#robotics | Logs for 2017-01-07

Back
[00:57:25] <rue_shop3> hmm, these cordless drill motors are sensitive about the freq of pwm used on them
[01:20:51] <rue_shop3> flipping!, there are 20A spikes comming off this motor, UNLOADED!
[01:22:32] <rue_shop3> irfZ44 @15A
[01:50:23] <rue_shop3> ok, the 4110 didn't help much, the motor is pulling 25A with a bit of load, and the battery I was using as a supply just died
[01:50:27] <rue_shop3> HMMMMMMMMMMMMM
[01:51:09] <rue_shop3> so I need to make a really insane motor driver
[02:48:39] <rue_shop3> it'd take like 6 50W 12V halogens in parallel to mimmic that load
[02:50:59] <SpeedEvil> Well, yes and no.
[02:51:29] <rue_shop3> the peak load
[02:51:39] <SpeedEvil> The initial current of 6 50W halogens would be 2100W
[02:51:45] <rue_shop3> without the induction
[02:51:56] <rue_shop3> thats a good point too
[02:55:23] <rue_shop3> I love it when your trying to find information on something, and the company name is so generic its impossable to isolate, and so is the part number
[02:55:30] <rue_shop3> Camtech CD-1
[03:43:11] <rue_shop3> but this is funny, the chip their using has SUCH a specific package, there is prolly only one chip that uses it
[06:20:49] <SpeedEvil> But if that's the case, it's probably cyustom anywya
[06:30:51] <rue_house> found it, its a custom package for a A1393
[06:31:05] <rue_house> but its a A1393, which is the important part
[06:31:26] <rue_house> sweeeeeeeeeeeeeeeeeeeet I got the multiplication code gen working
[06:31:57] <rue_house> to multiply 16 bits by 0.9:
[06:31:59] <rue_house> o += i; // Add dividend to accumulator (dividend /1.000000)
[06:31:59] <rue_house> i >>= 1; //Shift dividend right. Shift#1
[06:31:59] <rue_house> i >>= 1; //Shift dividend right. Shift#2
[06:31:59] <rue_house> i >>= 1; //Shift dividend right. Shift#3
[06:31:59] <rue_house> o -= i; // Subtract dividend from accumulator (dividend /8.000000)
[06:32:00] <rue_house> i >>= 1; //Shift dividend right. Shift#4
[06:32:02] <rue_house> i >>= 1; //Shift dividend right. Shift#5
[06:32:04] <rue_house> o += i; // Add dividend to accumulator (dividend /32.000000)
[06:32:06] <rue_house> i >>= 1; //Shift dividend right. Shift#6
[06:32:08] <rue_house> i >>= 1; //Shift dividend right. Shift#7
[06:32:10] <rue_house> o -= i; // Subtract dividend from accumulator (dividend /128.000000)
[06:32:12] <rue_house> i >>= 1; //Shift dividend right. Shift#8
[06:32:16] <rue_house> i >>= 1; //Shift dividend right. Shift#9
[06:32:18] <rue_house> o += i; // Add dividend to accumulator (dividend /512.000000)
[06:32:20] <rue_house> i >>= 1; //Shift dividend right. Shift#10
[06:32:22] <rue_house> i >>= 1; //Shift dividend right. Shift#11
[06:32:24] <rue_house> o -= i; // Subtract dividend from accumulator (dividend /2048.000000)
[06:32:26] <rue_house> i >>= 1; //Shift dividend right. Shift#12
[06:32:28] <rue_house> i >>= 1; //Shift dividend right. Shift#13
[06:32:30] <rue_house> o += i; // Add dividend to accumulator (dividend /8192.000000)
[06:32:32] <rue_house> the error is really really small
[06:32:38] <rue_house> have a program that generates the code
[06:32:42] <veverak> lal
[06:32:45] <veverak> :)
[06:33:00] <rue_house> 65507 *0.9 = 58958 should be 58956.300000 error -0.000026
[06:33:06] <rue_house> error is against the multiplier
[06:33:30] <veverak> sounds good enough
[06:36:45] <Ad0> is ROS any good?
[06:37:42] <rue_house> no just write code that does what you want ;)
[06:37:58] <rue_house> its 4am, have the robot get some sleep for me
[06:39:16] <Ad0> hehe
[06:39:33] <rue_house> well _I_ dont want to do it, I'm having fun
[06:39:51] <rue_house> which is getting challanging thru the foggy vision and the woozy head
[06:40:24] <rue_house> oh, you want the bits of precision 1 more than the bits your working with
[06:47:09] <Ad0> hehe
[06:48:33] <rue_house> ok, the software works, I have a routine to multiply a 16 bit number by 0.9 on a microcontroller
[06:48:42] <rue_house> http://www.piclist.com/techref/method/math/divconst.htm
[06:56:16] <veverak> Ad0: ros is good, depends on what
[06:57:24] <Ad0> AGV
[06:57:51] <Ad0> trying to think out an extremely low profile AGV
[06:57:59] <Ad0> which rules out LIDAR I guess
[06:58:50] <veverak> AGV?
[06:59:09] <Ad0> Automated guided vehicle
[07:00:13] <Ad0> you know like those amazon bots etc
[07:00:24] <veverak> Ad0: point with ros is that it creates certain knowledge overload you have to go through
[07:00:26] <Ad0> they "cheat" a lot though, using a grid of tags
[07:00:35] <veverak> and you have to learn how to make proper ros packages with all things they need
[07:00:44] <Ad0> yeah I have looked at it
[07:00:50] <veverak> which takes some time
[07:00:57] <Ad0> ok
[07:01:02] <veverak> on the other hand, if you learn it, you got access to a tons of allready written modules/stuff
[07:01:13] <veverak> kalman filters for positiiniong, lidar, slam
[07:06:50] <veverak> but, if you don't intent to use that stuff, it's easier to write it without it
[08:33:58] <Ad0> aye
[08:33:59] <Ad0> thanks
[09:52:48] <Ad0> omniwheel looks fun but I guess they aren't really powerful
[10:10:40] <z64555> navy forklift trucks have omniwheels
[10:26:40] <malcom2073> They're as powerful as a solid wheel of equivalant surface contact
[10:26:58] <malcom2073> (Which is usually not that significantly smaller than one of equivalant size)
[12:18:01] <lonecrow> rue_house don't forget MakerFaire meeeting 11am at gum boot cafee. today!!
[12:45:23] <rue_house> gum boot
[12:45:28] <rue_house> I know where that is
[12:45:35] <rue_house> what time is it
[12:45:41] <rue_house> ok, breakfast..
[12:47:04] <Tom_L> u gonna be late
[12:53:46] <rue_house> how did making breakfast take 15 mins?
[12:53:52] * rue_house looks down at his cerial
[12:54:14] <Tom_L> u runnin at 300 baud today?
[12:54:20] <rue_house> I would eat faster but I'd betterfinish my breakfast first
[12:54:38] <rue_house> no, the system clock is apparently double the advertized rate
[12:55:39] <rue_house> :) I have a program that generates integer fractional multiplication code :) what a nice thing to wake up to
[12:56:22] <rue_house> gumboot, hmmm, 5 mins should be enough
[12:56:34] <rue_house> maybe 7
[12:57:31] <rue_house> hmm, with this I could prolly write a PID filter
[13:01:19] <rue_house> the main problem I have with the speed controller code is wrapping my head around the output bias
[13:02:03] <rue_house> I suppose if I'm rebiasing it, the code should look something like pwm = 128 + error;
[13:03:03] <rue_house> that should make the 0 level of the output 6krpm
[13:03:14] <rue_house> I think those motors are 1000kv
[13:03:53] <rue_house> the operation is insane, with a bit of load I'm seeing the current hitting 25A
[13:04:07] <rue_house> this is under 12V
[13:04:20] <rue_house> 12V supply, about 10% duty
[13:04:54] * rue_house ties ec\ low
[13:05:55] <rue_house> 1kv
[13:06:03] <rue_house> 1000 rpm/volt
[13:06:54] <rue_house> :53 is when i need to definitly go
[13:07:10] <rue_house> I'm dressed, thats a good start
[13:07:50] <rue_house> it might be interesting to implement a code based rc filter
[13:08:17] <rue_house> you do it by the charge on the cap
[13:08:49] <rue_house> a drain resistor is just scaling q at regular time intervals
[13:09:06] <rue_house> a resistor is adding constant values on a regular time interval
[14:19:47] <fishbulb> who has some finished robot projects they can post
[14:42:18] <deshipu> fishbulb: what are you looking fro?
[14:42:19] <deshipu> for
[14:42:47] <fishbulb> something awesome?
[14:43:11] <fishbulb> I'm waiting for parts for my 3d printer to turn up so I can make an autonomous outdoors... thing
[14:43:30] <deshipu> I have this http://tote.rtfd.org
[14:43:39] <deshipu> made a fifth version just recently
[14:44:46] <fishbulb> are those small plane servos?
[14:45:08] <deshipu> yes, sg90, the cheapest 9g servos out there
[14:45:18] <fishbulb> haha I think I bought like 15 of them
[14:45:26] <deshipu> I was trying to make it as affordable as possible
[14:45:31] <fishbulb> like purple and translucent
[14:45:38] <deshipu> you only need 12 for this robot
[14:45:49] <deshipu> oh, purple would look nice with a PCB from OSHPark
[14:45:53] <deshipu> where did you get them?
[14:45:58] <fishbulb> I dunno, ebay
[14:46:08] <fishbulb> some special, I think they're all purple and seethrough
[14:46:26] <fishbulb> I haven't looked at them for ages, I got a bunch of metal gear big servos for roughly the same price
[14:46:46] <fishbulb> I haven't used any yet because quads/drones don't need them
[14:47:09] <deshipu> well, this is a quad :P
[14:49:33] <deshipu> I've recently also re-built a biped: https://hackaday.io/project/3736-bob-biped-robot
[14:49:46] <deshipu> it's much simpler
[15:10:21] <ace4016> hehe, i remember those little wind up toys that used the same motion
[15:12:28] <deshipu> they actually used a little bit different technique -- they had small "forks" on their feet, so they didn't have to wobble side to side to keep balance
[15:16:39] <fishbulb> I meant flying ones
[15:17:32] <fishbulb> is that your one on hackaday deshipu?
[15:17:58] <veverak> that hat makes it really qute
[15:19:21] <fishbulb> is the duck driving
[15:19:57] <deshipu> fishbulb: nah, the duck was only posing for a photo
[15:20:00] <deshipu> yes, it's mine
[15:20:05] <fishbulb> I have to wait for 3d printer parts before I can start the drivetrain of a something that can drive around outside
[15:20:30] <deshipu> I thought 3d printing is not too good for stuff like cogs
[15:20:57] <fishbulb> I'll use bearings and probably hobbyking cogs
[15:21:29] <fishbulb> but the RC "offroader" stuff is mega, mega expensive, and it would be stupid to base an autonomous wheeled thing on that
[15:22:15] <fishbulb> motor on wheel would probably be better
[15:22:37] <fishbulb> for steering
[15:22:47] <deshipu> get some of those segway-like toys that tend to burst into flames
[15:22:59] <deshipu> they have motors in the wheels
[15:23:22] <fishbulb> "wanted: one of those segway things that has burst into flames"
[15:23:31] <fishbulb> "I only want the wheels"
[15:23:49] <fishbulb> do you know how much those cost? the first iterations were like 1000 usd or something
[15:24:16] <deshipu> not segway
[15:24:19] <fishbulb> they're still hundreds and hundreds of dollars
[15:24:21] <deshipu> the chinese stuff
[15:24:29] <fishbulb> yeah the stupid "hoverboard" things that you stand on and they roll around
[15:24:47] <fishbulb> they're still very expensive and I think they might have even banned the imports into australia becuase of all those fires
[15:25:12] <fishbulb> I heard they banned them somewhere, it wouldn't surprise me if this retarded country did it
[15:25:21] <fishbulb> bandwagon backlash
[15:25:25] <deshipu> https://www.aliexpress.com/item/11/32759919140.html
[15:25:27] <deshipu> those
[15:25:55] <deshipu> you also get the battery, and the h-bridge, and the voltage regulator
[15:25:58] <fishbulb> 194 a piece
[15:25:59] <deshipu> a pretty good deal
[15:26:33] <fishbulb> that's more than it cost to build a good 3d printer
[15:26:48] <deshipu> seriously?
[15:26:52] <fishbulb> yes
[15:26:56] <deshipu> good luck
[15:27:07] <deshipu> and then you can 3d print yourself a motor and a battery
[15:27:34] <fishbulb> 160 was the parts for the 3d printer
[15:27:47] <deshipu> prusa is about 1000 if you are willing to do most work yourself
[15:28:01] <fishbulb> it's an open source design, all the parts cost 160
[15:28:05] <deshipu> and I wouldn't call that decent
[15:28:32] <fishbulb> I would and so do some professionals
[15:28:48] <fishbulb> what's with the arguing
[15:29:41] <deshipu> I just think you have a bit unrealistic expectations
[15:29:44] <fishbulb> I have better motors than those hoverboard toys but they're for a heavy lifting drone
[15:30:27] <fishbulb> and you base your observation on:
[15:30:32] <deshipu> there are some acceptable 3d printers appearing around at ~$200, but that's only recently, and they still require some modding to work well
[15:30:53] <fishbulb> of course you speak from experience having owned one?
[15:31:12] <deshipu> I speak from experience having listened to complaints and helped fix a lot of them
[15:31:43] <fishbulb> it cost 200 AUD for the parts for a prusa clone
[15:31:44] <deshipu> but I don't mind if someone on the Internet is wrong
[15:31:47] <fishbulb> that's like 160 USD
[15:31:57] <fishbulb> heh
[15:32:07] <fishbulb> you're not worth talking to
[15:33:40] <deshipu> the feeling is mutual
[15:36:29] <deshipu> fishbulb: so, do you have any finished robots?
[15:38:00] <veverak> finished robot
[15:38:02] <veverak> such a thing exists?
[15:38:15] <deshipu> sure, when you no longer feel like working on it
[15:38:39] <fishbulb> it's a like an exponential curve that approaches 1. yeah, but we just reached a mutual understanding
[15:39:58] <deshipu> I'm trying to change the subject to one on which we don't disagree
[15:47:32] <fishbulb> there was a flash sale on that anet a8, it's a clone of the prusa with shittier wiring and connectors (and powersupply) the first thing you print on them are upgrades for them
[15:48:18] <fishbulb> but unless you're moving a big thing on that y axis I dunno if the rigidity mods are necessary
[15:50:24] <fishbulb> does a drone that has a level of autonomy count as a robot? I think it does
[15:51:07] <deshipu> depends on what kind of autonomy
[15:51:18] <deshipu> generally, there is no agreement on the definition of "robot"
[15:51:58] <fishbulb> I think some kind of sensor(s), some kind of autonomy, and some kind of motion
[15:52:09] <deshipu> but people tend to agree that a machine that simulates some aspects of a human or an animal is a robot
[15:52:19] <fishbulb> a thing that reacts to stimuli and performs an action
[15:52:34] <z64555> that's still vague
[15:52:42] <z64555> "reacts to stimuli"
[15:52:49] <deshipu> for instance, chatter bots, while they don't really have sensors or actuators, are still "bots"
[15:52:59] <deshipu> because they try to pretend to be human
[15:53:00] <fishbulb> that's pure software
[15:53:06] <fishbulb> there's no hardware element
[15:53:12] <fishbulb> it woud have to move hardware in my opinion
[15:53:43] <deshipu> on the other hand, a washing machine, while formally having a lot of autonomy and automation in it, including sensors, feedback loops and actuators, is not considered a robot
[15:53:55] <deshipu> because it doesn't wash like humans do
[15:54:10] <z64555> according to deshipu's generalization, chat bots are robots as they emulate animal/human behavior, in the specific case of text communication
[15:54:26] <fishbulb> do you want a definition or a popular consensus
[15:54:37] <z64555> a definition is a popular consensus. :D
[15:55:01] <fishbulb> I mean if you asked uncle whoever if a washing machine was a robot, no, in my opinion, lots of them are
[15:55:12] <deshipu> that's why RC models are not robots -- unless they are ornithopters or other animal-shaped things
[15:55:14] <fishbulb> the ones with fuzzy logic and stuff? yeah
[15:55:21] <fishbulb> RC models are controlled by a human
[15:55:36] <fishbulb> if you add sensors and a level of autonomy, then.. I think that's a robot
[15:55:47] <deshipu> so are military drones
[15:55:54] <deshipu> yet we still call them "drones"
[15:55:56] <fishbulb> those are definitely robots
[15:56:08] <fishbulb> they fly themselves
[15:56:08] <deshipu> even though they are controlled by humans
[15:56:11] <fishbulb> the human points it
[15:56:24] <deshipu> well, quadcopters have automated stabilisation too
[15:56:26] <fishbulb> they perform the role of the pilot
[15:56:44] <deshipu> the difference is quantitative, not qualitative
[15:56:46] <fishbulb> a quad with just a normal RC controller? no that's kind of a toy, it can stabilise itself
[15:57:07] <fishbulb> because it's 100% impossible to control vectored thrust without a flight controller of some kind
[15:57:15] <z64555> nope
[15:57:17] <z64555> it's possible
[15:57:25] <z64555> difficult as fuck, but it's possible
[15:57:29] <deshipu> I think you can't really just look at a device and say if it's a robot -- you need a wider context, you need to know what it is for
[15:57:44] <fishbulb> show me somone who can... one that can fly around by itself, react to sensors, follow gps coordinates, land itself, take off by itself? that's a robot
[15:57:44] * z64555 drops an f-bomb by accident
[15:58:02] <fishbulb> show me anyone who can manually control 4 motors and fly one around
[15:58:16] <fishbulb> if someone like that existed they'd be all over the internet
[15:58:18] <deshipu> fishbulb: there are some amazing acrobatic rc pilots who disable the stabilization for some of the tricks
[15:58:47] <fishbulb> the vectored thrust is still controlled by the flight computer, they arne't controlling each of the rotor speeds
[15:58:51] <z64555> "manually control all 4 motors" See now you're overstepping there
[15:59:01] <deshipu> it's not a dlight computer, it's just a simple mixer
[15:59:11] <deshipu> s/dlight/flight
[15:59:20] <fishbulb> it is a flight computer, it contains code, just like the computers in the speed controllers, which are also computers
[15:59:30] <deshipu> your rc radio contains code
[15:59:36] <fishbulb> I don't have a radio
[15:59:42] <deshipu> your toaster contains code
[15:59:50] <fishbulb> actuallyI have a telemetry radio
[15:59:56] <SpeedEvil> I have a nokia n900 phone.
[15:59:57] <fishbulb> I don't have a controller though
[16:00:04] <deshipu> SpeedEvil: a computer!
[16:00:05] <fishbulb> I hate my n900
[16:00:05] <SpeedEvil> This has one user CPU core.
[16:00:20] <SpeedEvil> I counted the number of CPU cores in it, and got to about 13
[16:00:27] <z64555> fishbulb: you know anything about controller theory?
[16:00:36] <fishbulb> what controller
[16:00:39] <deshipu> it's "control theory"
[16:00:46] <z64555> deshipu: oh thanks
[16:00:54] <z64555> yeah, control theory
[16:01:28] <fishbulb> SpeedEvil, do you want an n900?
[16:01:54] <z64555> The mixer on the quad we're talking about is an open-loop controller
[16:02:13] <z64555> Aside from the radio, there are no sensor input to the controller
[16:02:14] <fishbulb> would you call an atmega a computer
[16:02:22] <deshipu> no, a microcontroller
[16:02:24] <SpeedEvil> fishbulb: I'd like a modern one.
[16:02:24] <fishbulb> ok whatever, I don't have those kinds
[16:02:27] <z64555> you're strawmanning. stop that
[16:03:04] <fishbulb> is an ESC a computer?
[16:03:11] <z64555> again, you're strawmanning
[16:03:22] <fishbulb> do you know what a strawman argument is
[16:03:38] <z64555> yes, it's an overgenerallization
[16:03:47] <z64555> which deflects the direction of the argument
[16:04:06] <fishbulb> let's argue about what a computer is
[16:04:13] <SpeedEvil> Let's not.
[16:04:13] <deshipu> let's not
[16:04:41] <z64555> a computer is something that takes in input and computes an output. done deal.
[16:04:42] <deshipu> let's have a productive discussion about something awesome
[16:05:05] <z64555> mechanical, electronic, hydrolic. all can be considered a computer
[16:05:17] <deshipu> ant-based
[16:05:56] <deshipu> or, if you are patient, you could even build one out of students
[16:06:05] <deshipu> but you'd need a lot of error correction
[16:06:20] <z64555> which involves more students working in parallel
[16:06:43] <z64555> it was common practice in the 60's to hire engineers to do calculations with slide rules and copious amounts of paper
[16:07:30] <deshipu> mechanical turk
[16:09:34] <z64555> there, we're done arguing what a "computer is"
[16:10:54] <deshipu> I wonder how difficult it will be to solder that atmega in a mlf package
[16:11:00] <deshipu> as opposed to tqfp
[16:11:52] <deshipu> last time I had to do a qfn, I just used my kitchen stove
[16:13:16] <z64555> my toaster oven can't get hot enough to do soldering, not without me modifying it. :/
[16:14:53] <fishbulb> are you guys either studying, or stuck in acedemia after studying?
[16:15:22] <deshipu> nope
[16:15:32] <fishbulb> why do you sound so burnt out
[16:15:39] <fishbulb> and why'd you mention "students can be a computer"
[16:15:42] <doomlord> anything new
[16:15:56] <deshipu> because they totally can
[16:16:06] <deshipu> given enough training, of course
[16:16:13] <z64555> and if they're mathematicians, they already are
[16:17:26] <deshipu> fishbulb: we had that discussion about what a robot is exactly, or what a computer is a lot of times already
[16:17:48] <deshipu> in the end, it doesn't matter how you call the thing that you build
[16:17:54] <deshipu> as long as you build it
[16:18:52] <deshipu> I noticed that as soon as something starts to work effectively and becomes popular, people stop calling it a "robot" and give it its own name
[16:19:12] <fishbulb> I need some kind of probably brushed motors
[16:19:31] <z64555> How will they be used?
[16:19:31] <deshipu> I've never heard about probably brushed motors, how do they work?
[16:20:09] <fishbulb> on wheels for something that goes outdoors
[16:20:20] <z64555> ._.
[16:20:28] <z64555> Speed, Torque, or Position.
[16:20:37] <fishbulb> to push the entire thing
[16:20:37] <z64555> which of the two do you need
[16:20:42] <fishbulb> torque and position
[16:20:47] <z64555> stepper.
[16:20:57] <fishbulb> too expensive
[16:21:06] <fishbulb> it has to carry like 2kg
[16:21:43] <z64555> then you'll need to use crummy rotary encoders and make a position controller
[16:23:12] <z64555> You'll have higher position control with a BLDC than a BDC, but you won't have as smooth a torque curve
[16:24:57] <z64555> an outrunner BLDC, that is
[16:25:44] <z64555> although I think some inrunner's have similar torque ratings to BDC's
[16:26:13] <z64555> depends on if the rotor is a magnitised mass or just the axle
[16:31:07] <fishbulb> I'm using gps
[16:31:26] <z64555> ok. how does that factor into your motor selection
[16:31:41] <fishbulb> becuse it doesn't have to go fast
[16:32:07] <fishbulb> and it's cheaper to gear down a motor to carry shit and get over stuff
[16:37:17] <z64555> k. so just get an electric scooter and hack it to your needs
[16:38:13] <deshipu> that's what I suggested
[16:39:09] <deshipu> I've been told that 3d printers are cheaper than electric scooter
[16:41:00] <veverak> "3d printer" starts to be too vaugue
[16:41:10] <veverak> for mine ideal I would need a tons of money and at least three persons to build it
[16:41:12] <veverak> ehmm
[16:41:14] <veverak> :/
[16:41:16] * veverak needs more funds
[16:41:29] <deshipu> I've also seen people use motors from cordless drills
[16:41:35] <deshipu> together with the gearboxes
[16:42:19] <deshipu> veverak: funds and time are just questions of priorities ;-)
[16:42:56] <veverak> got a point
[16:43:02] <veverak> but nah, too many projects to do/finish
[16:43:22] <deshipu> I'm a man with a one-track mind, too much to do in one lifetime
[16:44:00] <deshipu> veverak: clone yourself
[16:44:06] <veverak> deshipu: was thinking about that
[16:44:14] <veverak> but for now, it seems it's easier to be academic profesor
[16:44:28] <veverak> so many people that needs something from you (credits) and know how to do things
[16:44:29] <veverak> :))))
[16:44:32] <deshipu> they never get anything done
[16:45:12] <z64555> lol
[16:46:15] <deshipu> it's kinda funny, I have the funds for pretty much any project I could think of, and I would have the time -- what I don't have is the attention span and the motivation
[16:46:41] * veverak is more in phase that he can generate more projects that he can finish reasonably
[16:46:56] <deshipu> veverak: that just means you are alive
[16:47:14] <deshipu> it's always like that
[16:47:28] <veverak> deshipu: yeah
[16:47:36] <deshipu> but can you generate projects that actually can be finished?
[16:47:39] <veverak> but I figured that I am more into sociology/psychology than I thought
[16:47:49] <veverak> and started to get social skills a lot
[16:47:59] <deshipu> hehe, sensitive period
[16:48:06] <deshipu> sometimes it comes late
[16:48:16] <veverak> thinking that path "find people that will do your projects for you" makes sense
[16:48:18] <veverak> ;)
[16:48:37] <veverak> deshipu: I had always tendency to see how people think and tendency to affect their behavior
[16:48:48] <veverak> it's just that I tried to lock it and not to use it most of my life
[16:48:57] <deshipu> "if you want your people to build a ship, don't give them tasks and orders, but instead teach them to love the sea"
[16:49:19] <z64555> oooh, I love that quote
[16:49:20] <veverak> exactly
[16:49:22] <veverak> :)
[16:49:25] <z64555> stealing it
[16:49:35] <deshipu> I can't remember the exact quote
[16:49:41] <deshipu> but it was something like that
[16:49:49] <deshipu> there was a ship and a sea in it, anyways
[16:51:12] <veverak> anyway
[16:51:22] <veverak> point is in finding people that shares interest in what you are doing
[16:51:32] <veverak> they pose greater skills in areas you lack
[16:51:43] <veverak> and persuase them for cooperation ;)
[16:54:22] <z64555> if I can teach my self to love programming, I might be able to get more things done around here
[16:54:33] <z64555> :D
[17:02:12] <fishbulb> a 200 dollar scooter for "parts" is a shit suggestion
[17:02:36] <fishbulb> and I can show you the invoice for the parts to build a prusa i3 mk2 clone
[17:02:46] <fishbulb> it was 160 all up
[17:03:13] <fishbulb> it's an open source design, there was a flash sale on one called the "anet a8"
[17:06:19] <fishbulb> works fine, the only shit thing about it is the bed moves as the Y axis, so the whole thing needs parts printed out for rigidity if you're printing large things. it also needs an enclosure and lots of other stuff before it can handle a long print without crapping out
[17:07:13] <fishbulb> and a raspi as slicer/print server eliminates swapping out sd cards
[17:13:39] <z64555> How is the scooter a "shit suggestion?" You get a carriage with wheels, motor, gearing, and speed controls, and the whole lot is rating for much more than 2Kg payload
[17:13:51] <z64555> s/rating/rated
[17:14:33] <fishbulb> for 4 wheels that's 400USD
[17:15:07] <fishbulb> how much hobbyking stuff can you get for that?
[17:15:53] <z64555> how much hobbyking stuff is relevant?
[17:15:55] <fishbulb> the requirements are a rover thing for grass dirt and rocks
[17:16:08] <fishbulb> motors controllers and batteries
[17:16:25] <z64555> mkay. where's your frame coming from
[17:16:46] <z64555> If you're building it yourself, do you have the tools to weld it together
[17:17:07] <z64555> of if doing bolted construction, the drill press to make the holes
[17:17:12] <fishbulb> this robot is really pissing away time between when I can drive my car
[17:17:31] <fishbulb> muscle touring/road rally car that has been built and designed from the ground up
[17:17:47] <fishbulb> but I'm not welding together a robot frame, I'm using carbon tubes and epoxy
[17:18:38] <fishbulb> after the spaceframe is designed I'll probably 3d print some joiner things, if it needs a shell that'll probably be vacuum formed out of that thin hobby RC shell stuff
[17:19:20] <fishbulb> because making a frame like that would be easier than hand profiling carbon tubing to fit together in a spaceframe
[17:19:52] <fishbulb> joiny-bits for anything that joins 2 or more tubes at an angle
[17:20:37] <fishbulb> so brushed dc motors are pretty cheap compared to all the alternatives, since I don't want to spend 400 on the motors and controllers alone, I'd rather use some of the boxes and boxes of shit I already have from drones/planes
[17:20:42] <z64555> don't make a compression couple for more than 2 rods
[17:21:03] <fishbulb> I'm really just thinking for holding htem while epoxying them together
[17:21:05] <z64555> better to use a hub and clamps
[17:21:18] <fishbulb> I don't have that stuff
[17:21:27] <z64555> print em
[17:21:43] <fishbulb> I know how to make a spaceframe
[17:22:19] <fishbulb> the tubes would need to butt against each other
[17:23:10] <fishbulb> they're also ridiculously cheap from hobbyking, and pretty strong, I thought I'd be able to snap them with my hands, I also have dozens and dozens of arrow shafts but those are very expensive, that would be a total waste
[17:23:40] <fishbulb> if I can make some sort of grabber arm that would be good because nothing is very fun without FPV
[17:23:43] <z64555> make a joint like a plumbing elbow that slips over the tubes, or have it slip inside the tube
[17:24:03] <fishbulb> slip inside then epoxy the actual tubes together was what I was thinking
[17:24:07] <fishbulb> I haven't had a 3d printer before
[17:24:58] <z64555> eh, I'd expoxy to the joint piece, would transfer forces to it instead of just the expoxy
[17:24:59] <fishbulb> it's a totally new thing to be able to make small plastic items, and I love plastic. mostly in the form of fibre reinforced plastic but ABS is still cool
[17:25:13] <fishbulb> GLUE THE TUBES
[17:25:14] <fishbulb> done.
[17:25:14] <z64555> Unless you're thinking of beveling the rods?
[17:25:28] <fishbulb> that goes without saying if they're going to touch one another
[17:25:36] <fishbulb> notch them, not bevel
[17:25:57] <z64555> hm.
[17:26:22] <fishbulb> bevel is make an edge rounded
[17:26:40] <fishbulb> profiled in some way
[17:27:13] <z64555> that's chamfer, I think
[17:27:14] <fishbulb> so brushed motors connected to some kind of offroad wheels, then some kind of suspension which would most likely be swingarm like the rear wheel of a motorbike on all 4 wheels
[17:27:25] <fishbulb> and if it steers like a tank that's easier to control
[17:27:39] <z64555> oh, "miter" is the term I was thinking of
[17:28:00] <fishbulb> make the tubes stick to the other tubes
[17:46:39] <anniepoo> 8cP well, insane push at work is more or less under control
[17:46:47] <anniepoo> 8cD back to robot!
[17:47:06] <anniepoo> esp. since I'm snowed in at the shop!
[17:47:19] <z64555> have the robot dig out all the snow! :D
[17:48:18] <anniepoo> you know, I've thought about robotic snowblower, and indeed the local paper ran a cartoon about a robot snowblower today
[17:48:45] <anniepoo> but it'd have to be a big machine, comparable to a manual one,
[17:48:51] <anniepoo> and it'd be dangerous
[17:49:24] <anniepoo> cause if it weighed less it wouldn't be able to bite into the snow
[17:49:44] <anniepoo> but believe me, a solution for snow would make you very, very rich
[17:50:19] <anniepoo> to clarify - I live in an apt within about 60 seconds walk of the shop
[17:50:45] <anniepoo> and I'm wondering if I can get home tonight
[17:50:48] <anniepoo> 8c/
[17:51:28] <anniepoo> https://www.instagram.com/p/BO-t6xRAlJC/?taken-by=anneogborn
[17:51:35] <anniepoo> new pic of her snailiness
[17:57:20] <z64555> ah, its smaller than I had expected
[17:58:05] <anniepoo> well, this is the drive unit
[17:58:18] <anniepoo> the social robotics shell will sit atop it
[17:58:52] <anniepoo> the snail's shell will be about a foot in diameter
[17:59:06] <anniepoo> and a head and tail will each stick out about 9 inches
[17:59:38] <anniepoo> The tail's going to have a mechanism like a pull toy -
[18:00:06] <anniepoo> the robot pulls a trailer, crank on the wheels moves segments
[18:00:19] <anniepoo> Anybody with RC car experience?
[18:01:03] <anniepoo> I want good contact between tires and indoor floor - probably short pile carpet or hard floor
[18:03:26] <anniepoo> I'm thinking largish tires with knobby studs for 'off road'
[18:03:28] <anniepoo> 8cD
[18:29:21] <doomlord> r.e. self-driving cars: are there any attempts to control such things with remote cameras ,e.g. imagine having comprehensive coverage of a driveable area from traffic signal camera etc
[18:30:35] <Snert> rugged little bot. love it. call it Toad :)
[18:30:58] <anniepoo> It's gonna be a snail 8cD
[21:03:53] <anniepoo> you probably want a bigger diameter on the axle
[21:04:31] <BeautiCode> why a bigger diameter?
[21:04:33] <anniepoo> "gear ratio" is the number of turns of one end of a system to the other
[21:04:46] <anniepoo> motors turn fast, but aren't very powerful
[21:04:52] <anniepoo> they don't have much torque
[21:05:11] <anniepoo> you'll need your wheels to turn slower
[21:05:50] <anniepoo> the same linear amount of belt goes past the two pulleys
[21:06:03] <anniepoo> in 1 second
[21:06:13] <anniepoo> so the bigger pulley turns less turns
[21:07:05] <anniepoo> power is torque through angle.
[21:07:28] <anniepoo> another way to hook to a motor shaft is a coupler
[21:08:05] <anniepoo> this will have a 1:1 gear ratio, but lets you use a different sized shaft or whatever
[21:08:28] <anniepoo> you can make couplers for small motors out of a bit of small plastic tubing
[21:09:16] <anniepoo> just slip it on the motor shaft and the shaft you're connecting to
[21:09:27] <anniepoo> think of the kind of tubing in IV lines or aquariums
[21:09:29] <anniepoo> since it's small diameter, it can be hard for the tubing to grip
[21:09:52] <anniepoo> a little wire around the tubing, twisted tight, helps
[21:10:15] <BeautiCode> i think for this first project, i wanna keep it as simple as possible...i might go with the first solution
[21:10:32] <BeautiCode> to connect a rubberband to the shaft and the axle
[21:10:53] <BeautiCode> i have a guy who's good at woodworking, so i might talk with him about
[21:11:05] <anniepoo> start with cardboard boxes
[21:11:05] <BeautiCode> making an axle & wheels out of wood
[21:11:15] <anniepoo> you can make a robot from cardboard
[21:11:20] <BeautiCode> cardboard?
[21:11:28] <rue_shop3> you cant start finished
[21:11:37] <BeautiCode> isnt cardboard too...flimsy?
[21:11:41] <rue_shop3> you have to get experience to know what your doing
[21:11:44] <anniepoo> yah, like the boxes you get from the grocery store to move with
[21:11:48] <rue_shop3> flimsy wont be the main problem
[21:12:08] <rue_shop3> its the other things you dont know about that the cardboard and wood will teach you
[21:12:15] <anniepoo> yes
[21:12:17] <anniepoo> I'm with rue
[21:12:18] <rue_shop3> like major design flaws you didn't expect
[21:12:33] <rue_shop3> when the last limit is the cardboard and wood, change those materials
[21:12:39] <anniepoo> I need 3 mins
[21:12:42] <rue_shop3> dont fret the stuff you know about
[21:12:58] <rue_shop3> your gonna have more than a time with the things you dont know about
[21:13:09] <BeautiCode> hmm, you're right
[21:13:47] <rue_shop3> when you have a 3d printer, and you want 1 part, you end up with 3, 2 of them defective in ways you didn't see cause you werne't holding them
[21:14:13] <rue_shop3> and one you put up with the flaws with cause you dont want to make any more cuase you only wanted one
[21:14:56] <rue_shop3> I need a motor driver
[21:15:52] <anniepoo> yep
[21:16:12] <BeautiCode> i think i'll start with cardboard xD
[21:16:58] <anniepoo> I'm building a robot right now. It's basically a hobby type robot, but it's for professional work, and I'm a pro engineer with a complete machine shop and computer controlled lasers in the next room
[21:17:13] <anniepoo> and my robot's planned to have some corrugated cardboard in it
[21:17:33] <anniepoo> the robot's to interact with children. I can't hurt the kiddos with cardboard
[21:18:02] <anniepoo> and yes, totally what rue said
[21:18:49] <BeautiCode> looks like someone already built something similar to the project i want to build,completely from cardboard
[21:18:51] <BeautiCode> https://cdn.instructables.com/FMF/EIMZ/HYC61K21/FMFEIMZHYC61K21.MEDIUM.jpg
[21:19:20] <anniepoo> something you could do right now - take your timing belt, your motor with the timing pulley on it, some cardboard, duct tape, etc and make a wheel that turns
[21:20:37] <anniepoo> if you mess it up, eh, you're out 50 cents worth of duct tape and a couple nails
[21:21:41] <anniepoo> in 'from the earth to the moon' tv series, the engineers building the LEM made a cardboard mockup to test their idea
[21:22:47] <BeautiCode> i think im gonna do that
[21:23:29] <rue_shop3> how is that modeled so well after the bookkeeper at work?
[21:23:31] <BeautiCode> i have all the stuff...cardboard boxes, duct tape, etc.
[21:23:35] <anniepoo> how will you power your motor?
[21:23:50] <anniepoo> 8cD we're all pack rats
[21:24:18] <rue_shop3> get 200 switches from ebay, should run you about $2
[21:24:42] <rue_shop3> I suggest on-off-on DPDT switches
[21:24:59] <rue_shop3> you can hook them up to a motor to do forward-stop-reverse
[21:25:03] <rue_shop3> if your using dc motors
[21:25:24] <anniepoo> hang on
[21:25:29] <rue_shop3> its quite viable to use an arduino and servos now too
[21:25:39] <BeautiCode> im probably gonna use this smaller 3-6v motor, and power the whole thing with 4AA batteries
[21:25:41] <rue_shop3> damnit, I got radius and diameter mixed up again!
[21:25:44] <anniepoo> One of the most frequently cited works in social robotics is the tweenbot
[21:25:56] <BeautiCode> its the smaller one on the left in the pic i showed earlier
[21:26:34] <anniepoo> amazingly, this robot started a whole academic subject- human dependent robotics
[21:26:48] <anniepoo> which now has hundreds of papers and so on
[21:26:55] <anniepoo> http://tweenbots.com/
[21:26:58] <rue_shop3> daaaaamnit I cant just scale it down
[21:27:06] <rue_shop3> ugh, that took 2 hours to model
[21:27:17] <BeautiCode> how long are you two going to be online?
[21:27:20] <anniepoo> BeautiCode, check the robot out
[21:27:28] <anniepoo> I expect to be around for hours
[21:27:37] <anniepoo> I'm more or less trapped here
[21:27:40] <BeautiCode> that tweenbot thing looks cool
[21:27:48] <anniepoo> it has no computer on board
[21:27:59] <anniepoo> it's just a battery and a motor
[21:28:17] <anniepoo> you could make one tonight if you wanted and if you're somewhere warm try it tomorrow
[21:28:20] <BeautiCode> sounds pretty simple, ill start with something like that
[21:29:01] <BeautiCode> i have to go do something though, before i get started
[21:29:02] <anniepoo> heck, yes - reproduce tweenbot, with the whole thing with the flag
[21:29:09] <anniepoo> in some public area
[21:29:15] <anniepoo> it's a pretty amazing experiment
[21:29:47] <BeautiCode> tweenbot will be my next project
[21:30:08] <BeautiCode> im gonna start on it in a little bit though, but gotta brb for a little bit
[21:30:10] <anniepoo> lol
[21:30:12] <anniepoo> cool!
[21:30:24] <anniepoo> and you WILL learn some shop skills
[21:30:32] <anniepoo> how to cut straight
[21:30:36] <anniepoo> how to plan out a project
[21:30:44] <anniepoo> how fast a motor runs,
[21:30:51] <anniepoo> what a gear ratio is
[21:30:53] <anniepoo> etc
[21:30:59] <anniepoo> so it'll be cool
[21:31:31] <BeautiCode> i appreciate all the help you and rue have given me in the past 2 hours
[21:31:37] <anniepoo> 8cD
[21:31:41] <anniepoo> it's been fun
[21:31:54] <BeautiCode> i really do, thanks :D
[21:32:00] <BeautiCode> im gonna save this chatlog if you dont mind, so i can refer back to some things
[21:32:16] <BeautiCode> and ill be back on hopefully in like 2 hours
[21:32:33] <BeautiCode> so i can start working on the tweensy clone
[21:34:12] <BeautiCode> be back in a few
[21:36:20] <anniepoo> sure
[21:36:27] <anniepoo> tweenbot
[21:37:03] <anniepoo> elmers glue makes strong, rigid joints with cardboard
[21:45:46] <rue_house> and hot glue sets fast and givesyour robot that halloween effect when your finished
[21:47:03] <rue_house> nothing like finding out your motors underpowered cause it got stuck on a hot-glue-string
[21:47:24] <Tom_L> rue_house
[21:47:29] <rue_house> Tom_L,
[21:47:44] <rue_house> 42
[21:47:48] <Tom_L> how do you go about converting various html logs to a consistent format?
[21:47:57] <rue_house> perl
[21:47:58] <Tom_L> trying to make all the old logs look the same
[21:48:18] <rue_house> perl can pick off the templating via regular expressions
[21:48:23] <Tom_L> jt wrote a python file to convert them but it doesn't do them all
[21:48:39] <Tom_L> and it truncated some of them
[21:48:41] <rue_house> perl, regex
[21:48:48] <rue_house> perl
[21:48:52] <rue_house> regular expression
[21:49:31] <rue_house> you got me eating supper
[21:49:47] <rue_house> show me one line you want to convert, do NOT show me what you want to convert it to :)
[21:49:52] <Tom_L> well at least you're eating it instead of burning it
[21:50:09] <Tom_L> they vary quite a bit
[21:50:17] <rue_house> just show me a line
[21:50:49] <rue_house> you have 7 perogies of time left
[21:51:06] <Tom_L> http://tom-itx.no-ip.biz:81/~tom-itx/irc/logs/%23avr_old/2011-10-21.html
[21:51:07] <rue_house> 6.5 perogies
[21:51:38] <rue_house> <a href="#00:16:04" name="00:16:04" class="time">[00:16:04]</a> <span class="person" style="color:#746086">&lt;feurig&gt;</span> but thanks.
[21:51:38] <rue_house> <br/>
[21:52:01] <Tom_L> and hyperlinks
[21:52:17] <Tom_L> he's got hyperlinks opening in a new window
[21:52:22] <Tom_L> which is nice
[21:52:42] <rue_house> thats a few layers alright
[21:52:51] <Tom_L> but it truncated alot of files
[21:52:56] <Tom_L> i'm putting them all back now
[21:52:57] <rue_house> perl
[21:53:01] <rue_house> regular expressions
[21:53:07] <Tom_L> i don't know perl
[21:53:13] <Tom_L> not even a little
[21:53:34] <rue_house> and you dont know regular expressions, which is a whole nother language all to itself
[21:53:44] <rue_house> but it cant do this easy
[21:53:56] <Tom_L> can't?
[21:54:15] <rue_house> WHERE DID THAT T COME FROM!!!!! WHO PUT THAT THERE!
[21:54:26] <rue_house> it CAN do it easy
[21:55:08] <rue_house> (time) (name) (text)
[21:55:15] <rue_house> ok thats what it breaks down into
[21:55:16] <Tom_L> i've got ~10k or more files
[21:55:28] <Tom_L> not on all of them
[21:55:36] <Tom_L> some of the older ones are funny
[21:55:53] <rue_house> you have to write a perl script for each version of file you created
[21:56:14] <Tom_L> it would be the same with his python then
[21:56:21] <Tom_L> it gets some of them but not all
[21:56:28] <rue_house> but WAAAAY easier in perl
[21:56:55] <Tom_L> neither one would be easy for me
[21:56:56] <rue_house> been a while
[21:57:01] <rue_house> <a href="#00:16:04" name="00:16:04" class="time">[00:16:04]</a> <span class="person" style="color:#746086">&lt;feurig&gt;</span> but thanks.
[21:57:01] <rue_house> <rue_house> <br/>
[21:57:32] <Tom_L> about 3 days in that ver, i removed the color and set them all to red for the person
[21:57:59] <Tom_L> i think most of them are ok but there are a few odd ones
[21:58:18] <Tom_L> just how far back do you think i should keep?
[22:00:29] <rue_house> .+[(dd:dd:dd)].+;(a+)&.+>(.+)<.+
[22:00:30] <rue_house> I think
[22:00:30] <rue_house> I'll try later, your out of perogies
[22:00:30] <Tom_L> goin to the shop?
[22:00:30] <rue_house> I can write a stripper for each type of file, and you can write in the code to make it the new format
[22:00:30] <rue_house> I'll give you variables for the time, name and text
[22:00:30] <rue_house> I can prolly strip the urls back down to raw text
[22:00:30] <Tom_L> i honestly think it may be best to just leave them as they are
[22:05:36] * rue_shop3 tries to work out if his printer can take an 8cm high print
[22:17:00] * anniepoo starts thinking about architecture of software for the snail