#robotics | Logs for 2016-05-18

Back
[02:37:05] <rue_shop5> hahah the stepper motor driver I got working can run a motor I have up to 1650 rpm
[02:37:12] <rue_shop5> SLA7044 + tiny13
[02:37:18] <rue_shop5> SWEEET
[03:20:35] <LiohAu> SpeedEvil, deshipu : actually I don't even understand why are all these switch having intensity/voltage requirements
[03:21:16] <LiohAu> because what i'm looking for is just a simple switch to cut off my battery
[03:22:15] <rue_house> I do not suggest having the saftey disconnect a wire that you try to pull from a moving robot in the event of an emergency
[03:22:19] <rue_house> that didn't go well for me
[03:23:31] <LiohAu> well right now the wheels are not touching the floor to avoid this :d
[03:23:50] <LiohAu> and it won't change until I find that switch
[03:24:48] <LiohAu> I could use jet-ski things
[03:29:10] <rue_house> a heaphone jack, with a plug that shorts R and L internal leads, use the L and R terminals as the switch leads, put a string on, it, if you pull it out it beaks the circuit
[03:29:53] <rue_house> 1/4" headphone jacks can take more current than 1/8"
[03:35:53] <LiohAu> like 50amp ?
[03:36:01] <rue_house> nope
[03:36:13] <rue_house> but a jet-ski wont either
[03:36:35] <rue_house> a 50A switch is not " just a simple switch"
[03:40:20] <LiohAu> why
[03:40:21] <LiohAu> ?
[03:43:18] <rue_house> for the same reason fire isn't cold
[03:48:17] <deshipu> damn physics
[03:48:23] <deshipu> meddling in everything
[03:53:48] <LiohAu> I understand that depending on intensity wires are bigger
[03:54:30] <LiohAu> so I guess that depending on intensity switch "poles" should be more spaced when the switch is open
[03:54:52] <LiohAu> so that leads to different switch depending on intensity
[03:55:04] <LiohAu> but then why don't I find any product matching my request?
[03:55:48] <LiohAu> am I doing something wrong by using a 24v battery with 50A ? should I go to 48v ?
[03:56:59] <LiohAu> cars, boats, there are multiple objects using low voltage batteries, and I guess they should have the same needs as me right?
[04:07:11] <rue_bed> look for a camping battery selector switch
[04:07:25] <rue_bed> its not an emergency switch tho
[04:23:56] <SpeedEvil> LiohAu: high current DC willl readily arc across contacts, and continue to arc
[04:24:12] <SpeedEvil> destroying switches that would be OK at AC
[04:30:35] <LiohAu> SpeedEvil: ok
[04:31:01] <LiohAu> so do you think i'm wrong with my voltage? I should go to 48v to get a lower current ?
[04:31:49] <LiohAu> also, when selecting that emergency switch, should I consider the peak current (50A in my case?) or the regular current (half of 50A) ?
[04:53:55] <SpeedEvil> Switches have two ratings - the interrupting rating - what current they can safely break - and the maximum current
[04:54:13] <SpeedEvil> The interrupting current is usually lots less on DC than AC
[04:54:30] <SpeedEvil> If you want the switch to be able to turn it off when running, you need to break a high load.
[04:55:06] <SpeedEvil> If you are just wanting something to make sure it doesn't move when you're working on it, then the interrupting current can be lower
[04:56:34] <SpeedEvil> For example, digikey has 14000 toggle switches.
[04:56:35] <SpeedEvil> http://www.digikey.com/product-search/en/switches/toggle-switches/1114212?
[04:56:59] <SpeedEvil> Select 'DC voltage rating' of 28V and higher.
[04:57:35] <SpeedEvil> Then look at 'current rating'
[04:57:55] <SpeedEvil> And you see the column has 50A AC/DC as one
[04:58:02] <SpeedEvil> (you only care about the DC rating)
[04:58:28] <SpeedEvil> http://www.digikey.com/product-detail/en/nkk-switches/S732/360-2095-ND/1007152 - is one of two that are rated to work
[04:58:32] <SpeedEvil> but this is $50
[04:59:13] <SpeedEvil> http://www.ebay.co.uk/itm/130952-24-VOLT-24V-UNIVERSAL-STARTER-MOTOR-BULK-HEAD-SOLENOID-LUCAS-76732-SRB321-/281941626221?hash=item41a507cd6d:g:5oQAAOSwvc9WFqXV
[05:08:47] <LiohAu> SpeedEvil: why can I select 28v and higher?
[05:37:10] <SpeedEvil> I'm assuming you don't have a 24V PSU, but a 24V nominal battery
[05:37:28] <SpeedEvil> this may be 28V fully charged
[06:04:06] <LiohAu> 27.7v right now :p
[10:30:18] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:18] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:18] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:19] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:20] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:25] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:28] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:31] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:34] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:37] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:41] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:44] <rue_house> http://www.ebay.ca/itm/12V-CAR-TRUCK-BOAT-CAMPER-BATTERY-ISOLATOR-DISCONNECT-CUT-OFF-POWER-KILL-SWITCH-/281908739725
[10:30:47] <rue_house> buy that damn thing and be done with it!
[10:34:22] <rue_house> or were all gonna die of old age before you get the first 3 wires put togethor
[10:51:40] * z64555 lols
[11:18:04] * z64555 stares at his robot code
[12:03:55] <deshipu> robot code
[12:05:09] <z64555> *not-flying robot code
[12:05:15] <z64555> sorry
[12:05:40] <deshipu> coding robot code?
[12:06:03] <deshipu> z64555: can you remind me what your project is?
[12:06:26] <z64555> AirDragon, an experimental quadrotor
[12:07:01] <deshipu> that balsa dragon fly?
[12:07:24] <z64555> nah, it's an aluminum and plastic chassis
[12:07:46] <z64555> showed a dinky video of it some time back of it playing see-saw
[12:07:56] <deshipu> gimme :)
[12:08:40] <z64555> my internet is in the shitter at the moment, will link you to it as soon as I can find it
[12:09:00] <deshipu> thanks
[12:14:56] <z64555> here we go, searched IRC backlog
[12:15:00] <z64555> mentioned it back in Feb
[12:15:01] <z64555> https://www.facebook.com/allen.babb.1/videos/200727363326050/?l=4186910995976629021
[12:15:22] <deshipu> nice rig
[12:16:33] <z64555> I put 4Kg worth of mild steel weights on the thing to try to get the center of mass directly on the test rig's pivot point
[12:16:40] <z64555> it did not help. lol
[12:18:05] <deshipu> I'm sure that enough weight will stabilize any craft
[12:21:03] <z64555> nah, if anything it just made the motors work harder
[12:21:22] <z64555> what I need to do for the test rig is to get a proper gimble cage built
[12:22:34] <deshipu> so what's experimental about that particular quadcopter?
[12:23:20] <z64555> juuuusst about everything
[12:25:30] <z64555> it was supposed to be a project for an aerial bot that would follow waypoints, but at some point in time interest shifted to getting a fuzzy logic attitude controller on it
[12:25:46] <z64555> and then later "how the heck am I supposed to get this thing working"
[12:25:58] <deshipu> I see
[12:26:27] <deshipu> I'm sure it's very educational :)
[12:26:36] <z64555> oh very. :D
[12:34:23] <veverak> great
[12:34:35] <veverak> 3d colisison library passed at school
[12:34:40] <veverak> now real problems .... :)
[12:35:09] <deshipu> congrats
[12:38:31] <veverak> given that I have two spheres with r_a,r_b and center at points A,B. Now I have defined transformations transA,trasnB where each is 'move to position X and than rotate by quaterion Q in time 't'' and are not equal, but they have same 't'. Is there sane algorithm to check if during the transition, two spheres will collide ?
[12:38:50] <veverak> because given that there is 'rotation', the 'path' of the movement of each sphere is not line...
[12:41:32] <z64555> You can discretize the path into substeps, and do a collision compare for each time
[12:42:17] <z64555> Alternatively, if both paths are linear you can formulate a time when the two will collide
[12:42:59] <z64555> wait, I'm thinking of a different problem, sorry
[12:43:20] <z64555> go with the discrete path checking for now, there might be a sneaky way of formulating collision later
[12:47:46] <veverak> yeah, I was thinking of substeps, but was curious if there is actually "proper" version :)
[12:47:59] <veverak> P.S: for now I just don't solve it
[12:48:09] <veverak> I have list of transforamtions , and just iterate over it
[12:51:33] <veverak> but, while it's not simple change, it also doesn't sound really complicated
[12:52:28] <veverak> given that I can define actual position of each sphere by parametric equation
[12:52:36] <veverak> with one parameter
[12:52:39] <veverak> it should be doable :)
[12:54:19] <z64555> yeah.... yeah. there we go
[12:54:47] <z64555> Since both objects are spheroids, you can define another parameter, d, which would be the distance between the two centers
[12:55:48] <z64555> you could then create an equation where d = to the greater of the two spheres
[12:55:58] <z64555> and then find the time that d = to that
[12:56:55] <z64555> or times
[12:57:57] <z64555> you can also apply some linear rules to it to determine if the paths are parallel, intersecting, or the same path
[13:04:41] <veverak> or just say that d = ....t....
[13:04:45] <veverak> and derivate
[13:04:47] <veverak> :)
[13:04:54] <veverak> examine maximums/minimums
[13:04:56] <veverak> and maybe... :D
[13:05:10] <veverak> nah, but ideally I would need all intervals on which sphere collide
[13:06:37] <z64555> so you'd be looking for a set of t
[13:07:03] <veverak> z64555: yeah, well, because it's boundary tree from spheres
[13:07:07] <veverak> leafs are triangles
[13:07:10] <veverak> :)
[13:07:11] <z64555> which I'm pretty sure can be done with some linear algebra magics
[13:07:24] <veverak> z64555: if spheres collide on some interval, I have to inspect that interval on childrens of that sphere
[13:08:07] <z64555> internal, or external
[13:08:18] <veverak> interval
[13:08:32] <z64555> because if internal, the method I described would yield at most one vertex for a mesh
[13:09:15] <veverak> oh, external boundary :)
[13:09:21] <z64555> yes
[13:09:22] <z64555> but!
[13:09:37] <z64555> you can iterate forward a bit to see which parts intersect
[13:10:04] * veverak is getting list
[13:10:13] <veverak> I have transition A and transition B
[13:10:22] <veverak> I can make linear interpolation between them
[13:11:13] <z64555> you got this
[13:11:18] * z64555 turns attention to other things
[13:11:30] <veverak> :D
[13:12:28] <deshipu> we are not things
[13:12:31] <deshipu> we have feelings
[13:12:35] <deshipu> *emotions*
[13:13:25] <rue_shop5> if I make hte reset pin on a tiny13 an output and put it low, will the avr reset?
[13:13:39] <deshipu> depends on the fuses
[13:13:51] <rue_shop5> no reset repurpose fuse
[13:14:00] <deshipu> then I think it will just be ignored
[13:14:03] <deshipu> but I never tried
[13:14:29] <rue_shop5> one particular peice of code is scerwing up the ability for the programmer to reset it
[13:14:58] <rue_shop5> if I scratch a wire between reset and ground when I run avrdude, it'll sometimes go
[13:16:24] <rue_shop5> the ddr keeps it an input, and if anything i'm writing 0's to it
[13:32:26] <z64555> i sometimes worry my code looks too alien...
[13:39:05] <deshipu> nah, all code in brainfuck looks like that
[13:39:27] <z64555> ...but this is in C++ D:
[13:46:28] <deshipu> z64555: hey, an article for you! http://hackaday.com/2016/05/18/flying-with-proportional-integral-derivative-control/
[13:47:24] <z64555> Thanks, will look at it
[13:49:23] <deshipu> actually, that's just basic stuff
[13:49:27] <deshipu> sorry
[13:49:29] <krisy> what is
[13:49:42] <deshipu> PID control
[13:49:49] <krisy> oh i see
[13:50:06] <krisy> fun
[13:51:05] <krisy> What are you doing with it
[13:52:35] <krisy> hello?
[13:53:14] <krisy> got to go sorry
[13:56:03] <z64555> wonder what that was about
[13:56:15] <z64555> "hi!" few seconds later "bye!"
[13:57:07] <deshipu> sounds like the new generation of engineers :P
[13:59:53] * z64555 looks at code
[14:00:03] <z64555> hm. I might actually be close to doing another test run
[14:00:36] <z64555> need to get the ADC coded, then need to get the remote control input coded
[14:01:28] <z64555> also need to buy an actual transmitter and reciever pair...
[14:04:14] <z64555> oh. still missing a few things
[14:04:16] * z64555 sighs
[14:35:07] <z64555> I've got the AHRS setup for angle output, but plum forgot about angle rate as output.
[14:35:10] <z64555> :[
[16:23:06] <z64555> hm. ok. so I can use the rate value from the kalman filter
[16:23:13] <z64555> good. I think
[16:28:17] <deshipu> veverak: got that zsun wifi card reader to work -- I have root on it :D
[16:28:35] <deshipu> veverak: it's a small openwrt computer with wifi and sub host
[16:28:40] <deshipu> veverak: for ~$10
[16:32:54] <veverak> awesome
[16:33:00] <veverak> deshipu: linkz?
[16:37:34] <deshipu> veverak: https://wiki.hackerspace.pl/projects:zsun-wifi-card-reader
[16:52:26] <SpeedEvil> $10, and 2.5 grams (for the linux bit only, without the plug)
[17:07:24] <veverak> deshipu: cool
[18:01:34] <Snert_> wow, that busybox sure is getting around.
[18:01:42] <Snert_> commercial products even.
[20:36:56] <rue_shop5> what did you find it in?
[20:37:38] <Snert> bitcoin miners and access points made by enterasys. Now extreme networks.
[20:38:32] <rue_shop5> oo that could get interseting
[20:39:28] <Snert> the wireless controllers we have are just linux boxes with a little special sauce rolled in.
[21:13:16] <z64555> hm, attitude and heading can be encoded into a unit vector
[21:13:21] <z64555> ...but what about angle rates?
[21:17:32] <z64555> it might be better at this point to convert from the unit vector into radians
[23:30:00] <rue_shop5> whats it for?
[23:31:29] <z64555> dinky not-flying robot
[23:43:19] <rue_shop5> what are the goals?
[23:45:35] <z64555> to fly without crashing into anybody
[23:45:43] <z64555> :)
[23:46:25] <z64555> I've got the tail-end of things coded to where I wanted, now I need to code on the sensory stuff
[23:47:09] <rue_shop5> to fly or not to fly
[23:47:11] <z64555> my kalman filter's set up for theta and omega (angles and angle rates) but I need to give it units that make mathematical sense