#robotics Logs

Dec 19 2012

#robotics Calendar


01:16 rue_mohr tom
01:16 rue_mohr pricey tho
04:33 RifRaf killed me router again this arvo :/ same part as 2 months ago, the 3 phase frequency convertor that drives the spindle motor, sucks
04:34 RifRaf good that there is only 1 more work day this year, i wanted to make some robot bits in the holidays though :/
04:41 RifRaf i have this 4000w(8kw peak) invertor thats not working at capacity and the alarm sounds the whole time
04:43 RifRaf have opened it, is like 8 big transformers, 90'odd irf 120a fets and and buch of caps that look ok, where would i start looking for a fault, the 40 or so diodes are about the only place i cen get a decent reading at many similar points
04:46 RifRaf rue_mohr i bet you could find the fault quick but should i just desolder the fets and stuff and salvage
04:48 theBear yo ! yer measured across gate resistors while powered ? any measurable drop indicates a leaky fet, which among other things will screw up under load
04:55 theBear i suppose if it's 9 transformers, you can look for one that isn't being driven properly, compare input waveforms
04:55 RifRaf hello there thebear
04:56 theBear how yer doing ? i'm older, more bitter, an ex cripple pensioner, and never able to lift anything heavy again :)
04:56 theBear plus i was sober for a month or so, it was horrible, i can't recommend it
04:56 RifRaf i have still got to remove anothe 40 screws on the other side to unmount all the heats sinks, is gonna be a big reassembly job
04:58 RifRaf i saw you had a rough trot there for a bit, good to see your still here, i have just moved into new joint
04:58 theBear mmm fancy, still nice and private ?
05:00 RifRaf sorta, not quite as private but only about 600meters from me old place and its all normal like, there will soon be solar panels, robots and other things around the place
05:01 theBear fancy, and very nice
05:03 RifRaf so measure across the gate resitors eh, there are rows of resistors on one side of the pcb
05:05 RifRaf 8 x maybe 5w resostrs on each side of the unit in a row, could these be them, will put on me glasses and find some colours for a value
05:06 RifRaf red red blue gold
05:06 theBear probly not, usually 5w's are balancing resistors on source/drain, gate resistors will be 1/4w usually, 'cos they only ever see current when a fet is damaged
05:06 theBear err, that's BIG, that's like, 220meg ?
05:06 theBear or 22maybe
05:06 RifRaf nah mus be 22 something though
05:07 RifRaf ok there are small tiny resitors as well, might need me big glasses for that one
05:07 theBear can you see the pcb side, or just do resistance checks maybe
05:08 RifRaf i
05:08 RifRaf d say brown black black gold
05:09 theBear mm, that's in the ballpark, anything from maybe 100 to maybe 10k or so is possible
05:10 theBear and they'll go to one leg of every fet, and most of them should measure 0.000v when the thing is powered, regardless of what the fet is doing at that time
05:10 RifRaf ok so tomorrow when i an sober in the morning i will try measure a voltage drop across a bunch and see what i get
05:10 RifRaf tonight i will likely fry myself on the big wires in there
05:11 theBear maybe :)
05:12 theBear was in forster few months ago for a wedding... closest i been to yer for ages, nice little town
05:12 RifRaf is a chucky think, i reckon you could make some decent motor controllers with the remaining good fets
05:12 theBear no doubt, lots of them
05:13 RifRaf i so rarely go south these days as well, but if you ever make it this far north yer welcome to drop in, still many mad times
05:14 theBear ya know i'm still on the other side right ? but yeah, if i'm ever in control of the transport, yer on the list
05:14 RifRaf am so enjoying being in me own space again looking at the crap i used to make
05:14 theBear heh
05:15 RifRaf ah so still in wa, one day i'll make it that way
05:15 theBear meh, it's not special :)
05:18 RifRaf seems like the good money is still over that way thougt, a few mates still trying to get there
05:18 theBear mmm, still a lotta mining, but it's slowed down a shitload the last 6 months or so
05:19 theBear bunch of people i know come back to town instead
05:20 RifRaf yep just been happy doin me sign making thing these days, easier than computers and looking forward to using the tools for make something
05:20 RifRaf but i killed me router again today :(
05:20 theBear mmm
05:21 RifRaf 5 minutes before knockoff, second last day of the year, so alot of cleaning tomorrow for me
05:22 RifRaf and hope they get the parts from wa before next year
05:23 RifRaf i so wanted to start making replacement parts for foxy in the holidays but will have to wait now
05:25 theBear mmm... this place aint the same without a foxy
05:25 RifRaf i found the bits of her brain while unpacking, it will be a long sober night trying to fix it i think
05:27 RifRaf plus no real internet here, just wifi
05:29 RifRaf theBear good to catch up, one more real day of work and then might be around a bit while not out fishin
05:29 theBear mmm, nice to see yer around :) i'm usually here when i aint at work...
07:19 Tom_itx some cnc's don't like phase inverters
07:20 Tom_itx add a cap to it
08:17 anannie Does anyone know where I can learn more about PIDs? I've googled and read through the pages, but what I'm looking for is a good source that explains what's going on so to speak and how do I take the integral or the derivative of what I'm looking for
08:34 Triffid_Hunter anannie: http://brettbeauregard.com/blog/2011/04/improving-the-beginners-pid-introduction/
08:34 anannie thank you Triffid_Hunter, most PID tutorials don't go into the math itself and just handwave it away
08:36 anannie so we are taking the actual literal meaning of integration and differentiation, finding the change in value and the area underneath the curve... That's interesting.
08:37 Triffid_Hunter yep
08:37 Triffid_Hunter how fast are we approaching the setpoint, and how long have we been some distance from it
08:38 anannie Can I read an explanation about how this was derived so that I can get an intuition for it?
08:40 Triffid_Hunter anannie: it's obvious when you think about it.. when you're driving a car and you want to stop at a stop sign, you can't just stomp the brake when you get there.. instead you have to judge how quickly you're approaching it and reduce velocity early
08:40 Triffid_Hunter anannie: that's derivative, the D term
08:40 Triffid_Hunter anannie: if you stop a bit too early, then you have to see that you're too far from the line and not moving, so you creep up until you're happy with your position. that's what the I term does
08:40 anannie I then notice over time how the thing is changing, so I then make adjustments accordingly
08:41 Triffid_Hunter anannie: when all 3 terms are tuned properly, you can stop your car exactly on the line every time
08:41 Triffid_Hunter of course, it's not going to work in your friend's volvo or your boss's porsche, you have to retune if the dynamics change
08:42 Triffid_Hunter anannie: PID is not some magical panacea, you have to tune it for each situation where it's used
08:42 anannie It makes sense that way, but what I want to do is develop an intuition for it so that it's intuitive for me trii
08:42 anannie * Triffid_Hunter
08:43 Triffid_Hunter anannie: and it's actually relatively poor (but better than nothing) for some systems eg asymmetrical stuff like high-powered heaters that heat up very quickly but drop only slowly
08:43 Triffid_Hunter anannie: hence the car analogy
08:43 annie
08:44 Triffid_Hunter well good luck, I'm off to bed :)
08:44 anannie Is there a book you recommend where I can grasp the theory behind things?
08:44 anannie Thank you Triffid_Hunter
08:44 Triffid_Hunter anannie: keep looking at different tutorials until it clicks for you
08:45 anannie I'll write my own code for it as well and try to experiment with it to develop a line follower
08:46 Triffid_Hunter yeah good plan, line followers are twitchy as hell without some filtering
08:46 Triffid_Hunter also google kalman filter although that one's probably a bit heavy for a line follower
08:46 Triffid_Hunter really gone this time, good luck!
08:46 anannie bye!
08:46 anannie See you later Triffid_Hunter
09:00 rue_mohr http://www.ebay.ca/itm/Waterproof-DOOR-LOCK-ENTRY-ACCESS-CONTROL-KEYPAD-BACK-LIGHT-/261081929143
09:01 rue_mohr how could anyone not make money at $37?
09:01 rue_mohr and thats cheap, they go up over $250
09:21 elektrinis so is anyone using solidcam?
09:23 rue_mohr what does it cost?
09:23 elektrinis many dollart
09:24 elektrinis haven't bought yet, just evaluating a demo
09:24 elektrinis they give only 1 month and time flies
09:24 elektrinis 2 weeks past and still can't machine a very simple part
09:24 elektrinis https://dl.dropbox.com/u/68288580/cooler.jpg
09:24 rue_mohr if someone made a program they wanted people to use instead of making their own, they shouldn't ask money for it
09:26 rue_mohr why would you do that to a solid block of metal, have you not heard there is an environmental and energy crisis?
09:26 rue_mohr :)
09:27 elektrinis trolls everywhere
09:27 rue_mohr hey, I'm one of the master trolls
09:28 rue_mohr ERP who is ate for leaving for work!
09:36 elektrinis derp?
12:35 RazWelles ping, is anyone here doing uni research with robots, personalities, and/or HCI?
14:29 mac- hey
14:29 mac- could you help me to understand equation 5.3
14:29 mac- based on 5.1 and 5.2 ?
14:29 mac- http://www.robotyka.com/teoria.php/teoria.44
14:29 mac- from the link
14:29 mac- ?
14:42 mac- anyone ?
19:53 zhanx why does every computer I have brake at the same time?
19:53 zhanx break even
20:01 goRed sh
20:01 theBear thats how computers work
20:02 zhanx all 4 died at once, epic fun. at least I have spare parts
20:02 zhanx the hd failed out of the blue on my laptop
20:03 zhanx 1 desktop, motherboard is toast, 1 the psu died. the other I have yet to get to... moving sucks
20:06 Tom_itx it only gets worse
20:07 zhanx the fact my raid failed don't help also
23:47 rue_mohr hi
23:53 KimLaroux good morning
23:54 rue_mohr good afternoon
23:56 KimLaroux not working tomorrow?
23:56 KimLaroux Oh wait, you're on the west coast... so you're 2 hours earlier, right?