#robotics Logs

May 26 2015

05:06 veverak is there some opensource tool
05:06 veverak to make "thermal" analysis of part?
05:06 verak got something that cools itself at one end, and heats at ano
05:06 veverak P.S: yeha, nozzle for 3d printer
06:33 deshipu veverak: there is
06:33 deshipu veverak: openmv with the thermal shield
06:33 veverak ... :D
06:33 veverak deshipu: it's quite bad way of doing this
06:33 deshipu https://www.youtube.com/watch?v=rCJXCA8tLqw
06:33 deshipu but it's open, hey :)
06:33 veverak I need to check it's thermal properties are somehow right
06:34 veverak so I can create it
06:34 veverak or modify it for better thermal properties ;)
06:34 deshipu there are those cheap IR thermometers...
06:35 veverak http://i.imgur.com/9fX5CJr.png
08:09 deshipu http://www.cnet.com/uk/news/robot-learns-via-trial-and-error-like-a-human/
08:12 deshipu https://www.youtube.com/watch?v=oQasCj1X0e8 <-- full video of the learning process
08:42 copycat has anyone experience with using gazebo?
08:42 copycat i can't find much about it on reddit or hacker news
08:43 copycat what communities about robotic resources can i go to?
08:44 deshipu I think that gazebo is mostly an academic thing
08:44 Elleo the playerstage mailinglists would probably be the best place to go for gazebo related stuff
08:44 Elleo I used it a few years back, but I think its been largely rewritten since then
08:45 SpeedEvil I use gazebo
08:45 SpeedEvil mostly for storing gardening tools
08:52 copycat is robot simulation mostly an academic thing?
08:52 copycat so amateurs get their hands dirty and stuff?
08:55 Elleo I suspect simulation is probably more favoured by people interested in the AI side of robotics, whereas I think a lot of the people in this channel are more interested in the mechanical side of things
08:55 Elleo and you just happen to get a lot more AI type stuff going on in academia
08:55 deshipu copycat: well, the whole fun is that you get to play with actual moving things
08:56 deshipu copycat: you know, create something physical
08:56 Elleo but I come more from the AI/academic side myself, so might have a skewed view of things
08:58 copycat i see
08:58 copycat thanks for the answers
08:59 copycat what kind of robots do you guys work with?
08:59 deshipu I'm mostly building walking quadrupeds
09:00 copycat who finances the project?
09:01 deshipu myself
09:01 deshipu I go for the cheapest parts possible, and I keep it at small scale to reduce the costs
09:01 deshipu in other words, it's a hobby
09:02 copycat that's nice
09:02 copycat you don't get to satisfy such urges during work/school?
09:03 copycat i work with 2d gantry robots, i feel like a fraud everytime i tell people i work on robots
09:03 deshipu I work as a software engineer, so I don't get to play with physical objects :)
09:05 copycat cool
09:06 copycat what software runs on your quadrupeds?
09:07 deshipu I have several approaches. The most basic one just uses plain Arduino. There is one with PyBoard programmed in Micorpython, and two running Python code on Linux, one on RaspberryPi with Raspbian, and one on VoCore, with OpenWRT
09:08 deshipu Oh, and there is one that just has a bluetooth module, and the controlling code runs in Python on my PC
09:09 copycat do you have a blog or something where i can see how you chart your progress
09:09 deshipu https://hackaday.io/deshipu
09:10 deshipu I guess that's the closest to a blog that I have
09:31 SpeedEvil Simulation is really valuable.
09:31 SpeedEvil It lets you change things that you can't change easily on the robot thousands of times a second.
09:31 SpeedEvil And also evaluate control strategies in the same way
09:32 SpeedEvil see https://www.youtube.com/watch?v=Nuk3PhJzez0
09:32 SpeedEvil err - not that
09:32 SpeedEvil https://www.youtube.com/watch?v=pgaEE27nsQw
09:32 SpeedEvil that
09:34 SpeedEvil oh - he left
09:48 LiohAu_ SpeedEvil: I did not :)
09:48 SpeedEvil ah.
09:49 SpeedEvil Plus - for example - if you're using a robot to do a task - for example to construct something, it lets you do dry runs without finding that half an hour in, the elbow bangs into something and ruins it
09:56 LiohAu_ SpeedEvil: lol 2:43 :D
09:56 LiohAu_ poor dog
10:00 LiohAu_ (I was talking about the next video : https://www.youtube.com/watch?v=dRthdBr46cs )
10:02 SpeedEvil oh - neat
10:07 SpeedEvil Simulation seperates the problems.
10:07 SpeedEvil For example - the above video@4:25
10:07 SpeedEvil you're going to need a really robust robot with lightweight structure and very high performance actuators to do that.
10:07 SpeedEvil And you may break it when it hits an edge
10:08 SpeedEvil And it's got to have position sensing, object sensing, ...
10:33 deshipu SpeedEvil: I agree that simulation is very useful for coming up with new control methods and such
10:34 deshipu SpeedEvil: but most challenges of a hobby roboticist involve attaching things so that they don't fall off :P
10:36 SpeedEvil sure
10:36 SpeedEvil DEpendant on the level of your hobby
10:37 deshipu I'm still looking uphill at most stuff
10:37 deshipu but some day
14:46 AWSrobot Hey: I'm using the l298n motor driver module to power my motors
14:46 AWSrobot I hook it up to a 9V alkaline battery and it is unable to drive it
14:46 AWSrobot it outputs fraction of a volt on that motor output
14:47 AWSrobot but it's outputting larger voltages when it has no load - when it isn't connected to motor
14:50 Anniepoo is the battery nearly dead?
14:51 Anniepoo even a nearly dead battery will output it's normal voltage at no load
14:51 SpeedEvil 9V batteries have approximately no current drive
14:52 SpeedEvil if your motor wants more than about 100mA, it's not going to work at all
15:00 Anniepoo you can imagine a real battery as an ideal battery with a resistor in series. As the battery dies, that resistance grows
15:33 AWSrobot speedevil, what is "current drive"?
15:33 AWSrobot how do I find a good battery with a lot of "current drive"?
15:38 deshipu AWSrobot: alkaline batteries suck
15:38 AWSrobot Everyone tells me this
15:38 AWSrobot I believe them
15:38 AWSrobot but I don't know what to do about it
15:39 deshipu get some rechargable NiMH or NiCd batteries
15:39 AWSrobot user "SpeedEvil" in here said "9v batteries have approximately no current drive"
15:39 AWSrobot what's that even mean?
15:39 deshipu so, electric current has two values that describe it, at least for DC
15:39 deshipu voltage and power
15:40 deshipu or volatge and current
15:40 deshipu actually power is voltage times current
15:40 deshipu anyways
15:40 deshipu you have voltage and current
15:40 deshipu voltage is how quick the water in the pipe moves
15:40 deshipu current is how much water there is in a cross-section of the pipe
15:41 deshipu and power is how much water moves through the pipe per second
15:41 deshipu you can have high voltage (high speed of water) and low current (thin pipe)
15:42 deshipu or you can have low voltage (low speed) but high current (thick pipe)
15:42 deshipu with the same power
15:42 deshipu now
15:42 deshipu electric motors need current
15:42 deshipu and your battery can only provide so much of it
15:43 deshipu the motor tries to get it all, which makes the voltage drop rapidly, because there is simply not enough water
15:43 deshipu (or, in this case, electrons)
15:43 AWSrobot "watts" = "water"
15:43 AWSrobot ?
15:43 deshipu watts = power
15:44 AWSrobot in your metaphor I mean
15:44 AWSrobot ?
15:44 AWSrobot yes
15:44 AWSrobot okay
15:44 deshipu power = amount of water flowing through a piece of pipe per a time unit
15:45 AWSrobot so my 9v is able to put out a lot of volts
15:45 deshipu multiply that per time and you get total energy used
15:45 deshipu 9 of them, to be precise
15:45 AWSrobot ha
15:45 AWSrobot but has little internal capacity, low Ah?
15:45 deshipu yeah, it has low "C rating"
15:45 deshipu not capacity
15:45 AWSrobot but I was having the experience that a 9V wouldn't drive my motor *at all*
15:45 deshipu capacity is the total energy
15:45 SpeedEvil AWSrobot: 9V batteries can produce almost no current before their voltage collapses to 0.
15:46 AWSrobot I expected it to drive it momentarily before, as SpeedEvil says, the voltage collapses to 0
15:46 SpeedEvil AWSrobot: If you try to draw 1A, the voltage across most '9V' batteries will be .2V
15:46 AWSrobot SpeedEvil - I don't understand that. Why?
15:46 deshipu AWSrobot: imagine that the battery is a water tank
15:46 SpeedEvil AWSrobot: because the cells are tiny, and have little current supply ability.
15:46 SpeedEvil Why does a cardboard box flatten when you jump on it
15:46 deshipu AWSrobot: and that it has only a small hole to get the water out of it
15:47 deshipu AWSrobot: when you connect something that uses more water than can get through it, there is no stream of water anymore
15:47 deshipu this is actually where the metaphor breaks a little
15:48 deshipu AWSrobot: basically, the battery has internal resistance
15:48 deshipu AWSrobot: as the current grows, that resistance grows too, in non-linear way
15:49 deshipu AWSrobot: it's actually because there are certain chemicals depositing on the electrodes and not being removed fast enough
15:49 AWSrobot I understand it to be: it actually gives "watts," and normally it's giving 9V which gives a certain amount of watts, so if I increase the drawn amps, it proportionally decreases the drawn volts, if the amps*volts I'm pulling exceeds the watts available in this battery. Is this right?
15:50 deshipu AWSrobot: no, the battery has constant volts
15:50 deshipu AWSrobot: when you exceed the amps it can handle, the internal resistance just grows rapidly, and the voltage drops to almost nothing
15:50 AWSrobot okay.
15:50 Anniepoo no, what you say isn't right -
15:51 deshipu it's a simplification
15:51 deshipu different batteries have different ranges of volts they handle, depending on chemistry
15:51 Anniepoo ah, sorry, didn't read enough
15:51 Anniepoo nvm
15:51 deshipu but it's close enough for the alkaline batteries
15:51 deshipu AWSrobot: the 9V batteries seem to be especially bad -- have the lowest current
15:52 deshipu AWSrobot: the maximum current a battery can safely produce is measured by its "C rating"
15:52 deshipu AWSrobot: it's a multiplier of its capacity
15:53 deshipu AWSrobot: so, a 1000mAh battery with C rating of 2 can produce 2A of current
15:53 deshipu 2000mA
15:53 deshipu a 500mAh battery with C rating of 1 can produce 500mA
15:54 AWSrobot ok
15:54 deshipu some batteries have two C ratings, one for prolonged operation, and one for "burst" -- meaning that they can produce more current for a moment, but then they grow hot
15:54 deshipu (and eventually explode in flames and toxic fumes, yay)
15:56 AWSrobot so I'm looking for something like a battery pack with 6V, and since my motors "stall" at 2.0amps, and there are 2 of them, and I'm also powering my arduino and its sensors, I need a battery pack with 6V, and Ah*C rating = 5000?
15:56 AWSrobot (3v and 2.0amps per FA-130 motor = 6V and Ah*C = 5000)?
15:56 deshipu assuming your motors will be stalled most of the time, which is usually not the case
15:57 AWSrobot "stall" = max power, yes?
15:57 AWSrobot This will eat up my batteries, then?
15:57 deshipu also, to allow short bursts of higher current, you can put a large capacitor parallel to the battery
15:57 AWSrobot er, my motors too
15:57 deshipu that will prevent "brownouts" -- when the voltage drops and your arduino restarts
15:58 deshipu AWSrobot: stall = when the motor is completely blocked and doesn't turn at all
15:58 AWSrobot parallel to it? not serial?
15:58 deshipu AWSrobot: you can think about the capacitor as a buffer
15:59 deshipu it can provide very high current, but has rather limited capacity
15:59 deshipu so it would charge from the battery, and discharge when the battery's voltage drops
15:59 AWSrobot What would qualify as a "large" capacitor in this context?
16:00 deshipu 1000µF maybe?
16:00 AWSrobot ok
16:00 deshipu 500µF may be enough
16:00 deshipu depends on how large are the current spikes
16:00 AWSrobot and by parallel, you mean one line runs from the battery to the capacitor and from there to the power supply, and another line runs straight from the battery to the power supply?
16:01 deshipu no, by parallel I mean that the + of the capacitor is connected to + of the battery, and - of the capacitor to the - of the battery
16:01 deshipu also, put the capacitor as close to your arduino as possible, without any long wires
16:01 AWSrobot ok
16:01 deshipu otherwise the induction of the wires will make it a nice oscilating system
16:02 deshipu which you don't want
16:02 deshipu ah, another thing
16:02 deshipu make sure your wires are thick enough
16:02 deshipu if they are very thin, their resistance will also grow rapidly with the increase in current
16:03 AWSrobot ok
16:03 deshipu and there will be voltage drop
16:03 deshipu you can think abou it as a rubber hose
16:03 AWSrobot I didn't think this would matter over the small distances of a robot
16:03 deshipu until you only put a little water through it -- don't fill it all -- it has almost no resistance
16:04 AWSrobot I've been using the thin, flexible jumper wires that come in packages, and have maybe 2-3 inches of length to them, for almost everything
16:04 deshipu once you put more than it can hold, it will start to swell and push the water back and get hot
16:04 deshipu AWSrobot: those are good for signal, for power you either put several of them in parallel, or use something thicker
16:07 AWSrobot ok
16:08 deshipu generally it doesn't hurt to have thicker wires for power :)
16:08 deshipu unless you are constrained for weight
16:08 AWSrobot ok
16:08 deshipu it also makes sense to have the cables run from the battery to the devices separately
16:08 AWSrobot what's a good resource for finding the C-values for batteries? will they be on the data sheet or...?
16:09 deshipu then the currents flowing through those wires don't add up
16:09 deshipu AWSrobot: for LiPo batteries and the like they are always provided
16:09 deshipu AWSrobot: for alkaline batteries nobody bothers, because they have almost none
16:09 AWSrobot "almost none" = c-rating < 1?
16:10 deshipu AWSrobot: for others -- depends, there are some resources on the Internet where people do experiments and compare different batteries
16:10 deshipu AWSrobot: yeah
16:11 AWSrobot ok
16:12 AWSrobot this is for a class and it needs to be running pretty soon - seems to be sooner than I can get a new LiPo battery through ebay
16:12 AWSrobot can I use a cell phone battery?
16:12 AWSrobot what's their amp range
16:13 AWSrobot the one on my cell is "3.8V Li-Ion battery, 2100Ah"
16:17 SpeedEvil lolz
16:17 SpeedEvil Is it bigger than your leg?
16:17 SpeedEvil If so, then it's not.
16:19 AWSrobot ?
16:19 AWSrobot I see.
16:20 AWSrobot thank you
16:32 SpeedEvil Ah or mAh
18:19 veverak to which format, would you like your CAD to export technical drawings?
21:23 polthergeist Does this channel is for general robotics?
21:23 polthergeist or only arduino and 3d printed?
21:34 rue_house general robotics
23:22 zhanx eving
23:22 zhanx evening