#robotics Logs

Nov 01 2011

07:19 Tom_itx zhanx
07:19 Tom_itx http://tom-itx.dyndns.org:81/~webpage/rue/stlink_output.txt
07:20 Tom_itx http://tom-itx.dyndns.org:81/~webpage/rue/stlink_output1.txt
07:20 Tom_itx 2 attempts
08:20 zhanx so totally dont feel like working today
08:20 zhanx the first one is good, second not so much
08:21 zhanx did you close it right?
08:37 rue_bed its cause its dark prolly
09:08 rue_mohr "Although PID could work as a quick solution, it is not recommended for long-term use in a system exhibiting nonlinearity due to a dead-zone."
09:12 home guys
09:12 home I need to know what PID is
09:12 home is it used explictly for line following?
09:12 rue_mohr no
09:13 rue_mohr hmm
09:13 rue_mohr home, lets say, you have a motor with a gearbox
09:13 home ok :D
09:13 rue_mohr and the position of the shaft out of the gearbox is measured with a potentiometer
09:13 rue_mohr (variable resistor)
09:14 home ok
09:14 rue_mohr now, lets say you have an arduino that can control the motor, and read the position from the pot
09:15 home ok
09:15 rue_mohr add to that, software that takes a position from the computer, and tries to make the shaft of the gearbox go to that position
09:15 home yeah.
09:15 rue_mohr 'go to 22.5 degrees' kinda thing
09:15 home I think the problem is...they get out of that position?
09:16 rue_mohr you have a feedback loop, the controller commands the motor that changes the pot that changes the position to the controller
09:16 rue_mohr so the controllers job is to wastch the position and continiously update the power to the motor to make the shaft stay where its supposed to be
09:16 rue_mohr still with me?
09:17 home not sure
09:18 home http://www.ostan.cz/LineFollowerSimulator/
09:18 rue_mohr do you understand the arduino has to push the motor back and forth to get the gearbox to where the user wants it?
09:18 home only thing that I know about PID
09:18 home >.<
09:18 rue_mohr hold off on the line following for a min
09:19 home not really XD
09:19 rue_mohr hmm
09:19 rue_mohr ok
09:19 rue_mohr lets say your put in charge of the volume knob on a stereo
09:20 rue_mohr youre familiar with turn knobs right?
09:20 home yes
09:20 rue_mohr gooood!
09:20 rue_mohr ok, so someone says "turn it up to 7" and you have to push it around till you get to 7
09:20 rue_mohr yes?
09:20 home yep
09:21 rue_mohr ok someone says "turn it to 5", you have to push it the other way
09:21 theBear except that the knob doesn't have numbers on it, so each change you gotta judge how loud it is and get closer if you can
09:21 home yes
09:21 rue_mohr now it gets trickey, someone else tries to turn the knob to 6, you have to put enough force on it to keep it at 5
09:22 rue_mohr then the person goes away, and you dont have to push anymore to keep it at 5
09:22 rue_mohr this is the same thing, cause your taking the place of the arduino, the knob is the shaft of the gearbox
09:23 rue_mohr ok with that so far?
09:23 home yes
09:24 rue_mohr ok if the computer is doing the pushing it uses math to work out how hard to push
09:24 rue_mohr PID is a compination of 3 formulas
09:24 rue_mohr Proportionate
09:24 rue_mohr Integral
09:24 rue_mohr Derivitive
09:25 rue_mohr Proportionate means that the more wrong the position is, the harder it pushes the motor
09:26 rue_mohr shall I go on?
09:26 home Sure
09:26 home If I dont understand it I Will look it up XD
09:26 home or geet it clarified :)
09:27 rue_mohr Integral means that if there is an amount of error that isn't going away over time, to push harder
09:28 rue_mohr so if the volume was at 4, was supposed to be at 5, and the computer was pushing with (5-4=1) "1" and the knob was staying at 4, Integral would gradually up the force
09:29 rue_mohr ?
09:29 home i see
09:30 rue_mohr derivitive limits the speed of the movement
09:31 rue_mohr if the volume was at 1 and needed to be changed to 10, the computer might noramally push with (10-1=9) "9", but doing so might have the volume knob moving so fast that when it got to 10, it wouldn't be able to stop fast enough, it might accidently hit 11 before it had time to turn around and try to get back to 10
09:32 rue_mohr (obscure joke ref for those who got it)
09:32 e_mohr glances at theB
09:32 home Hmmm
09:32 eBear sm
09:33 home so rue_mohr would you icnrease the speed gradually?
09:33 home so that it does not go past 11? but hits 10?
09:33 home so it gets pushed by like.. 3
09:33 rue_mohr yes, it says that if your going really fast, to slow down
09:34 home I really need to try out PID
09:35 rue_mohr try out proportionate first
09:35 rue_mohr do you see how that applies to the steering of your robot?
09:35 home somewhat I guess.
09:35 theBear consider the 3 words.. if you are far from something you gonna put more effort into getting there...
09:35 home oh of course..
09:35 home and you slow down once you get there?
09:35 home as you keep going?
09:36 home ok well thanks :)
09:36 theBear you are constantly re-evaluating what to do "now"
09:36 home will have to research PID more..
09:36 home :/
09:36 home bye XD thx for the helpz
11:13 Tom_itx zhanx apparently not
12:39 KonguPanda hello robotics :)
17:23 coppermine anything wrong with using a small mini itx motherboard for a robot vs a microcontroller board?
17:55 coppermine i have the perfect self destruct method for my robot, release 50ml of water directly on the motherboard
17:56 coppermine using an actuator :)
17:56 GuShH coppermine: you'll find water isn't very conductive on it's own
17:57 coppermine its not something i want to test often
17:57 GuShH You can try it with your dmm
17:57 coppermine how about switching to a 24 voltage battery
17:57 coppermine haha
17:57 coppermine instead of 12v
17:58 GuShH both measuring resistance and current across the expected amount of water using the same voltage levels present in the motherboard
17:58 coppermine keep a 2AH 24V battery small
17:58 GuShH If you want self destruct, what are you trying to destroy exactly?
17:58 coppermine who wants to know?
17:58 GuShH data?
17:58 GuShH the machine itself?
17:58 coppermine yes
17:58 coppermine in case it goes in the wrong hands
17:58 GuShH explosives.
17:59 GuShH This reminds me of seal team 6 and how they "forgot" to rig the tail rotor ...
18:00 GuShH Either hes a former SEAL or he just spazed out.
18:08 ace4016 thermite explosion!
18:11 KonguPanda hello robotics people !
18:26 rue_mohr high
18:27 Tom_itx lo
18:28 Tom_itx long day
18:30 rue_mohr a half hour wire diagnosis just took me and a partner 3.5 hours
18:30 rue_mohr diagosis: cant get around it, have to run a new wire
18:31 Tom_itx tie on the end and pull
18:32 rue_mohr insuccicent tensile strength on the wire to pull out all the staples holding it behind the drywall for its 400 foot run
18:32 Tom_itx aww
18:33 Tom_itx time for decorative conduit
18:33 rue_mohr the customers dont want me ripping out all the drywall for the length of the wire run either
18:34 KonguPanda how can you diagnose a wire? it either conducts or not?
18:34 rue_mohr wrong
18:34 rue_mohr in this case, it conducts too well
18:34 rue_mohr 2 of its 6 conductors conduct between each other in the middle of its length
18:35 Tom_itx mice?
18:35 joink rue_mohr: do you know where in the cable stretch that is?
18:35 rue_mohr Tom_itx, no, prolly rats
18:35 rue_mohr joink, if I did, it'd be fixed
18:35 GuShH Heh maybe the staples?
18:35 Tom_itx could be
18:35 rue_mohr prolly, they used a gun to staple them, even pro's miss occasionally
18:36 GuShH Sounds like not fun.
18:36 rue_mohr no it'll take a few hours to run a new cable
18:36 joink rue_mohr: spin off from the company I'm working in now :P http://www.wirescan.no/
18:37 rue_mohr a TDR would tell us how far along the short is, but the wirepath is not accessable
18:37 KonguPanda I see :D
18:37 ShH has yet to experiment with his improvised
18:38 rue_mohr any customer who can afford that just gets us to put in a new wire
18:38 rue_mohr scope and?
18:38 GuShH pulse generator basically
18:38 rue_mohr whats the speed of electricity about 2m/ns?
18:39 ShH m
18:39 rue_mohr 300m/us
18:39 rue_mohr .3m/ns
18:39 rue_mohr cool
18:39 rue_mohr You have: 1c
18:39 rue_mohr You want: meters/ns
18:39 rue_mohr * 0.29979246
18:40 GuShH Similar to this in essence http://www.epanorama.net/circuits/tdr.html
18:40 rue_mohr about 1foot/ns
18:41 rue_mohr so between prolly about 10 and 400n
18:41 rue_mohr s
18:42 rue_mohr so, with m scope, I could get it down to about a 20' section
18:43 rue_mohr .1us/div min
18:43 rue_mohr /5
18:43 rue_mohr then again, I have a x10 on the horiz
18:43 rue_mohr if I"z lucky I could get down to 2'
18:44 rue_mohr 1/1ns
18:44 rue_mohr 1e9
18:45 rue_mohr mmm ghz...
18:45 Tom_itx ok now
18:45 Tom_itx what to do
18:46 rue_mohr microwave...
18:46 rue_mohr I'd need a counter with a gate that could do 1Ghz
18:47 Tom_itx why so quick?
18:47 rue_mohr to TDR a cable failure within 1'
18:48 Tom_itx flir
18:48 Tom_itx look for hot spots in the wall
18:56 GuShH Following Tom_itx's logic, I would hook the cable to a car battery and enjoy the fireworks.
18:56 GuShH While it sounds childish, this is how certain cable related issues are solved in the telecom industry
18:57 GuShH However they employ high voltages.
18:57 rue_mohr its a phone line
18:57 rue_mohr tho I can thinking if we ran some current thru it for a while
18:57 Tom_itx tell them it's time to go wireless
18:57 GuShH a high voltage pulse is sent out to vaporize the short
18:57 rue_mohr but the odds of starting a fire are too high
18:57 rue_mohr GuShH, again, the fire potential
18:57 GuShH well, in reality it's done to condition the cable
18:57 GuShH when there are sporadic crackling noises
18:58 rue_mohr I considered it
18:58 GuShH but a similar technique is used with milled PCBs for instance, high current to vaporize any copper strands shorting a trace
18:58 rue_mohr I'v done it
18:58 ShH is just sa
18:58 Tom_itx i found some board shorts that way
18:58 GuShH I'm still upset at ##electronics, they called me flyback and even suggested I was him.
18:59 rue_mohr actaully, to that effect, I'd like to say that using a microwave transofrmert to try to blow ALL the trances off a pcb is a bad idea and leads to a severly burned pcb and fire
18:59 Tom_itx they were gonna have to be jumpered anyway
18:59 rue_mohr I can tell your not flyback
18:59 Tom_itx and weren't easily spotted with the mask on them
18:59 rue_mohr GuShH, flyback isn't annoying (only) he's vulgar, the true meaning of it
19:00 GuShH I can almost picture the flaming PCB
19:00 rue_mohr pepsi can be offensive, but he's not vulgar
19:00 rue_mohr GuShH, I didn't have a camera to take a pic
19:00 rue_mohr !assist images/
19:00 tobbor Possibly http://eds.dyndns.org/~ircjunk/images/
19:00 GuShH pepsi has never been offensive to me, I offended him in fact.
19:00 rue_mohr http://eds.dyndns.org/~ircjunk/images/pcb_reuse1.jpg
19:00 Tom_itx http://tom-itx.dyndns.org:81/~webpage/scope/DSC_0001.JPG
19:01 rue_mohr http://eds.dyndns.org/~ircjunk/images/pcb_reuse2.jpg
19:01 Tom_itx that was one of them
19:01 rue_mohr its better to use a disc grinder to sand off all the traces, sometimes this even leaves you with plated thru-holes
19:02 rue_mohr haha, we ddos'd tom
19:02 GuShH I attack with dremel
19:02 GuShH Oh I thought that was me
19:02 rue_mohr remberber the old computer baords with huge grids of 14 pin chips?
19:02 rue_mohr I had that before veroboard
19:02 GuShH Upstream "depleted"
19:05 rue_mohr isn't that orlocks board for his telescope?
19:05 rue_mohr wait... no..
19:05 rue_mohr orlock made a ring heater
19:05 rue_mohr orlok, where are you
19:05 rue_mohr what do you think of my pcb reuse?
19:07 Tom_itx what's it for?
19:07 rue_mohr its a power control board
19:07 rue_mohr its *the* primary power board for the coffeebot
19:07 Tom_itx is that phenolic?
19:07 rue_mohr yea
19:11 Tom_itx you ever used gdb?
19:12 rue_mohr "yes"
19:12 rue_mohr I use ddd
19:12 rue_mohr ddd is a gui front end to gdb
19:12 Tom_itx i think that's what this is anyway
19:12 Tom_itx that output i pasted yesterday
19:13 rue_mohr you dont have to learn gdb's leet language, you just need to have patience with a 300 year old gui design
19:13 Tom_itx i don't think eclipse is the way to go with this
19:14 Tom_itx for a dev environement
19:14 rue_mohr no
19:14 rue_mohr xterm and nedit
19:14 rue_mohr and, if you insist, gdb
19:14 rue_mohr er ddd
19:14 rue_mohr nedit has powers of no other editor
19:15 rue_mohr search/replace and repeat keystrokes, it does those 2 things better than anybody
19:16 Tom_itx i just wanna figure something out for this
19:16 Tom_itx i've got the stuff for windows that will work
19:16 Tom_itx but i'm giving this an honest effort
19:17 rue_mohr hmm
19:17 rue_mohr do you have so-far notes?
19:18 rue_mohr I could try to follow them, but I dont have an arm to program
19:18 rue_mohr wait, maybe I do
19:18 rue_mohr whats the hammer
19:18 orlok i'm here
19:18 orlok been busy looking after the bub
19:18 rue_mohr the ethernet is all I have to program it via tho
19:18 rue_mohr bub?
19:19 rue_mohr wait no
19:19 rue_mohr the hammer has a programming adapter
19:20 orlok i have a 3 week old daughter now
19:20 orlok 3 weeks today
19:20 rue_mohr ooooh
19:20 rue_mohr how much you selling the scope for?
19:21 rue_mohr ARM9 (Hammer, Samsung S3C2410
19:21 rue_mohr so, I have an arm9
19:22 rue_mohr waaaaaaaaaaaaaaaaaaaaaait a min
19:22 rue_mohr I compiled linux for the hammer
19:22 rue_mohr Tom_itx,
19:23 Tom_itx phone
19:23 Tom_itx sry
19:23 rue_mohr what arm do you have
19:23 Tom_itx cortex 4
19:23 rue_mohr S3C2410A
19:23 rue_mohr ?
19:23 Tom_itx hold on
19:24 Tom_itx STM32F407
19:25 rue_mohr hmm
19:25 Tom_itx or F4 for short
19:25 rue_mohr I have compiled for the hammer board I have
19:25 Tom_itx where did you get one of those?
19:25 rue_mohr so the hammer tutuorials must worok for compileing arm stuff
19:25 rue_mohr long story
19:26 rue_mohr I wasn't supposed to end up paying fo rit
19:26 Tom_itx i remember sorta that
19:26 atom1 http://cu.rious.org/make/stm32f4-discovery-board-with-linux/
19:26 Tom_itx i was following that
19:28 Tom_itx and got the stlink to work
19:28 Tom_itx but i don't know how it works yet
19:28 rue_mohr I wonder if that means I can write to the arm chips on printer baords...
19:28 Tom_itx i haven't had time to read any of this yet
19:28 rue_mohr arm7 iirc
19:28 rue_mohr Tom_itx, where are you at then?
19:28 Tom_itx i have stlink compiled and you saw the output from that yesterday
19:28 rue_mohr is stlink the programmer?
19:28 Tom_itx i have some sort of arm gcc installed but not configured
19:28 rue_mohr "avrdude"?
19:28 Tom_itx basically i think
19:28 rue_mohr k
19:28 Tom_itx maybe debugger too
19:28 Tom_itx brb, spider alert
19:29 e_mohr hands tom a small jar and a 4 of hearts
19:30 Tom_itx ok
19:31 atom1 https://github.com/jsnyder/arm-eabi-toolchain/blob/master/Makefile
19:32 atom1 https://github.com/jsnyder/arm-eabi-toolchain
19:32 Tom_itx i think that's what i need
19:32 Tom_itx they call it arm-none-eabi iirc
19:33 rue_mohr I dont understand how I compiled for mine
19:33 Tom_itx there are alot of branches since this board came out i think
19:34 Tom_itx most are quite similar probably
19:34 rue_mohr there shouldbe a million arm compilers for linux
19:34 Tom_itx these samples have alot of stuff for the 'big' compilers
19:34 Tom_itx pragma stuff etc
19:34 Tom_itx i dunno all about it yet
19:34 Tom_itx but that would need to be taken care of in linux as well
19:36 Tom_itx <karlp> I got my arm toolchain from a codesourcery lite package, but should be much the same
19:37 Tom_itx he seems to be doing quite a bit with it
19:37 Tom_itx also in avr
19:39 rue_mohr http://tom-itx.dyndns.org:81/~webpage/
19:39 rue_mohr I want a standard debian pacakge
19:40 rue_mohr then I know its complete
19:40 rue_mohr your door is open tom
19:40 rue_mohr that or its 22c outside
19:40 rue_mohr its 15c inside here
19:40 Tom_itx it warmed up today
19:40 Tom_itx windows open
19:40 Tom_itx warm
19:40 rue_mohr 15c is the highest temp in the house
19:41 Tom_itx ubuntu is basically debian isn't it?
19:41 rue_mohr http://wiki.debian.org/BuildingCrossCompilers
19:41 rue_mohr yes
19:41 Tom_itx i'd go with that last link i posted
19:42 rue_mohr but your stuck, why would I follow you?
19:42 Tom_itx i'm pretty sure it's installed right, i just need to get my paths straightened out etc
19:43 Tom_itx and z gave me a makefile he was using but not set for my setup
19:43 Tom_itx so it failed
19:43 rue_mohr :( I'm so tired I dont even want supper
19:43 Tom_itx supposedly you can tie ^^ that one to eclipse or that was the initial intent i believe
19:44 Tom_itx that was his idea
19:44 Tom_itx i suppose i could install it on the server but the only link i have to it is putty etc
19:44 Tom_itx so that would end up being a pain
19:45 Tom_itx it's in the other room anyway
19:45 rue_mohr http://wiki.debian.org/BuildingCrossCompilers
19:45 Tom_itx yeah you posted that already
19:45 rue_mohr oh
19:45 rue_mohr wasn't that the other one?
19:45 Tom_itx <rue_mohr> http://wiki.debian.org/BuildingCrossCompilers
19:46 Tom_itx https://github.com/jsnyder/arm-eabi-toolchain
19:46 Tom_itx i think i may have installed that one
19:47 Tom_itx wait..
19:47 atom1 http://www.stf12.org/developers/ODeV.html
19:47 Tom_itx that's the eclipse stuff
19:48 rue_mohr half his example is arm, ther other have is ppc
19:48 Tom_itx the last one?
19:49 rue_mohr http://www.ailis.de/~k/archives/19-ARM-cross-compiling-howto.html
19:50 rue_mohr I keep nodding off tom, I thnink tonights a bad night fo me to help ya
19:51 Tom_itx that's fine
23:23 coppermine anything wrong iwth using a mini-itx motherboard for a robot?
23:23 coppermine vs a microcontroller
23:24 orlok size, power draw, IO\
23:24 orlok What devices is going robot going to use for sensors and movement, etc?
23:25 coppermine a motor controller
23:25 coppermine for the drive system
23:26 coppermine and relays
23:26 coppermine for other subsystems
23:26 coppermine via usb in terface
23:26 orlok USB interface...
23:27 orlok what sort of motor controller, and how will you interface with it?
23:27 coppermine it says in the manual
23:27 coppermine its easy
23:28 orlok You will likely find that you are spending a pile of money on a microcontroller anyway :)
23:28 coppermine microcontroller is $$$ ?
23:28 coppermine it can control a dc motor
23:28 orlok Only when its integrated into somebody elses USB control interface or motor controller
23:28 coppermine better than a relay
23:28 coppermine you can set sensitivty levels
23:28 orlok Yes, and it does that via a microcontroller.
23:28 coppermine for speeds
23:29 coppermine and the microcontroller is embedded in the motor controller
23:29 orlok What sort of motor controller?
23:29 coppermine one just for motors
23:29 coppermine nothing else
23:29 coppermine there's plenty out th ere
23:29 orlok i know.
23:29 coppermine depends on your amperage requirements
23:29 orlok I spent $120 on mine
23:29 coppermine that's about how much i will spend for all 4 of mine
23:29 coppermine for my 4WD bot
23:29 orlok handles two motors at 13A each, 12v
23:30 coppermine mine are 7A each
23:30 orlok bidirectional, plus it has a third unidirectional output
23:30 coppermine i have reverse too
23:30 orlok talks serial
23:30 coppermine mine too
23:30 coppermine but i prefer a linux cli
23:30 orlok pololu?
23:30 coppermine ya
23:30 orlok i have the trex
23:30 coppermine cool
23:30 coppermine good quality?
23:30 coppermine i didnt order yet
23:31 orlok yup, but expensive
23:31 coppermine cool
23:31 orlok the trex can be controlled via pwm or serial
23:31 coppermine im using a 20AH 12V Battery
23:31 coppermine its 14lbs
23:32 orlok i'm using cordless drill motors
23:32 coppermine i want something lighter but you cant recharge a lithium ion battery using solar
23:33 zhanx sure you can
23:33 zhanx :D
23:34 coppermine you can?
23:34 GuShH you need a charge management unit but you can.
23:34 zhanx yep
23:34 coppermine a charge controller yeah
23:34 coppermine but they are made for GEL,AGM,SLA
23:34 zhanx i have so
23:34 coppermine or wet cell
23:34 GuShH charge, discharge, etc.
23:34 coppermine havent heard for nicad
23:34 coppermine or lithium ion
23:34 zhanx i also have a 24v solar cell powering it
23:34 GuShH who uses nicad?
23:35 GuShH nicad is about the easiest chemistry to charge
23:35 GuShH and I do mean proper charge cycle.
23:35 GuShH not just a resistor.
23:35 GuShH bleh it's late, good luck.
23:36 zhanx coppermine you need to power the charge
23:36 zhanx think caps and a solar cell and go over board a bit