[21:06:04] <rue_house> oh, one of those is 3"
[21:06:19] <rue_house> that pins the paper against another roller
[21:06:27] <anniepoo> the roller is on one side of gap, fixed blade on other
[21:06:28] <rue_house> so the robot dosn't ahve to be heavy
[21:06:50] <anniepoo> rue, you're responding to?
[21:07:00] <rue_house> or.. how about two rubber rollers, except one of them has a blade embedded in it?
[21:07:17] <anniepoo> yes, that works
[21:08:00] <rue_house> and a 3d printed child guard
[21:08:01] <anniepoo> hard to make that blade in rubber, but we have a 4 axis mill, we could put the blade into steel roller
[21:08:36] <rue_house> no, just slot a rubber roller from a laser printer and put a break-away knife int eh slot, and glue
[21:08:38] <anniepoo> ironically, for all the machinery around here, one thing we don't have is a 3d printer
[21:08:49] <rue_house> ah
[21:08:59] <anniepoo> rue, you have to twist the knife
[21:09:03] <rue_house> sorry, I kinda assumed you were working on the kitchen table
[21:09:11] <rue_house> it can be all-at-once
[21:09:16] <anniepoo> http://theelginworks.com/
[21:09:19] <anniepoo> working there
[21:09:47] <anniepoo> Oh, I get you
[21:10:00] <anniepoo> only problem is then you're dooing all the cutting at once
[21:10:06] <anniepoo> Oooh... that's fixable
[21:10:12] <anniepoo> make stairsteps
[21:10:12] <rue_house> you know, the excavator dosn't really fit in that list properly
[21:10:17] <anniepoo> LOL
[21:11:14] <anniepoo> 8cD we're a couple insane engineers in the middle of nowhere
[21:11:34] <anniepoo> ok, rue - use your roller with blade in it,
[21:11:41] <rue_house> hey!, I'm in the middle of nowhere too!
[21:12:11] <rue_house> part of the reason I suggest the roller, is that I have a LARGE box of rubber rollers from printers
[21:12:15] <anniepoo> but only 1/2" of blade, then next section is 10 degrees away
[21:12:23] <rue_house> I ahve no idea what to use them for
[21:12:26] <anniepoo> so the sections of blade are straight
[21:12:38] <rue_house> but cutting the rubber off to get to the metal bar inside is not worth it
[21:12:47] <anniepoo> Oooh, yes, those would work well
[21:13:07] <anniepoo> Hmm... we just pitched a printer, but sadly didn't salvage those
[21:13:23] <rue_house> heh, I love how you work out what I'm talking about 10 sentences after :)
[21:13:30] <anniepoo> lol
[21:13:40] <Jak_o_Shadows1> With our excavator we can make holes in the ground.
[21:13:43] <Jak_o_Shadows1> Good.
[21:13:43] <rue_house> anniepoo, door to door! there are printers in closets people WANT to get rid of!
[21:13:44] <anniepoo> lol
[21:13:51] <anniepoo> true
[21:14:06] <rue_house> I have 9 lasers on the deck
[21:14:15] <rue_house> one of them has a really promosing looking touchscreen!
[21:14:37] <anniepoo> well, your last iteration was just a slit radial to the axis of the rubber rod
[21:14:46] <rue_house> anniepoo, ever made an imperial/metric drill guage?
[21:15:07] <rue_house> its only 2" it wouldn't take much force to cut paper
[21:15:12] <anniepoo> and I'm suggesting a riff - make the blade in short sections
[21:15:23] <rue_house> yup, I get ya
[21:15:26] <anniepoo> well, it's 600 grams to push scissors down , we tried that
[21:15:48] <anniepoo> I think hardwood would be a good choice for rollers
[21:15:52] <rue_house> but if your pinching between rollers its easy to gear up that much force
[21:15:58] <anniepoo> right
[21:16:29] <anniepoo> yes, I was thinking servo drive, but guess I can find some small metal gears in my junk box
[21:16:54] <anniepoo> I like that solution
[21:16:59] <rue_house> :| if you could only see all the gears I have...
[21:17:15] <rue_house> you saw the box back 17 years ago?
[21:17:20] <anniepoo> I think I have a worm gear on a cheap moter?
[21:17:24] <anniepoo> I think I have a worm gear on a cheap motor
[21:17:36] <Jak_o_Shadows1> I dunno about hardwood. Don't most die cutters like that, the small ones with interchangeable blades, have hard plastic sheets on both sides of the die? The roller itself is not all that hard
[21:17:55] <rue_house> http://ruemohr.org/~ircjunk/tutorials/mech/gearbox/dscn7580.jpg
[21:18:02] <rue_house> imagine that * 8
[21:18:33] <rue_house> I had to split it up
[21:19:12] <anniepoo> I could make tghe rollers from UHMW or something
[21:19:17] <rue_house> crown gears, 90 degree teeth, metal, plastic...
[21:19:18] <anniepoo> that's what cutting boards are
[21:19:38] <rue_house> use a layer of self-healing mat
[21:19:50] <TheBarKeep> hello
[21:19:56] <Jak_o_Shadows1> Did you split it up, and sort it?
[21:19:57] <rue_house> meo
[21:19:59] <rue_house> moe
[21:20:08] <rue_house> Jak_o_Shadows, yea, I had to
[21:20:25] <rue_house> rubbermaid tubs that big dont fit under the bench
[21:20:31] <TheBarKeep> what do you guys think is an unreasonable amount of backlash on a robot?
[21:20:43] <TheBarKeep> industrial
[21:20:53] <rue_house> brain surgery or rocket science robot?
[21:21:14] <rue_house> or marble handling?
[21:21:18] <TheBarKeep> industrial robot arm, like how many arc-min would you find unsettling
[21:21:27] <TheBarKeep> sure marble handling
[21:21:27] <rue_house> backlash?
[21:21:43] <TheBarKeep> yeah
[21:21:44] <anniepoo> the roller solution has another merit
[21:21:49] <rue_house> https://www.youtube.com/watch?v=j4uz_yha9uI
[21:21:54] <rue_house> like that?
[21:22:04] <rue_house> if so, about 1mm is ok
[21:22:08] <anniepoo> I can use the same motion to drive getting the paper out of the rear of the robot
[21:22:33] <TheBarKeep> 1 mm where?
[21:22:57] <TheBarKeep> you mean net uncertainty at the end effector?
[21:23:24] <TheBarKeep> 1 mm seems rather tight to me, personally
[21:23:41] <anniepoo> ok, self healing mats are PVC coated fabric
[21:23:45] <TheBarKeep> (cool robot by the way)
[21:24:14] <rue_house> thanks
[21:24:16] <TheBarKeep> is that yours Rue_house?
[21:24:28] <anniepoo> but UHMW is what food handling uses for antimicrobial and cut resistant
[21:24:38] <anniepoo> it's the white cutting boards, slightly waxy feel
[21:25:20] <TheBarKeep> Also, in that robot, the vibration appears to dominate way more than the backlash
[21:25:46] <TheBarKeep> for tool tip position uncertainty
[21:26:16] <rue_house> its one of the set
[21:26:32] <TheBarKeep> set for what?
[21:26:50] <rue_house> just world domination :/
[21:27:00] <TheBarKeep> haha
[21:30:07] <rue_house> I was working on a 3d printable one, but also trying to go larger
[21:30:36] <TheBarKeep> oh
[21:30:41] <TheBarKeep> that's cook
[21:30:43] <TheBarKeep> cool
[21:31:19] <TheBarKeep> I've seen a 3D printed 3 axis, plus gripper using Dynamixels
[21:31:22] <TheBarKeep> that was cool
[21:31:41] <anniepoo> back in a bit, my lacquer's probably dry
[21:52:13] <anniepoo> cutting 8cD
[22:06:49] * BeautiCode waves to anniepoo
[22:08:01] <anniepoo> howdy BeautiCode
[22:08:34] <BeautiCode> hey
[22:08:37] <BeautiCode> how are you?
[22:39:15] <anniepoo> hey, sorry,
[22:39:20] <anniepoo> machine called me away
[22:39:49] <anniepoo> I'm doing fine
[22:41:43] <anniepoo> doing lots of work on the snail
[22:43:02] <anniepoo> was busy as heck, now I'm relaxing a bit.
[22:43:19] <anniepoo> so there's time for some fun
[22:44:54] <anniepoo> ok, one more design puzzle
[22:45:02] <anniepoo> I will often be using the snail on a table top
[22:45:13] <anniepoo> I'd like to detect the table edge
[22:46:57] <BeautiCode> ah no worries
[22:47:10] <BeautiCode> the snail?
[22:47:13] <anniepoo> yes
[22:48:27] <anniepoo> maybe a small IR laser and IR photodiode mounted so they meet at table height 18 inches ahead of the snail?
[22:48:43] <anniepoo> 8c( Hiding apertures for sensors is becoming a major issue
[22:51:37] <BeautiCode> sounds like a good plan
[22:51:48] <BeautiCode> what is the purpose of the snail?
[22:52:30] <anniepoo> to educate children
[22:52:34] <anniepoo> 8cD
[22:53:19] <Snert__> cherry makes a switch with a wire actuator...really gentle spring. Mount it on bottom of smail.
[22:53:37] <anniepoo> couple issues with that
[22:53:57] <anniepoo> 1) short range
[22:54:07] <BeautiCode> ah xD
[22:54:55] <anniepoo> 2) the robot turns on it's treads, whose front end is a good 8" aft of the front of the robot, where it'd be mounted
[22:55:17] <anniepoo> if the robot turns, the sensor needs to caster quite a bit
[22:55:29] <anniepoo> and if it's sticking forward that's going to be hard
[22:55:38] <anniepoo> (guess I could give it a servo)
[23:04:53] <anniepoo> but that sure looks more reliable than some range sensor thing
[23:09:14] <anniepoo> Hmm....
[23:09:26] <anniepoo> maybe a belt and suspenders approach would work
[23:10:01] <anniepoo> use the range sensor to locate the edges for social reasons
[23:10:33] <anniepoo> and the feeler to back up the range sensor in case it fails
[23:11:23] <anniepoo> Oh - doh - shine IR LED down and just a little forward (not 18") and look for reflection in IR detector.
[23:12:06] <anniepoo> ok, looks like I'm going to experiment with IR LED's and lenses
[23:12:08] <Snert__> that might depend on the reflective surface too.
[23:14:27] <anniepoo> if I gate the LED it's going to have some reflection and see AC signal pretty much on anything but a deep IR black
[23:14:41] <anniepoo> but yup, that might be an issue
[23:15:32] <anniepoo> think I'll make a little prototype
[23:19:57] <Snert__> carpet would be vastly different from table top
[23:20:08] <Snert__> maybe it never will be on carpet though.
[23:22:50] <anniepoo> it will be on table top, carpet, linoleum, and maybe transition from one to the other
[23:23:50] <anniepoo> brb, time to lacquer
[23:23:57] <Snert__> lots of challenges :)
[23:28:33] <anniepoo> and the acoustic sensor might be confused by carpet
[23:28:46] <anniepoo> brb, gotta do a shop thing
