#robotics | Logs for 2016-08-25

Back
[00:21:31] <Casper> hmmm my welding booth... need a table... grrr... money money...
[01:07:27] <anonnumberanon> rue_house, and here I thought there was a fancy thingy
[01:14:29] <The_Jester> so looks like i will be using a dedicated 7805 to run the micro servos on my hexapod
[01:18:23] <rue_house> why regulate?
[01:19:04] <The_Jester> 8.4 lipo was overheating the servos, could feel them right though the chassis..
[01:19:08] <orlock> hmm
[01:19:17] <rue_house> er, yup
[01:19:20] <rue_house> ok
[01:19:27] <rue_house> use one of the chinese dc-dc
[01:19:29] <orlock> need square M3 nuts
[01:19:33] <rue_house> much more efficient
[01:20:42] <The_Jester> and the regulator on the chineese arduino ive got on it seems to oiverheat an suffer from thermal failure which is crashing the controller
[01:21:00] <rue_house> yup, its barely good enough for the avr
[01:21:19] <The_Jester> yup
[01:21:46] <The_Jester> http://www.ebay.com/itm/Sensor-Shield-Digital-Analog-Module-Servo-Motor-For-Arduino-UNO-R3-MEGA-V5-OI-HR-/252479613511?hash=item3ac8f52a47:g:g5YAAOSwIgNXmrLn
[01:22:15] <The_Jester> was looking at those but if the v+ rails are tied to the arduino 5v line its pretty useless
[01:22:48] <rue_house> yup
[01:24:13] <The_Jester> i might get one to play with though, oven if i have to rip up traces on the underside it would be handy for this bot
[01:25:56] <The_Jester> https://pinshape.com/items/7130-3d-printed-3d-printed-exoskeleton-hands
[01:26:10] <The_Jester> inspiration for a capture glove
[01:29:26] <rue_house> yea I started to print one, REALLY SMALL HANDS
[01:35:18] <The_Jester> hmm
[01:35:38] <rue_bed> yea, it was like, damn, this is NOT gonna work
[01:35:44] <rue_bed> but I had the same thought
[01:41:25] <The_Jester> i could probley whip something off in solid works
[01:53:59] <Casper> The_Jester: have you done mechanical strength simulation in SW?
[02:02:22] <Triffid_Hunter> The_Jester: keep in mind that many regulators want ~2v between input and output, and your battery probably drops to 6v when it's empty
[02:05:39] <Casper> some is even 3V...
[02:05:43] <Casper> yeah, some manufacturer suck
[03:52:14] <z64555> bleh, got a bit of acid reflux going on
[04:48:03] <arha> i'm curious about those elevator parkings. one-two platforms go underground with cars inside
[04:48:29] <arha> what kind of hydraulic cylinder has that kind of stroke? i'm thinking 2.5-3 meters, if you stack two sedans
[04:51:02] <arha> something like this: http://www.airportcarstorage.com.au/blog/wp-content/uploads/2013/01/665021_279671328820344_1090904228_o-624x468.jpg
[08:07:30] <v555> hi, what type of tire i need for robot that has a nema 17 motors?( i want to build a draw robot that has a pen so i can programm it for drawing )
[08:09:17] <deshipu> any type would work, I think
[08:10:30] <v555> deshipu, a thick tire won't influence the accuraccy?
[08:10:43] <deshipu> of course it will
[08:10:49] <v555> oh
[08:11:13] <v555> so thiner is better in my case?
[08:16:32] <deshipu> not necessarily, because thinner will have lower friction
[08:16:42] <deshipu> and may skip
[08:17:04] <deshipu> I suppose you need to experiment and find the golden middle
[08:20:07] <z64555> robot making materials are officially on the china slow boat
[08:20:22] <z64555> sooo. I might get them sometime next month
[08:20:36] <z64555> but hey! I got a good deal on em!
[08:21:00] <v555> deshipu, yep sad thing is that it can increase the cost.:(
[08:21:38] <deshipu> well, the costs of r&d are a substantial percentage of most new projects
[08:23:21] <deshipu> z64555: what kind of robot this time?
[08:27:14] <z64555> same one I've been working on, the quadrotor
[08:27:52] <deshipu> ah, ok, thought it's a new one
[08:28:04] <deshipu> ever considered a spider? ;)
[08:28:22] <z64555> Bought an I/O breakout shield for the MSP430 launchpad, which is compatible with the Stellaris launchpad that I have
[08:29:25] <z64555> yeah, thought about doing a walker years ago, then I saw the IK formulas the college was using
[08:30:14] <deshipu> hehe
[08:30:28] <deshipu> it's not bad with just 3 degrees of freedom
[08:30:41] <deshipu> high school trigonometry
[08:32:27] <z64555> somewhat tedious, though. I'm not an ME, so I haven't been taught the ways of the quaternion
[08:32:50] <z64555> or finite element analysis, for that matter
[08:33:40] <z64555> oh, I seem to remember some crazy looking walker/flier
[08:34:42] <z64555> it was a quad limbed robot that could actuate the rotor arms as the first joint of the legs
[08:36:06] <deshipu> sounds fragile
[08:37:15] <z64555> i think it had a mechanical stop to prevent them from going above level with the deck
[08:38:08] <z64555> the propellers were kind neat, the self-folding type
[08:41:57] <deshipu> I have those on my heli, still hurts when they hit you
[08:43:03] <z64555> lol, yes
[09:51:56] <v555> can someone tell me please if these wheels can be mount to nema 17? http://www.ebay.com/itm/1pair-70mm-Rubber-Wheels-Match-360-Degree-Servo-Wheels-Accessories-For-DIY-Robot-/191796492079?hash=item2ca7f6372f:g:86AAAOSwUdlWePIL
[13:46:21] <anonnumberanon> z64555, it's time to make a massive move on your quadcopter
[13:46:25] * anonnumberanon is bored
[13:49:17] <anonnumberanon> I may spend a few hours working on job search and then try to use cortex m4 with that cortex m3 guide..
[13:58:35] <z64555> anonnumberanon: well, I did figure out where I need to put the 5V plug, so there's that
[13:59:01] <anonnumberanon> oh man..
[13:59:45] <anonnumberanon> z64555, so what cortex is your micro? m3, m4?
[13:59:45] <z64555> need jumpers to make, and also need to figure out the IDE, finally
[13:59:59] <z64555> LM4F, so an M4 with an FPU
[14:00:05] <anonnumberanon> ah okay
[14:00:08] <anonnumberanon> that' s abig one
[14:02:01] * z64555 flips on the soldering iron
[14:04:31] <anonnumberanon> man that guide is 244 pages bug i kinda feel like going through it all
[14:04:42] <anonnumberanon> i did 50 pages yesterday in the plane
[14:05:00] <anonnumberanon> z64555, what are you building today?
[14:05:21] <anonnumberanon> s/in/on
[14:05:48] <z64555> jumpers.
[14:05:55] <z64555> utterly boring stuff
[14:08:30] <z64555> kinda disappointed that there isn't a wide variety of shields for the msp430/stellaris
[14:08:41] <anonnumberanon> meh...
[14:08:52] <anonnumberanon> eventually you'll want to make your own pcb anyway
[14:09:09] <anonnumberanon> and then there will be 0 help from shields and what not
[14:09:41] <z64555> make my own pcb? no not really. although I am considering making an i2c controlled ESC
[14:09:49] <anonnumberanon> just the chip, a crystal, a few resistors, a vreg, maybe (hopefully) a USB circuit... :)
[14:10:26] <anonnumberanon> draw the pcb for the minimum stuff you need for your application, then order it to be made
[14:10:29] <z64555> that's if I wanted to make a production model, which I am not
[14:13:48] <anonnumberanon> u got a point
[14:14:44] <z64555> woo, score!
[14:27:24] <anonnumberanon> can freeRTOS be used for stuff like quadcopters and self-balancing robots?
[14:27:45] <z64555> yeah
[14:28:36] <z64555> don't think it's really needed for super simple stuff
[14:28:52] <z64555> even with quads, they usually just use a single thread
[14:29:08] <z64555> and interrupt vectors for
[14:29:34] <z64555> like PWM, some digital comms
[14:46:34] * z64555 cannabalizes some hookup wire
[15:17:25] <anonnumberanon> yeah but the interrupt stuff with micros is some of the more complicated stuff
[15:17:38] <anonnumberanon> achieving real time
[15:17:51] <anonnumberanon> imo
[15:17:55] <anonnumberanon> not an expert
[15:17:58] <anonnumberanon> yet~!
[15:23:09] <z64555> its not complicated at all, you just have to be careful and not make any blocking calls
[15:24:17] <z64555> There's a table that contains function pointers that you set to the interrupt handler
[15:24:47] <z64555> each interrupt will call a specific function from the table, if the pointer is set and the module generating the interrupt is active
[15:26:09] <z64555> some of the lower powered micros don't have a PWM module, and instead generate PWM signals on their GPIO's through an interrupt vector
[15:26:32] <z64555> which is as simple as -> reset the timer. toggle the output pin. done
[15:29:08] <z64555> for doing a communications handler, you have an output vector and an input vector. Outgoing messages go on the output vector, and incoming messages on the input vector
[15:29:59] <z64555> When it's done copying input to the input vector, you can raise a flag/mutex to signal to the main loop that there's something on the input
[15:52:34] <anonnumberanon> you are ready, padawan
[15:52:35] <anonnumberanon> lol
[15:52:45] <anonnumberanon> main{}
[15:55:28] <z64555> hm?
[15:56:09] <anonnumberanon> smallest compilable code for arm
[15:56:44] <z64555> k
[15:57:16] * z64555 stops playing with his new phone and goes back to soldering jumpers
[17:43:09] * z64555 wonders where his battery went
[17:46:27] <z64555> and my uC has power
[17:50:05] <anonnumberanon> cewl
[18:01:30] <anonnumberanon> sup? so, an LED light up yet? the micro is running?
[18:02:08] <z64555> yeah, kinda hard to see on the new photo
[18:02:19] <z64555> oh, and I deleted one of your comments, looked like it was a duplicate
[18:03:03] <anonnumberanon> oh by the way so that I wouldn't have to worry about charging the battery while doing tests, I got this to be able to run my quad from mains:
[18:03:15] <anonnumberanon> wait are you using a 2S or 3S battery?
[18:04:44] <z64555> 3S
[18:05:45] <z64555> Yeah, I probably should set up a wired power supply for extended tests as well. Just need a good way to prevent it from affecting its movement
[18:06:36] <z64555> Are you sure your power supply will give enough amps?
[18:10:18] <anonnumberanon> z64555, something like this, but my internet is crawling to a halt right now I can't find the correct sweet cheap and fast shipping model http://www.ebay.com/itm/5VDC-75W-15A-switching-power-supply-5PIN-adjustable-voltage-LED-LIGHT-METAL-/331743936811?hash=item4d3d7ac12b:g:V~wAAOSwZd1VbQem
[18:10:44] <anonnumberanon> mine was 15A I believe
[18:10:53] <anonnumberanon> or 30A
[18:14:06] <anonnumberanon> Okay I believe this is the one we got https://www.amazon.com/220V-Switching-Power-Supply-Adapter/dp/B00GWF619G/ref=sr_1_11?s=electronics&ie=UTF8&qid=1472165026&sr=1-11&keywords=100w%2Bpower%2Bsupply
[18:14:35] <anonnumberanon> I was using it for my robot after the quad.
[18:15:47] <z64555> 5V? your quad must be smaller than mine
[18:16:15] <anonnumberanon> 5V lowest setting (it's a screw), the max iirc was 15 or 30 volts
[18:16:33] <z64555> ok, that reduces the amount of current, though
[18:17:46] <anonnumberanon> I'm not sure.
[18:18:47] <anonnumberanon> it's 150W and 150 divided by 12 V is 12.5 A
[18:19:34] <anonnumberanon> because when you're testing on a bench you're sending power of usually 30% let's say your motors will take 4A, 6 A to be safe.
[18:19:34] <z64555> efficiency rating might have an effect
[18:19:48] <anonnumberanon> You're using the supply at 50% of its potential
[18:20:06] <z64555> motors consume the most ampherage when they spin up
[18:20:27] <z64555> you *might* be ok. lol
[18:20:43] <anonnumberanon> uhm I was okay the entire time
[18:21:20] <anonnumberanon> man did we have fun...
[18:21:59] <anonnumberanon> I just needed maybe 1 more month and it would have been radio-controlled.
[18:24:50] <z64555> ah.
[18:37:44] <anonnumberanon> nice progress on the quad
[18:37:51] <anonnumberanon> also i think i have the same receiver
[18:38:09] <anonnumberanon> IMU should be centered?
[18:39:22] <z64555> doesn't matter if you're not going to be doing a lot of accelerations
[18:39:32] <anonnumberanon> did you hear that sweet and dramatic sound from the ESCs?
[18:39:59] <z64555> reason you center the IMU is so that the accelerometer doesn't pick up the rotational acceleration
[18:40:13] <z64555> No, don't have the micro programmed to do that yet
[18:41:23] <anonnumberanon> that's odd
[18:41:31] <anonnumberanon> even without a signal they should beep
[18:42:25] <z64555> mine don't activate unless it's got a lock on the PWM signal
[18:43:35] <z64555> They're auto-ranging type, which means I can overclock them up to around 110Hz vs. the normal 50Hz
[18:57:48] <rue_house> $$
[19:08:21] <anonnumberanon> rue_house, are you expecting jitter from z64555 's setup ?
[19:08:24] <anonnumberanon> :)
[19:12:42] <rue_shop3> you people dont care about jitter or resolution
[20:07:06] <anonnumberanon> hear that z64555 ^ ??
[20:55:50] <rue_shop3> damnit I left a tap on and lost 400L of water from the rain barrel
[22:09:31] <z64555> hm? something about jitter?
[22:10:40] <z64555> 110Hz is well within the ESC's capability. I was able to max it out at 200~250Hz
[23:48:50] <anonnumberanon> z64555, jitter is when your command square wave vibrates
[23:49:53] <z64555> amplitude or period?
[23:50:03] <anonnumberanon> just period
[23:50:30] <Casper> frequency instability
[23:50:55] <z64555> yeah, not worried about that at all
[23:51:30] <Casper> an oscillator may be 8MHz if averaged over a few minutes, but cycle to cycle it can vary widelly, like 10%, sometime even more..
[23:53:00] <anonnumberanon> Casper, That is negligible jitter. We're talking about programming jitter...
[23:53:30] <z64555> more on the wave of discrete-time control systems, but yeah
[23:53:34] <anonnumberanon> That can make you lose a few microseconds of precision.
[23:53:39] <rue_bed> beielve it or not
[23:54:00] <Casper> same thing actually
[23:54:10] <rue_bed> I had better performance from the 8Mhz internal osc than an external crystal for my tiny26 bit-banged serial
[23:54:38] <rue_bed> which was cool, cause I needed the xtal pins for adc
[23:55:00] <rue_bed> I need the reset too, but, I'll hold off on that till I'm sure the rest is good
[23:55:08] <rue_bed> Tom_itx, have any HV programmers left?
[23:55:17] <rue_bed> or did you do a buffer?
[23:55:22] <anonnumberanon> why don't you use a chip with only 2 pins, now that would be a thing
[23:55:46] <rue_bed> changes the current draw for half duplex serial comms?
[23:56:16] <rue_bed> maybe more than 2 current levels and uses symbols
[23:56:30] <rue_bed> throw modem tech at it
[23:56:33] <Casper> it's sad that atmel used a different protocol for HV programming
[23:56:38] <rue_bed> ah
[23:57:01] <Casper> they should have kept the same protocol, but simply swapped the /reset for +12Vreset...
[23:57:01] <rue_bed> its not just picking the rst up to 12 instead of 0?
[23:57:13] <Casper> nope, different initialisation
[23:57:18] <rue_bed> oh
[23:57:24] <rue_bed> tom did hv programmers
[23:57:38] <rue_bed> I'v not seen any hv programmers online
[23:58:03] <rue_bed> I wonder , I think I have a ... dragon? I wonder if that does hv
[23:58:07] <rue_bed> never used it
[23:58:11] <rue_bed> still in the eval box
[23:58:17] <rue_bed> I wonder what else is in there?
[23:58:31] <rue_bed> ok, one last trip to the shop tonight