Back
[03:07:58] <anniepoo> 8cD well, I set up an arduino to add sensors, so I feel like I actually got around to doing something
[03:12:10] <madeline> neat
[03:12:35] <anniepoo> we have a network of sensors, set up by my biz partner, just outta hackishness
[03:12:47] <anniepoo> I'm just setting some up to learn
[03:13:50] <madeline> sounds like a fun project
[03:20:04] <anniepoo> 8cD
[11:55:36] <MisterX> anyone know something about IMU standard?
[11:56:59] <MisterX> it seems that gyros only give information about angular velocity
[11:57:43] <MisterX> and why are they calling magnetometers for accelerometer and gyroscope...?
[11:57:51] <MisterX> it's not even close
[11:59:40] <MisterX> anyone know a precise gyro? i need a device that can detect it's current rotation relative to a gyroscope that has to be calibrated
[11:59:52] <MisterX> it's current angle*
[12:00:04] <MisterX> one that doens't flicker
[12:00:40] <Snert_> what is flickering?
[12:00:42] <SpeedEvil> ...
[12:01:10] <SpeedEvil> If you want a nice all-in-one IMU unit, analog devices does some quite nice ones for only around a thousand dollars.
[12:01:34] <MisterX> then it's angle is 0 and it is laying on still, then it prints anything from 1 to -1
[12:01:40] <SpeedEvil> http://www.analog.com/en/products/mems/inertial-measurement-units.html
[12:01:56] <SpeedEvil> MisterX: all IMUs or gyros have noise.
[12:02:04] <MisterX> when the gyro is laying till on the table*
[12:02:07] <SpeedEvil> yes.
[12:02:22] <Snert_> look for hysteresis or possible dead zone adjustments.
[12:02:45] <SpeedEvil> You need to sample consistently at a precise interval, and then input all the gyro/mag/accel data into a filter to get usable output.
[12:02:54] <SpeedEvil> And even then, it may be less good than you hope.
[12:03:04] <SpeedEvil> What are you hoping to do with this IMU?
[12:03:43] <MisterX> i want to use gyroscope only to autobalance an object
[12:03:52] <MisterX> instead of a mix of gyro and accelerometer
[12:05:43] <MisterX> it seems that gyros and accelerometers usualy is a 3axis and 3dimentional magnetometer
[12:05:54] <SpeedEvil> The gyro random walks a few degrees an hour typically
[12:06:01] <SpeedEvil> you need to cope with this
[12:06:06] <MisterX> one that can meassure movement in 3 directions in addition to 3 angular velocities
[12:06:17] <MisterX> hmm
[12:06:34] <SpeedEvil> Most IMUs will be kilometers out after a few minutes.
[12:06:45] <MisterX> i need a very precise one where maximum and minimum is just 5 angles
[12:07:17] <MisterX> and then have 180* or 90* divided on these 5 angles
[12:07:30] <MisterX> so it should be like 5.0001
[12:07:47] <MisterX> 5 degrees*
[12:08:10] <MisterX> need 180 or 90 steps from 0 to 5*
[12:08:53] <SpeedEvil> What are you actually trying to do?
[12:09:01] <SpeedEvil> Oh
[12:09:08] <SpeedEvil> balancing - why do you think you need that?
[12:09:28] <SpeedEvil> you have a slowly drifting angle, with a small amount of noise - that's quite enough to balance stuff
[12:09:35] <MisterX> uhm, i'm making something very very secret atm x-D let's just say it's a segway
[12:09:51] <MisterX> i'll post an image later on
[12:09:59] <MisterX> when it's complete
[12:10:13] <SpeedEvil> you don't need that for that class of device.
[12:10:28] <SpeedEvil> you simply feed the gyro output into a simple filter, and out pops the relative angle
[12:14:10] <MisterX> hmm i guess i'll try that.. until we have more precise gyroscopes that is x-D
[12:14:29] <MisterX> i think the flickering somes from the gravitational force from the ground
[12:15:42] <MisterX> and magnetometers gives different readings depending on the ground it stands on
[12:16:12] <MisterX> if the ground is leveld or not
[12:17:42] <MisterX> think most "gyros" can be calibrated to the current location it is on on the planet
[12:17:55] <MisterX> gyros that are in fact magnetometers
[12:19:00] <MisterX> isn't that false marketing by the way?
[12:19:12] <MisterX> selling a magnetometer as a gyroscope is kind of missleading.
[12:20:08] <MisterX> it's like buying a blue house but you get a red
[12:44:54] <z64555> uh
[12:44:56] <z64555> no
[12:45:29] <z64555> electronic compasses are often a pairing of magnetometers and gyros
[12:45:49] <z64555> the magentometer works at a slow frequency, and the gryos work at a faster
[12:46:31] <z64555> can also toss in an accelerometer to get a better approximation
[12:46:39] <z64555> to reference the gravity field
[12:47:17] <z64555> Don't do dead reckoning on an IMU
[12:47:30] <z64555> if you require any sort of significant precision for navigation
[12:48:00] <z64555> you will need a sensor system that has some sort of immovable reference, such as a beacon
[12:48:37] <z64555> If the beacons position is known, and you can dtermine the angles to the beacon, and if you can determine the range to the beacon
[12:48:50] <z64555> you can triangulate your position on the map with good certainty
[12:49:48] <z64555> you can also use a GPS, but it has its own downfalls
[13:03:10] <delinquentme> im looking at a datasheet for a LM35 and I have a hypothesis that physical impact might have messed up its measurement capacoty
[13:03:34] <delinquentme> so im wondering what keyword would I search the datasheet for to find out what the physical acceleration / deceleration limits are for this sensor
[13:05:16] <robopal> G force?
[13:06:42] <robopal> vibration
[13:15:15] <MisterX> but there's not a phisucal gyroscope in normal imus, right?
[13:15:24] <deshipu> MisterX: there is
[13:15:26] <MisterX> physical*
[13:15:27] <deshipu> MisterX: very small
[13:15:55] <deshipu> MisterX:
https://www.youtube.com/watch?v=PK05u9c3yWI
[13:16:19] <deshipu> or maybe
https://www.youtube.com/watch?v=eqZgxR6eRjo
[13:19:58] <MisterX> hmm interesting
[13:20:18] <MisterX> but it says mems gyrops are vibrating?
[13:20:27] <MisterX> https://en.wikipedia.org/wiki/Vibrating_structure_gyroscope#MEMS_gyroscopes
[13:20:39] <MisterX> it's not a wheel inside of it?
[13:22:03] <MisterX> think i need something like this
https://www.youtube.com/watch?v=mrGfc-3uv7o
[13:22:13] <MisterX> or i want atleast x-D
[13:22:49] <deshipu> no wheels, no
[13:26:07] <MisterX> hmm think i'll try to make one, and have it transfer power to two potentiometers or something
[13:42:34] <MisterX> this is silly... how do you find a gyro with a wheel in it?
[13:42:51] <MisterX> the producers here are producing thigs they don't understand it seems
[13:42:56] <MisterX> things*
[13:54:18] <MisterX> hang on here a second... isn't ther eonly 6 degrees of freedom in the universe?
[13:54:34] <MisterX> https://www.sparkfun.com/products/13284 what are the 3 last?
[13:55:11] <MisterX> rotate x, rotate y, rotate z and move x, move y, move z
[13:57:02] <MisterX> uhm that would be 12 degrees og freedom divided on 3 dimentions in 2 planes of existence
[14:18:09] <z64555> MisterX: MEMS gyros are a set of forks that vibrate at a specific frequency
[14:18:57] <z64555> torque sensed by the forks thereby affects the phase of the signal
[14:19:04] <z64555> or something like that
[14:22:16] <SpeedEvil> 9 degrees of freedom is simply the fact that the accel, mag and gyro are seperate and have three channels
[14:22:57] <z64555> ^
[14:23:25] <veverak> mag -> position, gyro -> rotation accel, accel -> position accel
[14:23:27] <veverak> ?
[14:23:29] <veverak> :)
[14:23:34] <z64555> no
[14:24:06] <z64555> mag is angle, gyro is rotational velocity, accelerometer is translational acceleration
[14:24:19] <z64555> but
[14:24:24] <veverak> rotation velocity, hmm
[14:24:51] <MisterX> only when it's not calibrated
[14:25:07] <z64555> the accelo can sense rotataional accerlation, too
[14:25:08] <MisterX> think i need one hand calibrate it every time it's shut on
[14:25:17] <MisterX> the gyro that is
[14:25:26] <z64555> nah
[14:26:10] <MisterX> the gyro sneds nagular velocity only when it's not calibrated
[14:26:19] <MisterX> one should be able to get it's current angle aswell
[14:26:25] <z64555> just let it stay still for a few moents for the kalman sensor to zero
[14:26:53] <z64555> i'm talkin about the sensor, not your silly little module in an enclosure
[14:27:05] <z64555> the actual chip
[14:27:26] <MisterX> making something similar to this, but this version is using 4 servos to lift weights
[14:27:38] <MisterX> or shift the weights position*
[14:27:39] <MisterX> https://www.youtube.com/watch?v=woCdjbsjbPg
[14:27:52] <MisterX> think i'll go for MPU 6050
[14:27:56] <MisterX> think he's using that
[14:28:45] <MisterX> the degrees/second are so low when it stands still
[14:28:52] <z64555> that works on torque
[14:28:52] <MisterX> need a very precise imu
[14:29:04] <MisterX> yeah, but it's the same principle
[14:29:05] <z64555> yours works on center of mass
[14:29:17] <z64555> not quite the same preciple
[14:30:09] <MisterX> i mean the way you find out how fast the motors should spin
[14:30:27] <MisterX> that value will control the position of the weight
[14:30:30] <z64555> speed isn't what is causing the stick to center
[14:31:05] <MisterX> but it's almost standing still so there's like 5 degrees of movement before it falls over
[14:31:13] <MisterX> acceleration*
[14:31:21] <MisterX> of the wheels
[14:31:52] <MisterX> 5 degrees as in 360degrees in a circle
[14:32:04] <z64555> that's only 2 degrees of freedom
[14:32:15] <z64555> controlled by 2 wheels
[14:32:39] <MisterX> no.. there's like 5 of 360 usable degrees in the circle
[14:32:51] <z64555> name them
[14:32:56] <MisterX> if it's angle is more then 5 degrees, it'll fall
[14:33:02] <MisterX> the current version that it
[14:33:14] <MisterX> because of how the weights are placed
[14:33:21] <z64555> oh, something different then
[14:33:44] <MisterX> yeah x-D not degrees as in 9 degrees of freedom(9DOF)
[14:33:46] <z64555> I was thinking you were still talking about degrees of freedom, but you where talking degrees of angle
[14:34:14] <z64555> anyway
[14:34:25] <MisterX> hehe they must have very confusing conversations in that production facility x-D
[14:34:39] <z64555> That stick in the video uses the torque of spinning up/down the reaction wheels to keep it centered
[14:35:05] <z64555> The wheels weren't spinning fast enough to get a significant gyroscope effect until near the end
[14:35:29] <MisterX> i'm using something very very similar
[14:35:38] <z64555> it didn't torque about vertical axis because of the power cable
[14:35:41] <MisterX> it's 4 of those wheels, but they are chopped in halv'
[14:35:45] <MisterX> halv*
[14:35:53] <z64555> yeah, you're doing center of mass
[14:36:15] <MisterX> they create torque when they start spinning
[14:36:28] <z64555> and a hellova lot of vibration
[14:36:35] <MisterX> so they must retract at different speed depending on how fast it is falling
[14:37:35] <z64555> hm, you'll figure it out eventually
[14:38:03] <MisterX> the finished product will have a mix of this balancing system i'm testing now, that wheel thingy and a 3rd joint in the middle in addition to a wheel on the bottom
[14:38:28] <z64555> lol.
[14:38:41] <MisterX> it will be impossible to make it fall x-D
[14:38:56] <MisterX> hopefully
[14:39:15] <z64555> more like impossible to make it stand still. lol
[14:39:31] <z64555> use a big ball bearing instead of the wheel
[14:39:55] <z64555> because reasons. :D
[14:40:12] <MisterX> good idea x-D
[14:40:17] <MisterX> need as small wheel as possible
[14:40:26] <MisterX> was thinking of using a skateboard wheel or something
[14:41:49] <SpeedEvil> is it for use on very smooth surfaces only?
[14:42:27] <MisterX> yeah
[14:55:40] <MisterX> is it possible that SPI will give a moroe accurate reading the i2C?
[14:55:47] <MisterX> interface
[14:56:38] <MisterX> does one of them have higherbandwith speed then the other?
[16:07:58] <deshipu> spi can be much faster than i2c, but neither will go faster than the sensor can read
[17:30:56] <MisterX> raspberry pi zero can do pretty much the same stuff as arduino uno, right?
[17:31:07] <veverak> not even close
[17:31:07] <MisterX> and more
[17:31:14] <veverak> rasberry pi zero is something that runs OS
[17:31:24] <veverak> arduino uno runs simple program
[17:31:54] <MisterX> raspberry should be able to do more stuff then, right?
[17:32:15] <veverak> yep
[17:32:46] <MisterX> cool
[17:33:43] <anonnumberanon> Wait, you can actually buy one of those?
[17:34:02] <anonnumberanon> what's on it anyway, raspbian zero?
[17:34:08] <anonnumberanon> deshipu, do you have one yet?
[17:34:13] <anonnumberanon> is it qt?
[17:35:09] <anonnumberanon> rue_bed2, how do they do accurate PWMs with those behemoths that have an OS?
[17:36:25] <veverak> sucks
[17:36:32] <veverak> but there is something based on DMA for raspberry
[17:37:31] <veverak> https://github.com/metachris/raspberrypi-pwm
[17:38:05] <veverak> hmm
[17:38:22] <veverak> need to check if by any chance, we don't have some subject related to "covariance matrix" at school
[17:40:27] <anonnumberanon> DMA is for accessing GPIO?
[17:42:52] <veverak> dunno about details
[17:42:55] <veverak> you know what DMA is?
[17:43:00] <anonnumberanon> I still don't see how it's used for GPIO.
[17:43:01] <anonnumberanon> http://www.embedded.com/electronics-blogs/beginner-s-corner/4024879/Introduction-to-direct-memory-access
[17:43:16] <anonnumberanon> No I don't know I'm trying to find out.
[17:44:44] <anonnumberanon> The introduction of this article doesn't compare it to using GPIO on a microcontrollre therefore I chose not to continue reading it :)
[17:45:06] <veverak> DNA is direct memory access -> something else is accessing the memory of device
[17:45:15] <veverak> completely bypassing the CPU of that thing
[17:45:28] * anonnumberanon goes back to his Eagle tutorial
[17:45:34] <veverak> -> you can avoid interrupts that happens during normal CPU launch
[17:45:43] <veverak> I dunno "what something else" actually is in this case
[17:45:44] <anonnumberanon> ah I see veverak that sounds interesting
[17:46:57] <veverak> yeah
[17:47:34] <veverak> anonnumberanon: from description of the code
[17:47:44] <veverak> it seems that DMA controller got direct access to registers
[17:47:46] <veverak> hmm
[17:50:13] <anonnumberanon> seems like raspberry pi in general is a huge bloat accumulated over an otherwise quite powerful and useful ARM microcontroller
[17:50:36] <veverak> yep
[17:50:49] <orlock> anonnumberanon: Check out the gl.inet gear like the domino
[17:51:07] <orlock> anonnumberanon: or some of the other openwrt based devices on the same chipsets
[17:51:40] <anonnumberanon> gl-inet
[17:51:45] <anonnumberanon> .com
[17:52:09] <orlock> i've got several of them, they are nice
[17:52:46] <veverak> anonnumberanon: I am using i2c-pwm controller for now
[17:53:12] <orlock> veverak: Nice.. cost and how many pwm channels?
[17:53:19] <Tom_itx> got an OT question...
[17:53:33] <anonnumberanon> veverak, ah so it can be used for that..
[17:53:42] <Tom_itx> newer laptops come with windows 10 installed. is there a way to install 7 and get drivers for it for those?
[17:53:44] <veverak> (rpi is just talking to it via i2c)
[17:53:49] <anonnumberanon> Would like to look at the waves through an oscilloscope to see jitter if any..
[17:54:03] <veverak> orlock:
https://es.aliexpress.com/item/Smart-Electronics-PCA9685-16-Channel-12-bit-PWM-Servo-Driver-I2C-Interface-for-Arduino-Raspberry-Pi/32464046768.html?spm=2114.13010608.0.54.4un2B7
[17:54:06] <Tom_itx> the only drivers i see available would be for 8.1
[17:54:15] <orlock> Tom_itx: All depends on the laptop
[17:54:22] <Tom_itx> yeah i know.
[17:54:24] <Tom_itx> it's a dell
[17:54:36] <Tom_itx> and the only drivers they show are for 8.1 or 10
[17:54:54] <anonnumberanon> Tom_itx, try ##windows they're pretty nice there about these sorts of questions.
[17:54:59] <orlock> 8.1 should work for 7.. i _think_
[17:55:07] <veverak> Tom_itx: would avoid doing that at all
[17:55:09] <Tom_itx> that's kinda what i was wondering
[17:55:20] <veverak> from "drivers" perspective, using windows 7 will be harder and harder to use
[17:55:20] <orlock> but yeah, do what he said
[17:55:26] <Tom_itx> software this person needs doesn't run under 10
[17:55:29] <orlock> We are having the same issue
[17:55:44] <Tom_itx> veverak i'm sure of it
[17:55:45] <veverak> yay
[17:55:53] <orlock> Tom_itx: New laptop with USB-c?
[17:55:58] <Tom_itx> i'm not sure
[17:56:12] <veverak> Tom_itx: I would use virtual machine than, but that may not be preferrable
[17:56:17] <Tom_itx> she's a photographer and needs photoshop but it doesn't run under 10
[17:56:28] <veverak> wait
[17:56:31] <Tom_itx> veverak that won't work. she's no guru
[17:56:31] <veverak> wat?
[17:56:52] * veverak does not believe that photoshop won run on windows 10
[17:57:02] <Tom_itx> it doesn't
[17:57:05] <orlock> I would not be suprised
[17:57:10] <orlock> considering the shit we are going through
[17:57:11] <Tom_itx> it does all sorts of weird things
[17:57:13] <anonnumberanon> bullshit
[17:57:19] <anonnumberanon> it's really buggy?
[17:57:21] <orlock> my vmware console is fucking up under win10
[17:57:30] <Tom_itx> i have it on 7 and it's fine
[17:57:43] <Tom_itx> photoshop lightroom 6 to be exact
[17:58:23] <Tom_itx> but the laptop is new enough i don't think i can get 7 drivers
[17:58:34] <Tom_itx> unless i look for chip drivers
[17:58:43] <Tom_itx> and i don't really wanna go down that road
[17:59:16] <veverak> http://blogs.adobe.com/crawlspace/2015/07/photoshop-and-lightroom-compatibility-with-windows-10.html
[17:59:26] <anonnumberanon> bbl
[17:59:34] <anonnumberanon> Eagle calls
[17:59:36] <veverak> Tom_itx: I don't think that something as adobe photoshop won't run on windows 10
[18:00:14] <veverak> yeah, they have issues, but it's "Release blogpost"
[18:02:03] <veverak> Tom_itx: I mean, have you tried to properly debug why there is problem with photoshop?
[18:02:35] <Tom_itx> no i just was made aware of it this afternoon :)
[18:03:02] <Tom_itx> i suspect it's all the baggage that comes with a windows install
[18:03:19] <veverak> if it were some "crappy" software without proper support. OK
[18:03:30] <veverak> but something as Photoshop should be working on windows 10
[18:03:32] <Tom_itx> i hear ya
[18:03:41] <veverak> anyway, if she bought it
[18:03:44] <veverak> they have support
[18:03:46] <veverak> :)
[18:03:53] <Tom_itx> i hear that too :)
[18:03:54] <veverak> so they should help her make it work on windows 10
[18:03:58] <veverak> :)
[18:05:37] <Tom_itx> this is ver 6.7
[18:05:43] <Tom_itx> i would guess it should work
[18:06:49] <Tom_itx> i wonder if they have an update to it
[18:07:11] <Tom_itx> as my product box shows 7 8 and 8.1 as supported products
[18:10:18] <Tom_itx> carry on...
[19:14:30] <anonnumberanon> huh.. time to maybe use function pointers
[19:52:07] <Loshki> Hard to believe photoshop won't run at all on 10. Easy to believe it's riddled with buggy behavior, though.
[22:14:00] <rue_house> hmm
[22:14:23] <rue_house> Loshki, yea, its hard to beleive that windows 10 can even successfully boot
[22:16:02] <rue_house> so I guess I have to apply the *other* part of that softare and work out the position again
[22:16:08] <rue_house> see how close it hits
[22:16:22] <rue_house> maybe knock down the angular resolution and see how it performs
[22:16:54] * rue_house cackles 'last spoken, channels mine now suckers!'
[22:17:03] * rue_house listens to the crickets
[22:18:05] <rue_house> arg, what ahppened this summer
[22:18:17] <rue_house> I remember planning things I wanted to do in spring