#robotics Logs

Oct 28 2015

00:00 zhanx 2048nrgb leds
00:01 zhanx With touch sensor
00:01 orlock I'm not too happy with the wiring
00:02 orlock http://en.crypt.net.au/TStar/20151027_223659.jpg
00:02 orlock i probably should have done the power point to point or something
00:02 zhanx I am not happy with a Xmas present Givin out on Monday fornem to make
01:53 Devilholk Tom_itx: Yeah, had to install 12.04 because it didn't work in later versions >.<
03:25 pzyko hello
05:23 LiohAu_ deshipu: here?
05:23 LiohAu_ you said that a strong scara arm requires "expensive servos"
05:23 LiohAu_ are you sure about that?
05:24 LiohAu_ I don't think scara arms uses servos at all.
05:24 LiohAu_ sounds like cheap scara arms uses steppers, and better ones uses dc motors (brushless for better ones).
05:31 deshipu LiohAu_: change 'servos' to 'actuators'
05:31 LiohAu_ ok
05:32 LiohAu_ I would really like to reproduce one of that arm seen on kickstarter
05:33 LiohAu_ there's a 1st thing I don't understand: most of the professional scara arms have the Z axis on the end effector.
05:33 LiohAu_ while the makerarm or flx.arm on kickstarter have the Z axis at their base.
05:35 LiohAu_ from my point of view, i think it's smarter to follow the makerarm/flx.arm way, because you don't add the weight of the Z axis actuator at the end of the arm, right?
05:42 veverak Z actuator on other design doesn't have to lift the weight of the entire arm, right ?
05:43 veverak :)
05:49 LiohAu_ veverak: in the professional ones, the Z axis only lift the end effector (and potential grasped object)
05:49 veverak yeah
05:49 LiohAu_ but you have the motor at the end of the arm.
05:50 veverak I would believe mo professional ones
05:50 veverak :)
05:50 veverak *more
09:10 aethersis hi, can somebody help me find inverse laplace transform of G(s) = 1/(1+5s)^5 ?
09:10 robopal wolframalpha?
09:11 Snert before my coffee, even?
09:13 aethersis I have to do it by hand
09:16 robopal what are you, a farmer?
09:17 aethersis a student
09:17 robopal we will never manage to colonize the galaxy if we keep doing things by hand, tell the professor thats what we got the computers now :P
09:17 robopal yeah ok I know
09:19 sherlock gotta learn how to do it by hand before a computer does it for you so you understand what the fuck it does
09:20 sherlock not so you know how to do it
09:20 rue_house aethersis, hmm
09:21 rue_house G(s) = 1/(1+5s)^5
09:21 rue_house you want f(t)?
09:23 rue_house G(s) = 1/(5s+1)^5
09:23 rue_house hmm
09:24 rue_house the power of 5 is throwing me
09:26 rue_house its not supposed to be ^2?
09:27 rue_house G(s) = 1/(5jw+1)^5
09:30 aethersis no, it's 5
09:30 rue_house what kinda loop looks like that?
09:30 aethersis no idea lol there's the function given
09:30 aethersis actually wait
09:30 aethersis it's in laplace-carson domain
09:30 aethersis so it's G(p) = 1/(1+5p)^5
09:31 aethersis so in laplace it's G(s) = 1/(s(1+5s)^5)
09:31 aethersis and it says to find step response
09:31 rue_house ah
09:32 aethersis the question is does it mean I simply have to solve it for G(s) = 1/(s(1+5s)^5) or should I multiply by 1/s lol
09:32 rue_house you messed up your brackets
09:32 aethersis because L{1(t)} = 1/s
09:32 rue_house my step fn is f(t)[u(t-a) - u(t-b)]
09:33 rue_house which isn't helping
09:33 aethersis L{1(t)} = 1/s
09:33 rue_house I havn't done this stuff since '98
09:34 aethersis so most likely it means I have to multiply
09:34 rue_house wait
09:34 rue_house G(p)
09:34 rue_house G(p) = 1/(1+5p)^5
09:34 aethersis yeah but it's not in laplace domain
09:35 rue_house so
09:35 aethersis it's in some crazy laplace-carson shit the professor who taught my professor was using
09:35 aethersis so in laplace it looks like this: G(s) = 1/(s(1+5s)^5)
09:35 rue_house G(s) = 1/s(1+5p)^5
09:35 aethersis yes
09:36 rue_house hmm, my notes aren't good enough for this one
09:37 aethersis lol
09:37 rue_house zoh...
09:37 rue_house root locus...
09:37 aethersis does this kind of thing: G(s) = 1/(1+sT) have some common name in english?
09:37 rue_house bode plot...
09:38 rue_house well
09:38 rue_house how about
09:39 rue_house G(s) = 1/(1^2+(sqrt(sT))^2)
09:39 rue_house er, well
09:39 rue_house e^(-at) = 1/(s+a)
09:40 rue_house so a = 1 and ...
09:41 rue_house G(s) = 1/(sT+1)
09:41 rue_house G(s) = 1/(Ts+1)
09:42 rue_house e^(-at) = 1/(s+a)
09:42 rue_house G(s) = 1/(Ts+1)
09:43 rue_house G(s) = 1/T(s+1/T)
09:43 rue_house G(s) = 1/T * 1/(s+1/T)
09:44 rue_house G(T) = T * e^(-1/T*T) ?
09:44 aethersis dunno xD
09:44 aethersis I think I will come to her and ask her about it
09:45 e_house snic
09:45 rue_house I'm not much help, like I say, it was forever ago
09:45 rue_house I have a book, but I'm in the middle of breakfast
09:46 rue_house aethersis, its for a feedback system then?
09:46 rue_house mechanical stuff? electronic?
09:46 aethersis the most screwed up thing is that it doesn't state what kind of system it is
09:46 aethersis it's just solve this analytically lol
09:46 aethersis screw that kind of problems!
09:47 aethersis it drives me nuts because wolfram alpha gives some crazy solution for this
09:47 rue_house aslo, learn how to ask ramstien alpha
09:47 rue_house I dont know how to get it to do that kinda problem
09:48 aethersis I know how
09:49 aethersis but it's bullshit xD
09:49 aethersis imo it's hard as hell to solve it analytically
09:49 rue_house hah, I'm gonna port my project from an arduio mini pro to a arduino uno, which means going from a mega328 to a mega328
09:49 rue_house I recon I wont have to change much
09:49 aethersis most likely not
09:50 aethersis but I've never figured arduino out, I didn't even know it comes with its own libraries, so I was programming it like an ordinary atmega controller xD
09:50 rue_house I just use them as premade avr carriers
09:50 rue_house the whole board with clock and everythign is cheaper than I can get the avr chips
09:51 rue_house almost all arduino libraries are broken somehow
09:51 rue_house people just work around the problems
09:52 aethersis I have an idea
09:52 aethersis I will fuck this stupid laplace
09:52 aethersis and solve it like any other ordinary differential equation
09:52 aethersis XD
09:52 rue_house G(s) = 1/T * 1/(s+1/T) <- I think I was close
09:53 aethersis yes but it's ^5
09:53 rue_house ignore it
09:53 rue_house wrap it with a term
09:54 rue_house if you have to, make the 1 1^5
09:56 rue_house multiply by s/s, whatever you have to do :)
09:56 aethersis thank you for your help anyway
09:56 rue_house multiply by T^4/t^4
09:56 rue_house ok
15:12 ukpa st
15:19 eedEvil sha
18:02 LiohAu what does hybrid stepper motors means?
18:02 LiohAu why "hybrid"?
18:04 veverak nah
18:04 veverak hate games
18:04 Tom_itx can be wired multiple ways
18:05 Tom_itx ie series or parallel wiring
18:16 Snert__ for unipolar versus bi-polar?
18:17 Snert__ backwards compatibility is what I was thinking for old driver circuits.
19:59 rue_house lder?
20:00 rue_house older!?
20:00 rue_house different
20:01 SpeedEvil No.
20:01 SpeedEvil Hybrid stepper means electric/gasoline.
20:01 e_house just stares at SpeedE
20:06 rue_house this mega328 cant handle this
20:07 rue_house I wonder if I have one of my mega32 baords around
22:10 anonnumberanon that glorious feel when your Neato robot-scavenged 2 packs of NiMh at 7.2V goodness respond well to charge.
22:14 anonnumberanon I'm trying to find a power supply in this room for 5V or 6V, 1.5A for a servo, so far I found an lm317...