Back
[00:00:38] <rue_shop4> gpio
[00:17:55] <z64555> really?
[00:18:12] <rue_shop4> yup
[00:18:21] <z64555> why? timer doesn't consume as much clock cycles
[00:18:38] <rue_shop4> I suppose every io op takes 7 cycles, so the delay isn't as much as I thought it might be
[00:19:09] <rue_shop4> which is a good call to have a subsystem do any io, not tie up the processor
[00:21:10] <z64555> yes.
[00:21:25] <rue_shop4> I recond the dma can prolly be used for io
[00:21:43] <Jak_o_Shadows> Yeah, I've seen an example that does it
[01:37:27] <pdgeorge> Not an active chat?
[01:38:25] <rue_shop4> hu?
[01:38:31] <pdgeorge> nm
[01:38:34] <rue_shop4> building robots
[01:38:43] <rue_shop4> BUILD MORE ROBOTS
[01:38:45] <Jak_o_Shadows> BTW rue: The ##stm32 channel gets distracted a lot. The couple of times I've asked in #libopencm3 (which I realise you're probably going to program just in registers, etc) has been more helpful
[01:38:50] <orlock> OK!
[01:38:58] <rue_shop4> what sense resistor should I use?
[01:39:17] <orlock> Rue:
http://www.gl-inet.com/product/gl-mt300n-engineering-board/
[01:39:18] <pdgeorge> I'm in a 'building robots' limbo atm.
[01:39:26] <Jak_o_Shadows> The pretty one
[01:39:40] <orlock> Building a 300W sumo bot here
[01:39:43] <rue_shop4> I'm working on a force feedback servo board, 8 channel
[01:39:49] <pdgeorge> I've been asked "What would your ideal honours project be" and I'm stuck. I literally can't think of one... I've done so many robotics projects and nothing stands out
[01:40:08] <rue_shop4> just make a robot arm and have it do a canned cycle
[01:40:24] <pdgeorge> ew...
[01:40:26] <rue_shop4> orlock, hu? why would I get that?
[01:40:52] <Jak_o_Shadows> pdgeorge, the thing with honours projects is that they are ideally meant to be new research.
[01:41:09] <Jak_o_Shadows> Your cohort won't recognise the challenges in something that been built by lots of people.
[01:41:18] <rue_shop4> oh, then you can get points for making soemthing that dosn't work
[01:41:35] <Jak_o_Shadows> So robot arms & robot hands are out. Unless you do the maths behind workspace exploration. That'd be a good project.
[01:41:45] <rue_shop4> as long as the concept is presented as being unbelievably amazing
[01:42:12] <pdgeorge> Jak_o_Shadows: Yeah, that is what I am having trouble with. I've already been involved in one research thing (Royal Children's Hospital + Nao robots research thing) but it wasn't for honours or anything. I was just involved heavily.
[01:42:21] <rue_shop4> does it have to be mechanical?
[01:42:29] <z64555> doing off and on production of a quadrotor, if anybody asked
[01:42:32] <rue_shop4> can it be software based?
[01:42:33] <orlock> ahh
[01:42:34] <Jak_o_Shadows> "A theoretical approach to modelling the workspace of a serial manipulator using blargh". Swap out serial with parallel for more fun.
[01:42:37] <orlock> AUSSIE AUSSIE AUSSIE!
[01:42:45] <rue_shop4> yea, swarm bots are always a nice topic
[01:43:07] <pdgeorge> Doing robotics and mechatronics + computer sciences. So theoretically should be both. But doesn't matter if it's more heavily on the software side
[01:43:20] <orlock> pdgeorge: Which state?
[01:43:31] <pdgeorge> orlock: Australia, Vic
[01:43:32] <rue_shop4> if I make my driver cablable of driving in both directions and freewheeling and braking, when do I use what...
[01:43:44] <Jak_o_Shadows> Oooh, ooh. I got an idea: Forward kinematics of a stewart-gough platform using rotary position sensors
[01:43:58] <Jak_o_Shadows> So you get to build one, but also model it
[01:44:00] <rue_shop4> pdgeorge, cause I got some voice recognition stuff
[01:44:10] <orlock> pdgeorge: Unsure if we have spoken before.. But i work for a company in that location that specialises in making biotech robots
[01:44:26] <z64555> surgeon hands?
[01:44:36] <rue_shop4> oops, glitch
[01:44:49] <rue_shop4> I'm gonna make a force feedback robot arm with this board
[01:44:50] <z64555> it wasn't me
[01:44:55] <rue_shop4> if I can work out the motor driver...
[01:45:09] <rue_shop4> the size up which just went up 8fold
[01:45:17] <z64555> force feedback, so the thing vibrates whenever it hits something?
[01:45:19] <pdgeorge> orlock: probably not sadly. I've got a red streak in my hair to make me more memorable so maybe?
[01:45:37] <z64555> sorry, been messing with a lot of joystick io
[01:45:53] <rue_shop4> somethnig that tunes its vibrations to vocal noises?
[01:46:10] <rue_shop4> dynamically
[01:46:57] <rue_shop4> is it a technical school or university?
[01:47:42] <pdgeorge> I'm at a university.
[01:47:43] <rue_shop4> ARG, do I have the driver switch from drive to coast or drive to brake...
[01:48:02] <rue_shop4> oh, your screwed, if your lucky they thank you for your money as they kick you out the door
[01:48:21] <Jak_o_Shadows> rue, our Uni's aren't quite so money hungry yet :P
[01:48:23] <rue_shop4> make a burger flipping robot, its likley to be usefull in a career
[01:48:30] <pdgeorge> I've done technical too. (TAFE)
[01:48:33] <rue_shop4> ok
[01:48:51] <pdgeorge> For a final year project in TAFE I made a car that can be controlled over the internet.
[01:49:15] <rue_shop4> I need a robot that can stream video of a fire panel to me, and push the reset button when I ask it to, cause we cant keep employees at work
[01:49:18] <pdgeorge> with a reverse sensor that automatically stopped when reversing if something got too close behind it.
[01:49:42] <rue_shop4> I'm going to base it on a coat rack
[01:49:48] <pdgeorge> That should be VERY easy
[01:49:53] <rue_shop4> yea
[01:50:08] <rue_shop4> I think even basic opengl can keep track of where a button is
[01:50:30] <rue_shop4> BUT SHOULD MY DRIVER DRIVE/BRAKE OR DRIVE/COAST!!!!!!!???????
[01:50:40] <rue_shop4> god damned tristate systems
[01:50:43] <pdgeorge> I mean... there's two ways to press the reset button. Literally press it (have a 'hand' that presses the button) or re-wire the button so that it can be electronically pressed by a gpio pin of an external computer.
[01:51:04] <rue_shop4> cant rewire each fire panel I work on
[01:51:29] <rue_shop4> and modifying fire panels for annual inspections is sorta against the rules
[01:55:00] <rue_shop4> oh, If I put the current sense resistors like that... I can do soft AND hard braking...
[01:57:02] <rue_shop4> ok, all H bridge combinations result in 7 power supply shorting methods, 5 coasting methods, 2 braking methods, and 2 drive methods
[01:57:31] <rue_shop4> and I can make the braking into 1 hard and 1 soft
[01:57:52] <rue_shop4> the uc will have 1 pin for pwm, 1 for direction...
[01:58:08] <z64555> sounds like you're overthinking a tad
[01:58:20] <rue_shop4> I ALWAYS OVERTHINK
[01:58:40] <rue_shop4> I potentially have 3 damping levels
[01:58:57] <z64555> so the question is what do you need
[01:59:07] <rue_shop4> generic position control
[01:59:10] <z64555> *what do you need right now
[01:59:27] <rue_shop4> generic position control
[01:59:35] <z64555> ok, then pick the easiest one that'll accomplish that
[01:59:36] <rue_shop4> with force feedback
[02:00:41] <rue_shop4> I have 16 pwm channels
[02:00:45] <rue_shop4> I have 8 motors
[02:01:26] <rue_shop4> and atleast 8 io
[02:01:36] <rue_shop4> 16 adc channels
[02:01:43] <rue_shop4> well 18, but 16
[02:02:03] <rue_shop4> (position, current)
[02:02:23] <rue_shop4> my goal is to set up a serially conveyed force feedback servo demo
[02:02:36] <rue_shop4> I'v gutted a few cheap servos
[02:03:03] <rue_shop4> servo ----- position ---> <---- current ----- servo
[02:03:49] <z64555> I'm sorry, but I can't be of much help right now. that's a bit over my head while half-asleep
[02:03:51] <z64555> :(
[02:04:00] <rue_shop4> its just me
[02:04:07] <rue_shop4> I walk a lonley road
[02:04:16] <rue_shop4> THAT NEEDS MORE TEA
[02:04:30] <z64555> Lone Road Tea is a good company name
[02:07:20] <rue_shop4> I guess I should see how fast a max232 can switch a fet
[02:07:59] <rue_shop4> oh damn
[02:08:14] <rue_shop4> I forgot two channels are the wrong direction
[02:08:23] * rue_shop4 grumbles
[02:09:24] <rue_shop4> if I have a mosfet totem pole thats 6V apart with connected gates, what is the swtiching overlap
[02:09:41] <rue_shop4> I think its 2V
[02:10:09] <rue_shop4> can I just put an inductor on the power feed to eat the spike?
[02:11:07] <rue_shop4> max232 will generate -10V, at 6V, thats 16V which isn't over 20, so thats ok
[02:11:29] <rue_shop4> wait wait wait
[02:11:43] <rue_shop4> I can put the P on the bottom and the N on the top
[02:12:04] <rue_shop4> cause the 232 will generate voltages that can still drive them between those rails
[02:12:51] <rue_shop4> then there is 0 corssover
[02:12:54] <rue_shop4> crossover
[02:14:13] <rue_shop4> hmm thats only 4V drive on the N channel, but I'm using a logic level for that, so...
[02:16:21] <rue_shop4> says 60A
[02:18:11] <rue_shop4> .1V drop
[02:18:22] <rue_shop4> this is crazy, but I think it works
[02:18:26] <rue_shop4> no drive overlap
[02:19:18] <rue_shop4> so, what size current sense do I use
[02:19:36] <rue_shop4> whats the internal vref on a avr
[02:19:48] <rue_shop4> 1.1V
[02:20:08] <rue_shop4> so, say 2A, at 1V
[02:20:24] <rue_shop4> 0.5ohms
[02:20:27] <rue_shop4> hmm
[02:20:30] <rue_shop4> k
[02:20:44] <rue_shop4> root@blackie2:/files/objects/downloaded/Human_Hand_Bones_-_Thumb/files# ohm -i 2 -v 1
[02:20:44] <rue_shop4> Wattage is: 2.000000
[02:20:44] <rue_shop4> Current is: 2.000000
[02:20:44] <rue_shop4> Voltage is: 1.000000
[02:20:44] <rue_shop4> Resistance is : 0.500000
[02:21:10] <rue_shop4> gonna need 1/2W resisotrs min
[02:21:12] <rue_shop4> that sucks
[02:22:28] <rue_shop4> so 2.2A full scale
[02:23:25] <rue_shop4> iir ci got a ton of 0.39 and 0.47 ohm resisotrs
[02:30:27] <rue_shop4> I got a lot of 2W 0.5R
[02:30:44] <rue_shop4> that takes up to 2A
[02:31:12] <rue_shop4> I got some 1/8w 0.1R
[02:31:39] <rue_shop4> root@blackie2:/files/objects/downloaded/Human_Hand_Bones_-_Thumb/files# ohm -r 0.1 -p 0.125
[02:31:40] <rue_shop4> Wattage is: 0.125000
[02:31:40] <rue_shop4> Current is: 1.118034
[02:31:40] <rue_shop4> Voltage is: 0.111803
[02:31:40] <rue_shop4> Resistance is : 0.100000
[02:31:50] <rue_shop4> not enough votlage for the avr adc tho
[02:32:30] <rue_shop4> guh, these drivers are big
[02:33:39] <rue_shop4> 4 to-220 transistors, a 2w resistor, and a max232, along with 4 diodes, and the 5 caps for the max232
[02:34:20] <rue_shop4> I wonder if I should just let the body diode of the fets take the surge
[02:34:25] <rue_shop4> its 2A at worst
[02:36:16] <rue_shop4> hmm forward voltage of 2.5V
[02:36:26] <rue_shop4> at 6V, that only leaves 1V across the load
[02:36:35] <rue_shop4> thats 2w
[02:36:36] <rue_shop4> ugh
[02:40:57] <rue_shop4> maybe I should make my own charge pump
[02:40:59] <rue_shop4> ugh
[02:41:49] <rue_shop4> how many channel on a 1488 or 1489?
[02:42:14] <rue_shop4> 1488 is the driver
[02:42:22] <rue_shop4> 10mA limited
[02:42:32] <rue_shop4> 4 cahnnels
[02:45:08] <rue_shop4> max230 would be more ideal
[02:46:35] <rue_shop4> haha $2.50ea FORGET THAT
[03:49:17] <rue_shop4> ok, it works
[03:49:42] <rue_shop4> 2us load switching
[03:49:59] <rue_shop4> dosn't appear to be any fet overlap
[03:54:12] <rue_shop4> I see some 10mA spikes, but I think they are the gate switching charge
[04:05:01] <rue_shop4> hmm, the P channel fet still isn't performing as well
[04:05:13] <rue_shop4> 0.6V opposed to the N channel at 0.1V
[04:06:01] <rue_shop4> which is odd, its got a better drive
[04:06:06] <rue_shop4> -11V
[04:06:38] <rue_shop4> yea, it should do better than that
[04:07:12] <zzz64> I seem to remember something about P and N channel configurations having different performances
[04:07:57] <zzz64> maybe check their data sheets to see if they're different?
[04:08:05] <zzz64> (aside from the obvious)
[04:13:24] <rue_house> they are, but according to the irf4905 datasheet, the voltage drop with -11V on the gate should be lots less than 0.6V at less than an amp
[04:15:14] <zzz64> what's your biasing circuit look like?
[04:17:06] <rue_house> its a max232 :)
[04:17:15] <rue_house> thats hacked to output +-11V
[04:17:34] <rue_house> it can drive the fets to 20Khz no problem
[04:17:54] <rue_house> I only plan to use a few hundred hz for the motors
[04:18:00] <rue_house> maybe 440, or 880
[04:18:50] <rue_house> maybe 554.37, as a tech joke
[04:21:13] <zzz64> huh, that chip is as old as I am
[04:21:21] <zzz64> well, chip line
[04:21:25] <rue_house> yup
[04:21:29] <rue_house> great fet driver
[04:21:41] <rue_house> the 555 and LM386 are too
[04:22:39] <zzz64> neat
[04:23:16] <rue_house> I suppose that I'll get symmetric operation of the motor when I have a full H bridge, right now it drives one way better than the other cause its a half bridge
[04:23:27] <rue_house> means I lose 0.7V total
[04:24:06] <rue_house> whats the max operating voltage for a HS3003?
[04:25:21] <rue_house> 6
[04:25:24] <zzz64> eh, no idea. Google doesn't know what it is
[04:25:42] <rue_house> S3003
[04:26:04] <zzz64> servo?
[04:26:30] <rue_house> so, I might make a modified regulator,... or just use a switching one adjusted to 6.5 or so volts
[04:26:31] <rue_house> yea
[04:26:36] <zzz64> yeah, it says 4.8 to 6V
[04:26:42] <rue_house> hmmm
[04:27:03] <zzz64> nominal voltage most likely 5V
[04:27:41] <rue_house> and I'm losing 0.7V on the fets, and another 0.5-1V on the current sense
[04:27:51] <rue_house> so, I could be at almost 8
[04:27:55] <zzz64> heh. it moves faster at lower voltages but stronger at higher
[04:28:11] <rue_house> yea, I'm trying to keep the origional drive torque
[04:29:02] <rue_house> I'm just overhauling it with a 8 channel new brain
[04:29:15] <rue_house> that can do on/off and torque control
[04:29:26] <zzz64> neat
[04:29:56] <rue_house> individually adjustable pwm freqs for playing music will come later
[04:30:16] <zzz64> lol.
[04:30:28] <rue_house> if it has to make noise, it might as well sound good
[04:32:58] <rue_house> the whole setup is still much larger than I wanted
[04:34:25] <rue_house> I wonder if I can use a ULN2308 and pullups
[04:34:29] <rue_house> and get away with it
[04:36:48] <rue_shop4> maybe, the turnon would always be inherintly delayed
[04:38:35] <rue_shop4> I'd get 2 H bridges in 16 pins
[04:38:42] <rue_shop4> er 18
[04:39:57] <rue_shop4> but I'd need an inverter.... or would I
[04:41:10] <rue_shop4> 12V-1.4V/2.7k
[04:41:39] <rue_shop4> 4mA
[04:43:34] <rue_shop4> I think it'll work
[04:44:31] <rue_shop4> 2am, gnight
[11:19:17] <rue_shop4> ok
[11:19:17] <rue_shop4> so
[11:19:29] <rue_shop4> uln2803 with pullup resistors
[11:19:41] <rue_shop4> I have 2us switching right now
[11:19:47] <rue_shop4> I want turnon slower than turnoff
[11:20:05] <Tom_itx> did rue_shop2 & 3 get fired?
[11:21:02] <Tom_itx> capacitance will slow it down won't it?
[11:21:07] <rue_shop4> other computers, not sure
[11:21:23] <rue_shop4> so will pullup resistors on open collector gate drivers
[11:21:42] <rue_shop4> I need to know if resistors will work or if I need active current sources
[11:34:14] <rue_shop4> huh
[11:34:33] <rue_shop4> well, a uln2803 with a __1k__ pullup is 10x better thant he max232
[11:34:42] <rue_shop4> goes to 200khz easy
[11:34:48] <Tom_itx> what you making?
[11:34:55] <rue_shop4> motor driver :)
[11:35:12] <rue_shop4> I'm making an 8 channel servo controller for gutted servos
[11:35:29] <rue_shop4> I took out the control baords and put a 5 wire connector on with the motor and pot on it
[11:35:44] <rue_shop4> cause I want to be able to turn them off, and do torque control
[11:35:52] <rue_shop4> (force feedback)
[11:36:22] <rue_shop4> and I want to use the 16 channel led pwm controllers
[11:47:04] <rue_shop4> ok that works
[11:47:32] <rue_shop4> use a standard N fet to switch to ground a logic level N fet to switch to +
[11:47:49] <rue_shop4> the supply rail is 5-7V and I'm driving with 12V, which works
[11:48:01] <[VEGETA]> so you will feed it a 50hz pwm signal to control it?
[11:48:08] <rue_shop4> and the uln 2803 driver
[11:48:17] <rue_shop4> no, I got rid of the standard brain
[11:48:35] <rue_shop4> I'm getting an avr to control 8 gutted servos
[11:48:44] <rue_shop4> with extra features
[11:49:11] <rue_shop4> like off, position reporting, current reporting and current control
[11:50:19] <rue_shop4> the H bridge drop is now under 0.1V, which is great
[11:50:39] <rue_shop4> I wish I could have done it in a dip8 package
[11:53:47] <[VEGETA]> you need pwm signals to control the servo.
[11:54:20] <rue_shop4> not after you take the controller out and bring the motor and pot out on a special cable
[11:54:26] <[VEGETA]> arduino has some libs to generate it... but if you want a universal method i will give it to u
[11:54:41] <rue_shop4> nope
[11:55:12] <[VEGETA]> here is my video tutorial, I hope it benefits you:
https://www.youtube.com/watch?v=lL39vzxn3bA
[11:55:18] <rue_shop4> hmm, this will take 32 fets, this wont be smal
[11:55:25] <[VEGETA]> I don't think I understood your method though
[11:55:31] <rue_shop4> I'm taking the controllers out of the servos
[11:55:39] <rue_shop4> and controlling the motor myself
[11:55:50] <[VEGETA]> it will be a normal dc motor then
[11:55:54] <rue_shop4> yea
[11:55:54] <[VEGETA]> not a servo
[11:55:56] <rue_shop4> and a pot
[11:56:08] <rue_shop4> and when I use an avr to control it, its a servo again
[11:56:09] <[VEGETA]> how will you control it via code?
[11:56:17] <rue_shop4> normal servo loop
[11:56:24] <rue_shop4> and I'll use serial control
[11:56:39] <rue_shop4> this baord will be bigger than I want tho
[11:56:40] <[VEGETA]> make a video tutorial about it... looks great really
[11:56:58] <rue_shop4> I need force feedback
[12:00:25] <Tom_itx> current sense
[12:06:34] <rue_shop4> TLC5940, maybe I dont need all the ULN2803 drivers
[12:07:45] <rue_shop4> 12 bit :)
[12:08:09] <rue_shop4> I'd need to make sure the dot correction was cranked
[12:13:42] <rue_shop4> yea, it looks like I just need them as inverters
[12:15:56] <rue_shop4> It might almost be better at that point to use pn2222
[12:16:01] <rue_shop4> or... 2N7000
[12:31:24] <rue_shop4> yea, thats a crowded board
[18:45:12] <rue_shop4> 554.37Hz * 4096
[18:45:34] <rue_shop4> 2.27Mhz
[22:04:16] <rue_shop4> ok the timer is generating the clock for the pwm genorator
[22:09:42] <z64555> woo
[22:15:40] <Jak_o_Shadows> I just swapped my logic analyzer software from USBee to sigrok
[22:15:42] <Jak_o_Shadows> seems okish?
[22:16:10] <z64555> worst commercial one-liner, ever
[22:16:23] <z64555> :P
[22:20:09] <ladonsvenom> m
[22:20:51] <ladonsvenom> Does anyone here participate in FIRST Robotics
[22:20:54] <Jak_o_Shadows> I can write my own protocol decoders, which is nice I suppose
[22:23:02] <Tom_itx> US 1st country in history to win 1000 gold medals
[22:26:51] <Jak_o_Shadows> http://www.stats.govt.nz/browse_for_stats/population/estimates_and_projections/olympics-2016.aspx
[22:27:36] <Tom_itx> phelps has won more medals than 108 countries in the olympics
[22:28:33] <Jak_o_Shadows> Yeah. Phelps is kinda amazing
[22:28:53] <Tom_itx> holds quite a few records
[22:32:59] <Tom_itx> Jak_o_Shadows, will your LA work with saleae software?
[22:33:05] <Tom_itx> i kinda like it
[22:36:13] <Jak_o_Shadows> Dunno. It's just one of these:
http://www.ebay.com/itm/24MHz-8-Channel-USB-Logic-Analyzer-8-CH-Logic-Analyzer-for-Arduino-MCU-/191685084604
[22:36:36] <Jak_o_Shadows> To be honest, probably, from what I understand of these logic analyzers
[22:58:42] <ladonsvenom> test
[23:00:04] <ladonsvenom> test
[23:06:45] <Jak_o_Shadows> hi
[23:58:03] <theBear> rue_house, interested-other-parties,
http://imgur.com/a/VtDMi i show yer any of this the other week ?