**** BEGIN LOGGING AT Wed Jul 29 02:59:58 2015 Jul 29 10:09:56 nerdboy: I'm givin' her all she's got Captain! Jul 29 10:10:52 nerdboy alexanderhiam, Scotty was AFK these past two days more than he'd like. Time to polish the engine now that he's back Jul 29 10:11:08 https://github.com/nvisnjic/PyBBIO/blob/master/bbio/libraries/MPU9250/MPU9250.py Jul 29 10:11:31 made an MPU library for PyBBIO Jul 29 10:12:00 and manage to get it working via SPI, but had to use the internal I2C to communicate to the magnetometer. It works though Jul 29 10:12:51 * neemo was not happy with the lack of detail in the datasheets, it's like they've actively tried not to mention how to talk to the magnetometer and mentioned eeeverything else Jul 29 10:14:06 but anyway, it works, which is the point. Have to still add the selftest functions for the library, though the compensation algorithms fix those as well Jul 29 10:15:07 alexanderhiam I've modeled the library by using the BMP183 one and still have some cleaning to do Jul 29 10:15:24 but point me what to fix so it fits in PyBBIO nicely Jul 29 10:16:35 s/point out/ Jul 29 10:19:32 for the compensation part, I've read a bit more into this library -> https://github.com/richards-tech/RTIMULib/tree/master/RTEllipsoidFit Jul 29 10:20:44 and found a nice confirmation that the algorithms I've used so far work in the real world, since this library uses the same algorithm (Yury's) to fix the data points Jul 29 10:21:09 Now I should look into how to expand Yury's work to include the time-invariant compensation i Jul 29 10:21:13 I'll need to fit in Jul 29 10:22:40 so have to package the fitting/correction into Python or link it to some other solver Jul 29 10:22:52 and finish work on the API, now in Python Jul 29 10:23:46 nerdboy: alexanderhiam: Ready for the ideas / comments :) Jul 29 10:24:36 I'll write up a report for the week (pretty much what's written above) and ping back here once it's morning in the US Jul 29 10:26:07 oh yeah, added a short example for the library at https://github.com/nvisnjic/PyBBIO/tree/master/examples Jul 29 10:26:43 That's it for the update Jul 29 12:06:40 neemo: MPU9250 driver looks good Jul 29 12:10:16 so the mag is basically a totally separate sensor on the auxiliary I2C bus? Jul 29 12:16:36 InvenSense doesn't seem to sell a standalone magnetometer... they must have just bought an I2C magnetometer IP core and slapped it on the die Jul 29 13:54:28 <_av500_> test Jul 29 15:30:14 alexanderhiam: that's precisely what they did Jul 29 15:32:37 and while I couldn't care less on how they physically arrange their sensors, after rereading the datasheet about 30 times, I can certainly say they could've done a better job at describing how someone is supposed to access that slapped about hardware Jul 29 15:33:47 the fact that the registry map is half written by InvenSense and that the AK part was copy-pasted from some different AK datasheet didn't help one bit Jul 29 15:34:18 but alas it works, that the important part :) Jul 29 15:35:49 s/that/that's even Jul 29 15:44:57 gm! Jul 29 15:47:26 moin Jul 29 15:48:44 gm Jul 29 15:50:24 as long as the approach is valid, sensors can can get swapped out later Jul 29 15:55:14 nerdboy: yup Jul 29 15:55:41 <_av500_> 5min to 5min Jul 29 15:55:46 have to write a second library for the PNI mag as well Jul 29 15:56:04 but I don't think I'll have half the issues I had with MPU9250 Jul 29 15:57:57 mpu9250 issues? Jul 29 15:58:05 hello all Jul 29 15:58:08 * nerdboy re-architecting embedded build infrastructure Jul 29 15:58:35 will look at the blog posts when train stops Jul 29 15:58:44 hello vvu Jul 29 15:58:59 Hello all Jul 29 15:58:59 * nerdboy will look at the log when cmake build barfs Jul 29 15:59:05 hello Jul 29 16:00:19 ds2: had a bunch of them with interfacing to the MPU via SPI Jul 29 16:00:41 primarily on how to get a hold of the magnetometer which is connected internally via I2C Jul 29 16:01:04 since when using SPI, you can't use I2C bypass to access the AK mag directly Jul 29 16:01:17 * nerdboy flings another donut at moto-timo Jul 29 16:01:53 got it solved, but wasted a bunch of time digging through the not so well documented datasheets Jul 29 16:02:07 also, 'ello all Jul 29 16:02:20 best to do plenty of pre-buy evaluation Jul 29 16:02:53 yeah Jul 29 16:03:12 some vendors are too cheap to "license" I2C name and hack their own protocols... Jul 29 16:03:22 don't buy from them Jul 29 16:03:24 gm all Jul 29 16:03:42 <_av500_> jkridner: aloha Jul 29 16:03:49 well the point was I knew it *could* work, since I've been playing with the RTIMUlib and it did SPI just fine Jul 29 16:03:57 gm jkridner Jul 29 16:03:59 wasn't happy when it refused to work for me :P Jul 29 16:04:01 gm Jul 29 16:04:22 gm Jul 29 16:04:24 <_av500_> jkridner: 1min Jul 29 16:04:24 Hi jkridner Jul 29 16:04:35 neem0: how did you solve it? Jul 29 16:04:42 * nerdboy waves his hand as the actual morning timezone guy... Jul 29 16:05:00 ds2: with a lot of trial and error Jul 29 16:05:01 0904 Jul 29 16:05:29 neem0: knowif hte same technique works on the mpu6xxx's? Jul 29 16:05:31 ds2: and read throug more than a couple of different solutions, from which some worked half way through Jul 29 16:05:32 is she testing a mac? Jul 29 16:05:46 ds2: should work, I think they're completely the same Jul 29 16:05:48 even simpler Jul 29 16:05:54 nerdboy: PST? Jul 29 16:06:08 since I think the 6xxx are the ones without the AK magnetometer on them, correct me if I'm wron ds2 Jul 29 16:06:13 daylight stupid time, but yeah... Jul 29 16:06:34 ds2: ei mpu9x = mpu6x + AK magnetometer sandwiched on-die Jul 29 16:06:40 <_av500_> aloha Jul 29 16:06:41 <_av500_> its time Jul 29 16:06:48 <_av500_> 1) Welcome everybody Jul 29 16:06:56 neem0: IIRC, that stuff falls into the grey area for DMP support Jul 29 16:07:12 if you have one you could probably git pull and test directly, I don't think they've bothered to change any of the regs Jul 29 16:07:15 neem0: yes but there might be 'revision' issues Jul 29 16:07:23 <_av500_> ds2: neem0: later Jul 29 16:07:23 true Jul 29 16:07:29 <_av500_> 2) Who's here? Jul 29 16:07:33 I'll order one cheap here from taobao and can test later Jul 29 16:07:38 _av500_: ay ay Capt' Jul 29 16:07:41 * rohith is here Jul 29 16:07:42 here Jul 29 16:07:43 here Jul 29 16:07:45 here Jul 29 16:07:45 here Jul 29 16:07:52 here Jul 29 16:08:07 here Jul 29 16:08:18 <_av500_> and I saw arianepaola_ ? Jul 29 16:08:43 yup... hopefully we don't need a search party... Jul 29 16:09:20 * nerdboy can remember searching for lost TA in northern Baja Jul 29 16:09:20 <_av500_> arianepaola_: ping Jul 29 16:09:36 * _av500_ can remember you telling that last week :) Jul 29 16:09:56 we still have not issue Brunton's yet Jul 29 16:10:10 and we're missing 2 Jul 29 16:10:23 2? Jul 29 16:10:31 <_av500_> we miss arianepaola_ so far Jul 29 16:10:40 <_av500_> and there she is Jul 29 16:10:40 sorry, missed one... Jul 29 16:10:44 <_av500_> arianepaola: hi Jul 29 16:11:05 <_av500_> 3) Reports Jul 29 16:11:09 <_av500_> I count 8 Jul 29 16:11:12 <_av500_> andthere she was Jul 29 16:11:22 <_av500_> reports are all in Jul 29 16:11:31 * moto-timo sneaks in the back quietly, then drops his coffee on the floor Jul 29 16:11:31 7 plus one dual-particle Jul 29 16:12:00 <_av500_> about 2.5 weeks coding time left Jul 29 16:12:13 <_av500_> I hope everybody and their mentor is aware Jul 29 16:13:06 <_av500_> so any urgent or blocking issues better be brought up and resolved quickly Jul 29 16:13:36 <_av500_> so: Jul 29 16:13:42 <_av500_> 4) URGENT or BLOCKING issues? Jul 29 16:13:56 nope Jul 29 16:13:58 me Jul 29 16:13:58 no Jul 29 16:14:09 no Jul 29 16:14:12 no Jul 29 16:14:14 all green, solver will be immenent for next week though Jul 29 16:14:34 no Jul 29 16:14:48 s/immenent/imminent/ (facepalm) Jul 29 16:15:27 <_av500_> ebadawy/jkridner you aware? Jul 29 16:15:45 arianepaola_: any issues ? Jul 29 16:16:07 jkridner: can we talk after the meeting to arrange what should be done for the next two weeks, and also about what I have done so far? Jul 29 16:17:02 * nerdboy waiting to see if -flto can handle cmake linking on 1G machine... Jul 29 16:17:12 jkridner: I beleive that there were some issues about how the index page should look like. Jul 29 16:17:45 ebadawy: I can hang out a bit after the official meeting, but I have limited bandwidth as I'm currently on a mobile phone. Jul 29 16:17:53 and traveling. Jul 29 16:18:19 shubhangi: are you up to speed now? Jul 29 16:19:08 jkridner: too bad, my be we could have that meeting after you get a good connection. Jul 29 16:19:09 Abhishek_: still a bit behind schedule Jul 29 16:19:22 may* Jul 29 16:19:35 Abhishek_: extra hourse this week Jul 29 16:19:40 hours* Jul 29 16:20:08 vmayoral|pc ahcorde imuguruza: The build process on the Erle Brain should be mostly same as building on linux, right? Jul 29 16:20:44 Abhishek_: moving forward though... not stuck Jul 29 16:21:19 <_av500_> as useual, I urge everybody to use this channel all the time to interact Jul 29 16:21:27 <_av500_> its here to help you Jul 29 16:21:40 <_av500_> not just during the weekly meeting Jul 29 16:21:41 rohith: apm? Jul 29 16:21:43 maybe if we served cookies... Jul 29 16:22:12 rohith: if you are using Erle-Brain you have to use "make pxf" Jul 29 16:22:23 ahcorde: I meant replacing make linux with make pxf should do it. Jul 29 16:22:25 (mmmmm pepper cookies) Jul 29 16:22:30 ahcorde: Yes Jul 29 16:22:31 * neem0 got one left Jul 29 16:22:48 ( vmayoral|pc: finally flew my erle copter... still have a long ways to go to learn how to really use it. ) Jul 29 16:23:05 jkridner: good hearing that! Jul 29 16:23:25 jkridner vmayoral|pc: I plan to make one after GSoC ends :) Jul 29 16:23:41 * nerdboy adds firefox to no-flto env file Jul 29 16:24:45 vmayoral|pc: I saw a flight video on youtube sometime back. Seemed very "snappy" :) Jul 29 16:24:53 _av500_: are we in that time of the meeting for open discussion, or is there a specific topic? Jul 29 16:25:06 <_av500_> jkridner: URGENT and blocking issues Jul 29 16:25:20 <_av500_> which should not be in this meeting at all Jul 29 16:25:29 <_av500_> but resolved continously Jul 29 16:25:34 the 2 presentations? Jul 29 16:25:38 <_av500_> yes, next Jul 29 16:25:47 <_av500_> appar and neemo today Jul 29 16:25:52 <_av500_> 15 min each Jul 29 16:25:53 rohith: please use cross-compiling to compile faster -> https://docs.google.com/spreadsheets/d/14a5uqoWpgd5zgemcP9-el_m-VuWL-5PKaE94KZM4tUc/edit#gid=0 Jul 29 16:26:12 <_av500_> ahcorde: rohith vmayoral|pc : please resolve that after the meeting Jul 29 16:26:13 yup Jul 29 16:26:18 ups Jul 29 16:26:25 <_av500_> ahcorde: no problem Jul 29 16:26:29 apaar: wanna take the lead? Jul 29 16:26:29 ebadawy: let's talk after the meeting Jul 29 16:26:33 sure :) Jul 29 16:26:34 ahcorde: will look at it Jul 29 16:26:38 <_av500_> 5) status reports Jul 29 16:26:41 __av500__: okay Jul 29 16:26:45 * nerdboy is available for cross-toolchain grumpiness Jul 29 16:26:47 <_av500_> apaar: your 15 minutes of fame Jul 29 16:26:55 * apaar begins Jul 29 16:27:04 <_av500_> focus on progress, assume we know the basics Jul 29 16:27:05 pru bridge : we have managed to get the driver + userspace C API + python API's up and running :) Jul 29 16:27:15 <_av500_> nice Jul 29 16:27:25 * karki_ listens Jul 29 16:27:25 rohith: this link https://github.com/erlerobot/erle_gitbook_basics_guide/blob/master/en/Cross_compiler/README.md Jul 29 16:27:26 there are now 10 channels but each channel is unidirectional.Channel size has to be specified via the init sysfs file and each channel exposes a ch_read and ch_write sysfs file in userspace. Jul 29 16:27:47 a dev can transfer data by simply including pru_bridge.h opening a channel and doing a pru_write or pru_write function call. Jul 29 16:28:01 Also exposed downcalls vis a sysfs just have to input the 6 paramemters + pru no but this needs testing/working on Jul 29 16:28:12 via* Jul 29 16:28:41 I sepent the last week or so looking at the performance side Jul 29 16:29:00 There seems to me no jitter on pru side ie it is realtime Jul 29 16:29:30 got a result of 352 cycles for 4 bytes including cycles for calculating time delay Jul 29 16:29:46 So pru takes around 60-80 cycles per byte which is .5 microseconds Jul 29 16:30:06 this can be optimised/reduced a bit Jul 29 16:30:22 Driver seems to be taking 5-15 micro seconds per byte depending on size passed at once Jul 29 16:30:38 this is only for the C library Jul 29 16:30:52 how big is the PRU code to do this? Jul 29 16:31:51 the firmware file along with the test code comes to around 2-3 Kb Jul 29 16:31:55 i think Jul 29 16:31:57 * apaar checks Jul 29 16:32:14 wait.... isn't the latency a bit too much Jul 29 16:32:31 karki :on which side? Jul 29 16:32:44 pru Jul 29 16:32:57 60 cycles for a read?? Jul 29 16:33:12 we'll definitely need to improve on that Jul 29 16:33:17 karki : no write but read is also very similiar Jul 29 16:33:20 so as much as 25% of code space!?! Jul 29 16:33:35 it got messed up as i am parsing byte by byte Jul 29 16:34:21 ds2: how can i reduce it?? Jul 29 16:34:26 ds2 : I believe there is a lot of space for optimization, plus I think appear has the binary size wrong - he's probably calculating the ELF Jul 29 16:34:58 karki_: ah Jul 29 16:35:20 yes i was talking of the executable the code is 587 bytes Jul 29 16:35:56 Ohh so about 5% of code space Jul 29 16:36:01 worth trying a few things probably... link-opts on top of -Os maybe? Jul 29 16:36:23 ds2 :code can definatly be reduced... Jul 29 16:36:25 25% is a big dent. 5% is less so Jul 29 16:37:19 Also about pru side performance what karki what did you have in mind as a decent latency? Jul 29 16:38:08 ds2 would be more familiar with what a decent latency would be. But all I'm saying is that a 60-80 cycle overhead is a bit too much Jul 29 16:38:44 apaar: that seems really slow. that gets more efficient with larger payloads, no? Jul 29 16:38:46 * jkridner might have missed quite a bit of presentation the last 10 minutes Jul 29 16:39:11 yes larger the size better it gets but i have not tested beyond 4 bytes Jul 29 16:39:18 :o Jul 29 16:39:26 <_av500_> you should do so Jul 29 16:39:28 <_av500_> I guess Jul 29 16:39:42 * apaar means 4 bytes at a time i ran it for 2MB of data Jul 29 16:39:59 apaar - test with few 100s of bytes in a go Jul 29 16:40:06 atleast Jul 29 16:40:15 <_av500_> 128 :) Jul 29 16:40:47 2MB in 4 byte chunks is a good test Jul 29 16:40:48 yes i will but i am sending the time lag withing the data so i ahve to figure out a way.Testing has been done for large sizes i just dont kmow the time delays/latencies Jul 29 16:41:07 but it can't be your only test Jul 29 16:41:32 I should look into the code for optimisations Jul 29 16:41:42 ok cool that will not be very hard :) also any suggestions to reduce latency? Jul 29 16:42:15 could try some auto-vectorization stuff Jul 29 16:42:17 Preempt kernel? Jul 29 16:42:21 ds2 : do you think it makes sense to do read/write ops in hand written assembly Jul 29 16:42:33 ideally 0 cycle latency Jul 29 16:42:34 <_av500_> nerdboy: good one Jul 29 16:42:58 not full -O3, just the the "right" ones... Jul 29 16:43:23 * apaar will look into it Jul 29 16:43:29 karki_: it all comes down to the app. for the stuff I am looking at, yes. cuz I need to interleave it with the main task to maintain timing. If you are doing stuff that happens say ever 10mS, C is fine Jul 29 16:43:47 graphite might help with that Jul 29 16:43:57 karki_: put it another way, is the dead wait time between events greater then time to do the xfer. if not, asm. Jul 29 16:44:00 <_av500_> time is up Jul 29 16:44:09 <_av500_> please continue after the meeting Jul 29 16:44:16 <_av500_> apaar: thanks! Jul 29 16:44:19 ds2 : will look into it :) Jul 29 16:44:20 Thanks Jul 29 16:44:23 thanks :) Jul 29 16:44:29 apaar: I will have a look as well... Jul 29 16:44:35 <_av500_> neem0: your launch slot Jul 29 16:44:40 apaar : you are now officially famous Jul 29 16:44:54 * apaar celebrates Jul 29 16:45:07 <_av500_> +1 Jul 29 16:45:09 apaar: done? Jul 29 16:45:13 <_av500_> neem0: 3..2..1..go Jul 29 16:45:29 alrighty Jul 29 16:45:35 * nerdboy pulls the 'remove before flight" tags Jul 29 16:45:56 So, We've had the intro to BeagleSat a while back Jul 29 16:46:16 End goal -> Beaglebone based CubeSat platform Jul 29 16:46:23 what's new Jul 29 16:47:20 Went into a bit deeper into the correction algorithm and it seems possible I can optimize the fitting part by adapting the Ellipsoid fitting algorithm from our good friend Yury (obvious Russian math guy) Jul 29 16:47:52 Matlab code here -> https://github.com/nvisnjic/BeagleSat/blob/master/3D_fit_simulation/ellipsoid_fit.m Jul 29 16:48:41 I've also got a good confirmation that this stuff works quite well by finding it used in a different library with a similar purpose Jul 29 16:49:00 Namely RTIMUlib -> https://github.com/richards-tech/RTIMULib/blob/master/RTEllipsoidFit/ Jul 29 16:49:25 and you wrote some code... Jul 29 16:49:40 I've used it for testing before but only later realized that they use the same algorithm, albeit without the time variant correction part, that's the part I have to figure in Jul 29 16:49:49 yup getting to that now :) Jul 29 16:50:29 I've been working on the C API and made a basic design for the LKM module that's supposed to run the fitting algorithm later on Jul 29 16:51:41 Although we (my mentors :) ) decided that is going a bit to slow, so opted to make a first prototype for the data fitting and processing in Python since its quicker Jul 29 16:52:02 and can plug in nicely with Alexander's PyBBIO framework Jul 29 16:52:36 which we did, and now there's a MPU9250 library for PyBBIO ready to use Jul 29 16:52:39 code here -> https://github.com/nvisnjic/PyBBIO/blob/master/bbio/libraries/MPU9250/MPU9250.py Jul 29 16:53:03 It's still a bit rough, since I have to add the selftests and tidy it up Jul 29 16:53:31 but it's ready to do what it's supposed to, ei. get data to feed to the solver and fitting algorithm (which conveniently makes the selftests obsolete) Jul 29 16:54:27 next stops are tying the solver to the data input library and packing it together in an API which can be accessed easily Jul 29 16:54:54 is that realizeable in roughly 2 weeks? Jul 29 16:55:03 first in Python and then once we have something that works, going back to C for stability and portability (and because it's better that way) Jul 29 16:55:15 ds2: Python stuff, with a solver is doable in a week Jul 29 16:55:21 the solver is the trickiest part now Jul 29 16:55:44 it works for time invariant stuff, but thats not much more than the RTIMUlib library Jul 29 16:56:21 which is nice but not the thing I set out to make, since the time invariant correction is what really helps the CubeSat to get good and clean data easily Jul 29 16:57:01 so currently, I think the setup is to do the Python stuff as soon as possible and put as much back into C and the LKM Jul 29 16:57:20 and continue working on it after GSoC to finish it up properly Jul 29 16:57:40 adding support for other sensors in the future as well Jul 29 16:58:18 That's the overall idea and where we're at now Jul 29 16:58:49 questions? Jul 29 16:58:58 should I go into detail on some specific part Jul 29 16:59:17 <_av500_> questions anybody? Jul 29 17:00:13 13 min seems like an interesting number Jul 29 17:00:25 well if no one has any, I've got one Jul 29 17:00:52 Anybody have experience with eigenvectors? Specifically for how an eigenvector matrix translates to an Ellipsoid rotation matrix? Jul 29 17:02:11 neem0: Have a friend working with openCV and I remember that he is dealing with eigenvalues and eigenvectors alot Jul 29 17:02:13 Everyone off early to the cookies and milk table? Jul 29 17:02:25 not sure if it could help though. Jul 29 17:02:43 ebadawy: yeah, specifically I'm going through the code for the fitting I linked above Jul 29 17:02:49 https://github.com/nvisnjic/BeagleSat/blob/master/3D_fit_simulation/ellipsoid_fit.m Jul 29 17:02:55 (link again) Jul 29 17:03:05 <_av500_> neem0: thansk for the presentation Jul 29 17:03:09 <_av500_> I need to bail out Jul 29 17:03:12 <_av500_> soup is ready Jul 29 17:03:20 hmmm Jul 29 17:03:25 <_av500_> everybody, thanks and keep up the good work Jul 29 17:03:29 I could get you in touch with him and see if he can help. Jul 29 17:03:36 and while everything is perfectly fine and understandable, I'll need a rotation matrix from that eigen() function to get a complete fitting algorithm Jul 29 17:03:40 sure Jul 29 17:03:53 I mean, I'm asking in general, any links on the topic or something Jul 29 17:03:55 books even Jul 29 17:04:15 I'm happy to expand my horizons, and google didn't give me much to chew on on that correlation Jul 29 17:04:20 there's software for you when the python stuff is stable-ish Jul 29 17:04:39 _av500_: Thanks for the meeting and enjoy the soup! :) Jul 29 17:04:43 non-linear ellipsoidal stuff... Jul 29 17:04:50 nerdboy: well, the data input is stable Jul 29 17:04:55 I mean it works Jul 29 17:05:12 self calibrations gonna be fixed in the morning, but the fitting algorithm does that better than a selftest ever could Jul 29 17:05:15 neem0: I realy have no idea :) .. I'll check him up and see if he could help Jul 29 17:05:36 nerdboy: guess it makes most sense to work on the solver now, and when it works pack it with an API for general audiences Jul 29 17:05:42 self-test makes the python interface decoupled Jul 29 17:05:49 nerdboy: or did you have something else in mind? Jul 29 17:06:02 ahcorde: Bookmarked the link. Will use it as a reference Jul 29 17:06:11 nerdboy: come again? Jul 29 17:06:17 let's have a chat with alexanderhiam when he's available Jul 29 17:06:33 self-test here meaning hardware self test for the accel and gyro for the MPU9250 Jul 29 17:06:39 nerdboy: yeah, sure Jul 29 17:06:41 finish the python self-test and don't depend on the data fitting Jul 29 17:06:42 ! Jul 29 17:06:57 nerdboy: alexanderhiam -> when are you guys available Jul 29 17:07:03 nerdboy: roger that Jul 29 17:07:06 the latter is bonus when used/available Jul 29 17:07:09 rohith: we just send you the wrong one Jul 29 17:07:13 is the 21 one Jul 29 17:07:19 *2ยบ one Jul 29 17:07:49 ahcorde: This one right? https://github.com/erlerobot/erle_gitbook_basics_guide/blob/master/en/Cross_compiler/README.md Jul 29 17:07:57 https://github.com/erlerobot/erle_gitbook_basics_guide/blob/master/en/Cross_compiler/README.md Jul 29 17:08:00 imuguruza:^ Jul 29 17:08:03 yeap Jul 29 17:08:28 ebadawy: thanksa bunch, if he understands what the eigen() func does in Matlab and what the output means (evecs???), that would be grand! Jul 29 17:08:28 thanks a bunch even Jul 29 17:08:53 neem0: we need to re-assess priorities from last week Jul 29 17:09:07 nerdboy: ok Jul 29 17:09:16 ahcorde imuguruza: I realized that since the first one was a Google Docs spreadsheet :P Jul 29 17:09:17 nerdboy: what do you think I should focus on? Jul 29 17:09:26 exactly Jul 29 17:09:36 the MPU9250 data input in python is as good as done Jul 29 17:10:05 we need to look at what's there now Jul 29 17:10:19 last stuff is cosmetics, useful mostly for PyBBIO users only (ei ones not using my correction stuff, although we can integrate that into PyBBIO later, alexanderhiam ?) Jul 29 17:10:54 vamyoral|pc: How should I go about the project once I get the vehicle code compiling? Jul 29 17:10:54 nerdboy: solvers? or python input? Jul 29 17:11:12 vmayoral|pc:^ Jul 29 17:11:15 stable/decoupled python interface Jul 29 17:11:54 can you get enough test data via python interface for algorithm testing? Jul 29 17:12:08 nerdboy: yup, no prob Jul 29 17:12:21 I can work on that tomorrow after I pack up the self-tests in the lib Jul 29 17:12:23 via i2c or spi? Jul 29 17:12:26 spi Jul 29 17:12:50 SPI input works fine now, once I've done the black voodoo hopping InvenSense decided was neccesary Jul 29 17:12:50 vmayoral|pc: I have mentioned trying out mavros in the report. Jul 29 17:13:00 maybe work on python test/data interface for right now Jul 29 17:13:11 nerdboy: Ok Jul 29 17:13:22 i can begin transfering some of the API ideas to python Jul 29 17:13:29 and finish the PyBBIO library Jul 29 17:13:41 yeah, we have a dirty hack for sht10 on same bus with i2c devices Jul 29 17:13:43 but I'll need a solver to hook into all of that in the next week Jul 29 17:13:51 hi DiegoTc Jul 29 17:14:00 hi ebadawy Jul 29 17:14:13 jason gave the OK of your work Jul 29 17:14:15 is great Jul 29 17:14:37 you're almost finishing Jul 29 17:14:37 nerdboy: I mean this feels like a hack, but it's the official way InvenSense designed the device, SPI to the module, than I2C internally to the magnetometer Jul 29 17:14:38 :) Jul 29 17:14:43 solver we will have, young padawan... Jul 29 17:15:05 nerdboy: they failed to document the process and spectacularly failed at that Jul 29 17:15:08 haven't got his feedback yet. did he mailed you? Jul 29 17:15:23 nerdboy: I see masta, I shall speak of it no more Jul 29 17:15:25 :D Jul 29 17:15:52 * neem0 makes slow handwave gesture Jul 29 17:16:11 ebadawy: jkridner mentioned that he was travelling and may not be online Jul 29 17:16:26 * nerdboy should use chromebook next time for pirate hat/beard effects Jul 29 17:16:53 * moto-timo is in his own proprietary SPI hell Jul 29 17:16:53 rohith: yeah Jul 29 17:17:35 sht1x sucks, newer sht2x are actual I2C Jul 29 17:17:42 vmayoral: ping Jul 29 17:18:20 * nerdboy owns a "sensor stick" with sht10 and 2 other non-funky devices on the same pins Jul 29 17:18:26 DiegoTc: this week I suppose to do some tutorials, any ideas on how I should begin doing that? Jul 29 17:18:34 that's not fun... Jul 29 17:18:44 mmm Jul 29 17:18:50 ebadawy: how many weeks are left? Jul 29 17:18:55 2.5 Jul 29 17:18:59 I'd day 2 Jul 29 17:19:01 nerdboy: sensor... shtick? Jul 29 17:19:31 2 weeks (14 days) Jul 29 17:19:34 ? Jul 29 17:19:37 don't like to have a late work till the last day, prefer to leave it for any possible bugs Jul 29 17:19:42 yeah Jul 29 17:20:48 nerdboy: sht10? why? Jul 29 17:21:13 no idea Jul 29 17:22:03 sht10's are $$$$ :( Jul 29 17:22:10 one of those tindie "look ma, i made this board" things Jul 29 17:23:06 was a learning experience for me, no the least of which was what to look for before buying... Jul 29 17:23:13 *not even Jul 29 17:27:36 ebadawy: well you have two options then. Casa A & Case B Jul 29 17:27:45 Case A is to include the stop and run button Jul 29 17:28:03 Case B is to start preparing the site as the new home page of bone1010 Jul 29 17:28:06 bone101 Jul 29 17:28:43 for B I think jason have some concerns about the changes of the index page Jul 29 17:29:00 Ok Jul 29 17:29:01 we didn't finish the discussion about it last time Jul 29 17:29:09 * DiegoTc dreams goes awau Jul 29 17:29:22 ebadawy: well Jason is going of vacations Jul 29 17:29:25 this week Jul 29 17:29:31 but I quite got what he wants Jul 29 17:29:41 me too ,, 3 days off :) Jul 29 17:29:45 really Jul 29 17:30:05 i'm going on vacations but it's because need time for another project :( Jul 29 17:30:17 yeah, but I will try to have some work on them though :( Jul 29 17:31:05 he want this page http://beagleboard.org/Support/BoneScript Jul 29 17:31:17 modified Jul 29 17:31:30 and instead of showing the tutorials that are on the left Jul 29 17:31:32 _av500_: Who is up for presentation next week? Jul 29 17:31:41 he wants then on the new card system Jul 29 17:31:49 <_av500_> rohith: damn, forgot that Jul 29 17:31:55 <_av500_> rohith: can you do it? Jul 29 17:32:15 _av500_: Sure :) Jul 29 17:32:15 DiegoTc: so I am not making new tutorials? Jul 29 17:32:38 <_av500_> ebadawy: you also ok for next week? Jul 29 17:32:47 * _av500_ picks volunteers Jul 29 17:33:35 _av500_: no really, I may even have some difficulties to attend the meeting! Jul 29 17:33:45 <_av500_> ok Jul 29 17:34:08 <_av500_> azizulhakim: ping for next week? Jul 29 17:41:16 * ebadawy wonders if DiegoTc still there Jul 29 17:47:19 DiegoTc: will catch up with you later. Jul 29 17:47:28 _av500_: ping Jul 29 17:50:19 <_av500_> azizulhakim: ok for next week? Jul 29 17:50:25 <_av500_> to present Jul 29 17:50:36 yes, sure Jul 29 17:50:38 <_av500_> ok Jul 29 18:25:38 _av500_: How was the soup? :) Jul 29 18:26:04 nerdboy: alexanderhiam Anyway, ping me for a meetup time here or via email, anything works Jul 29 18:26:20 I'll be a bit out tomorrow morning, but should be back by US morning time :) Jul 29 18:28:47 * nerdboy is about to get sucked into another RfI response... Jul 29 18:29:00 just waiting for alexanderhiam to say something Jul 29 18:38:43 something Jul 29 18:39:08 * alexanderhiam is preoccupied and mostly afk today... Jul 29 18:45:15 this evening maybe? Jul 29 18:54:45 check email, i claim highest anti-priority today... Jul 29 19:00:43 * neem0 notice taken on email Jul 29 19:01:07 will follow instructions when wakeup time comes, going to shuteye mode now Jul 29 19:02:14 * nerdboy will shortly get bored with proposal stuff and look at some python code Jul 29 20:21:01 neem0: This is the mail of my friend alaaelnouby@gmail.com , his name is Alaa El Nouby you can send him your concerns and hope he could help. Jul 29 21:08:49 _av500_: what benchmark you wanted to see? Jul 30 00:06:57 neem0: *pushing and keeping us up to date please Jul 30 00:07:18 otherwise no cookie/soup for you... :( **** ENDING LOGGING AT Thu Jul 30 02:59:58 2015