#robotics Logs

Sep 05 2015

00:06 rue_shop3 yea
00:07 rue_shop3 split it and put a spring between the two peices
00:07 rue_shop3 take apart a cd player that uses a rack and pinion head drive
05:54 deshipu https://cdn.hackaday.io/images/143291441447673213.jpg
06:10 robopal what is it?
06:10 robopal bananameter?
06:25 deshipu bananabomb
06:25 deshipu I need to add red and green wires
06:26 robopal and blue
06:33 deshipu blue?
06:38 robopal https://www.youtube.com/watch?v=uMQrloEFio4
06:48 deshipu hmm, ok, I remembered wrong, I thought it was green
07:00 deshipu ok, replaced with blue
08:46 nfk i'm looking into IDE/tools for robot design and simulation and feel a bit overwhelmed by the bunch of options i found, any recommendations?
10:22 nfk after more detailed examination, is there anything comparable to coppelia but that's fully under free software licence?
10:22 nfk gazebo seems similar but also less GUI oriented
10:32 Hyratel brother ordered a new USB port for his phone - it came with a set of screwdrivers and plastic prybars
10:33 theBear we don't use those things... maybe they are aimed at little kids and fools, maybe the concept doesn't work for anything else than that or extremely specific setups, like a dept within a much larger robot designing/building company
10:33 rue_house a green bannana timer?
10:33 theBear Hyratel, no soldering iron ? that gonna be tricky
10:34 Hyratel har. i have one of my own
10:35 theBear hope yer almost half as good as me then, most people ain't real good soldering something huge like a microusb, only they call it "tiny" instead
10:35 rue_house no guitar pick?
10:36 Hyratel god, this is gonna be FIDDLY
10:36 Hyratel theBear, and it's not even the decently-large wires of a normal micro USB
10:40 rue_house according to the directions your supposed to throw the thing out and buy a new one
10:41 rue_house nfk, you want to do robot simulation?
10:43 nfk rue_house, design
10:43 nfk and i'd love to be able to simulate/text how well they work without building them
10:43 nfk the only 3d printer at uni is from what i have heard from my classmates who worked with it last year POS
10:44 ace4016 most 3d printers are PoS
10:44 ace4016 but i guess they're good enough (tm)
10:44 nfk yeah, but that one is special, it failed to even print out a basic arduino sized matchbox
10:44 ace4016 heh
10:46 nfk and if the question was if i'm a total noob when it comes to robotics then the answer is absolutely
10:47 ace4016 i kind of fear the newbies will disregard subtractive manufacturing for their parts...
10:47 nfk substractive?
10:47 ace4016 milling, cutting, drilling, etc.
10:47 ace4016 the "normal" stuff
10:47 nfk i though the point of 3d printing was to already design it with holes where needed
10:48 nfk and cnc workbenches are rare here
10:48 ace4016 kinda
10:48 nfk and people who actually understand how to use them are even rarer
10:49 ace4016 additive manufacturing can create more complex shapes and can be cheaper to make prototypes for very custom designs
10:49 ace4016 but it's always been slow and lacking in tolerance/resolution (may change in the future)
10:49 nfk besides if the currently ephemeral project is greenlit the final product is almost certainly gonna need machining anyway since I do not expect plastics to actually handle the load
10:50 ace4016 hehe
10:50 nfk anyway, any recommendations for a toolchain from design to simulation?
10:51 ace4016 nay on the simulation (if logic)
10:51 rue_house you know that you cant design in a simulator and then expect soemthing to work in real world, right?
10:51 ace4016 most CAD tools that do dynamics are good enough for the mechanical properties of a robot though
10:51 nfk i'd very much appreciate it as before greenlitting the project it's gonna need some formal proof that it's even worth trying to build
10:52 ace4016 what are you trying to simulate btw exactly?
10:52 nfk secret for now
10:52 nfk and i doubt it's gonna get any more public before it's published not this year if at all
10:52 ace4016 well, i meant in generic, high level terms
10:52 nfk a robot
10:52 nfk doing stuff
10:52 nfk in a semi-controlled environment
10:53 ace4016 are you trying to simulate the logic and how it behaves in its intended environment?
10:53 nfk from algorithms to mechanics
10:53 nfk wear probably does not need to be simulated
10:53 ace4016 how complex is the environment? normal air with obstacles and such? underwater? space? atmospheric flying?
10:54 nfk nothing that crazy
10:54 ace4016 also, just an FYI about engineeing options: simulating is one of the most expensive methods of designing :P
10:54 nfk as for obstacles and exact terrain i'll know in a short while
10:54 nfk what else can i use to verify that my ideas are sane?
10:56 ace4016 cheap prototype; partitioned design boards/parts (taking the components and prototyping them seperately); feasibility analysis (usually taking things that already exist out there and seeing how well they perform and what you'll need to "overcome" to meet your design requirements)
10:57 ace4016 the partitioned prototyping also shows how well you can componentize your designs :P
10:57 nfk also i'm not sure about funding in the initial stages so that's another reason why i'd like to avoid having to build something at first
10:57 ace4016 so that you can swap and replace and such
10:57 nfk i have worked with android before so i kina know i'd very much not want to fund even a basic protype myself
10:58 ace4016 as i don't know a hint of the scale you're working with; can't say
10:58 nfk buying a chinese quadcopter kit would probably be cheaper than even a half working prototype
10:59 nfk in fact, i suspect the required logic boards to power it would cost more than the cheapest quadcopter kit
10:59 ace4016 but if you do go the simulation route, most of your time will be setting up the physical environment and figuring out schemes to simulate the sensor inputs. actuation is REALLY easy in simulation though
10:59 nfk s/power/run/
10:59 ace4016 and considering that a computer is running the sim, the logic bit isn't bad
11:02 nfk ace4016, i'll change the question, how would you go about designing a robot?
11:02 nfk e.g. with what would you start and what tools you'd use?
11:10 ace4016 first, i'd collect the design parameters (e.g. requirements, needs, etc.)
11:10 ace4016 i'd do a risk and feasibility analysis on each breakdown of said requirements
11:11 ace4016 then, based on budget, schedule, and funding, i'd figure out a way to design the various needed components
11:12 ace4016 some things can be done with some paper and pen analysis (or matlab or what have you), some are easier to build the little circuit/actuator/part, and some may require i simulate
11:13 ace4016 simulating seems like the best option, but you waste a lot of time (schedule) on setting up the environment, and as rue_house mentioned, what you see in the sim may not match what happens in real life at all
11:13 ace4016 but, it's definitely an option
11:14 ace4016 as i don't know what you're building, what scale and complexity it is, or anything, i couldn't say what you could break down into easy prototypes and sims and such
11:15 ace4016 tbh, if you want to simulate a physical world and have an actor behave in that world, you could pick up a free game engine and it would meet your needs for a sim, assuming basic physical complexity
12:21 theBear Hyratel, quit being a pussy ! try a mfc connector, same width but 11 oddly spaced conductors.. a disposable straight razor blade literally does not fit between the closest sets of pins
12:45 theBear cheer up, it's just a saying
12:45 theBear i don't htink you're a cat, not really
14:45 deshipu what if he *is* a cat
14:45 deshipu on the internet nobody knows you are a fridge