**** BEGIN LOGGING AT Tue Jul 18 03:00:04 2017 Jul 18 03:10:32 hi nerdboy: Jul 18 03:11:01 I have completed the receiver side to do and would surely push code today Jul 18 03:11:33 also I dived into last year wave generation using PRU0 Jul 18 03:11:50 what's it like? :D Jul 18 03:12:11 it is cumbersome, and editing it requires lot of debugging Jul 18 03:12:41 also I wanted to ask in mail nerdboy: said that pwm drivers are not in good shape Jul 18 03:13:05 that's not quite what i said... Jul 18 03:13:33 but while using linux beaglebone 4.4.54-ti-r93, I was able to enable pwm Jul 18 03:13:41 sorry abt misinterpreation Jul 18 03:13:57 you can do it by hand Jul 18 03:14:53 but what does it mean that '#PWM's change on every bootup' Jul 18 03:15:23 I was using example from bb.org-overlays from here Jul 18 03:15:24 https://github.com/beagleboard/bb.org-overlays/blob/master/examples/cape-unversal-pwm.txt Jul 18 03:15:28 maybe with things that have been updated, but there's nothing dts/overlay-wise that really does anything useful Jul 18 03:17:01 i was trying to say (in the email) that *i think* the gpiochipN paths are static, with a given kernel/version and dt setup Jul 18 03:17:31 okk so it changes with kernel maybe Jul 18 03:18:01 but cant we during final stage make our own kernel image like beaglelogic type Jul 18 03:18:08 is it complex?? Jul 18 03:18:14 stuff like adafruit-beaglebone-io-python and bonescript has code to lookup the sysfs paths Jul 18 03:18:15 ot complex at all Jul 18 03:18:27 that is more of a philisophical/political question Jul 18 03:19:11 what i said, eventually a dts overlay and setup guide Jul 18 03:19:44 and also I am having a slight prob, my bbb cant connect to any git repository, it shows connection failed port no 443 etc. any idea?? Jul 18 03:19:55 if you use pwm you should probably use the the kernel interface Jul 18 03:20:47 depends on your network setup/hardware Jul 18 03:21:01 or if someone can help me to provide bbb internet access from mac, currently I have to use windows for this internet thing and it is very distasteful Jul 18 03:21:02 sometimes the route is funky Jul 18 03:21:17 I am able to ping 8.8.8.8 and every other thing Jul 18 03:21:38 if you have 2 default routs you should delete the one that shouldn't be there Jul 18 03:21:48 nerdboy: would you please tell me what do you mean kernel interface?? Jul 18 03:22:09 smthing abt setup?? Jul 18 03:22:19 the driver modules? headers/exported symbols, etc? Jul 18 03:23:01 okk Jul 18 03:24:39 the (broken) prudaq example code uses fileio on the sysfs entries Jul 18 03:25:16 a dts with pwmN aliases would make things cleaner Jul 18 03:27:20 which is kinda why i brought up pwm, since the working/cumbersome code avoids it Jul 18 03:27:54 you wanna rewrite it in pru C? Jul 18 03:28:09 write a cleaner version i asm? Jul 18 03:28:15 s/i/in/ Jul 18 03:29:05 i don't really care, but if you use asm you need to comment it really well Jul 18 03:29:49 but you do need something to test with and build on Jul 18 03:35:54 and you have several examples to choose from Jul 18 04:03:37 mine asm just switches mux and samples channels, pwm enabling is to be done by usin file io Jul 18 04:22:16 better get to testing then... Jul 18 04:33:39 something with enough info so i can test it Jul 18 05:36:55 thetransformerr[: whatever branch/code you want to use, i would forget dts/overlay for now and use a setup.sh like master has Jul 18 05:37:28 config-pin blah actually works Jul 18 17:58:29 hi nerdboy: Jul 18 17:59:06 I was working upon receiver side, at what clock rate should we sample with prudaq Jul 18 18:01:22 if we assume distance to be 15 cm and speed of sound to be 330 ms then it takes 4.5 milli seconds and Jul 18 18:02:22 with rate of 1 msps we get 4545 samples, with 400000 we get 1818 samples Jul 18 18:02:56 and also how did you find work around for ui-pruss module Jul 18 18:13:01 you didn't read the setup doc? Jul 18 18:13:57 use 4.9-bone/bone-rt kernel and disable the uboot_overlays Jul 18 18:15:25 without a correct dts, you need to use config-pin after loading universalh cape which frees up most of the pins Jul 18 18:16:08 i have black wireless Jul 18 18:16:15 uio-pruss is default on bone kernels, rproc is default on TI kernels Jul 18 18:16:51 i think it still uses hdmi pins so you need to disable that ^^ Jul 18 18:17:09 my kernel is 4.4.54-ti-r93 Jul 18 18:17:17 I have disabled them Jul 18 18:17:39 uio-pruss is disabled in ti kernels Jul 18 18:18:05 i tried it but it didn't work even after blaclisting the rproc stuff Jul 18 18:18:39 run the kernel script in /opt/something Jul 18 18:18:58 like it says on the bb.org overlays readme Jul 18 18:19:21 update kernel?? Jul 18 18:19:23 script Jul 18 18:19:29 yup Jul 18 18:19:54 it uses apt-get, easier than greping Jul 18 18:20:41 just follow the doc, that's why it's there Jul 18 18:22:05 copy the setup.sh script from master and add your pwm setup stuff there Jul 18 18:23:48 okk setup.sh from src/ Jul 18 18:25:30 from src/prudaq/firmware Jul 18 18:25:33 i done it and I am now removing earlier stuff Jul 18 18:26:04 the commented out stuff, it just sort of confuses :| Jul 18 18:29:31 in what, uEnv.txt? Jul 18 18:29:55 setup.sh Jul 18 18:30:26 delete it then, but i would leave a note about not using overlays Jul 18 18:31:02 yeaH i have left it, citing the reason for using config pin Jul 18 19:50:48 try (40 kHz x 5) so start around 200 khz or so Jul 18 19:51:41 you'll be sampling a bunch of harmonics and ringing without filtering Jul 18 23:08:38 thetransformerr[: did you use spice to model your input chain? Jul 18 23:08:58 http://www.ecircuitcenter.com/Circuits/instamp1/instamp1.htm <= this looks good Jul 18 23:09:12 takes 3 opamps though... Jul 18 23:09:57 but, easy to scale the output to 0-1.8V or so, then apply prudaq vref to center it Jul 18 23:10:32 *VOFF(set) in the spice model Jul 18 23:18:07 what op-amps are being considered? Jul 18 23:22:28 spice thingie says OPAMP MACRO MODEL, SINGLE-POLE Jul 18 23:23:00 probably depends on what he can get his hands on... Jul 18 23:23:23 should watch out for non ideal characteristics of the op-amps Jul 18 23:23:30 the resistors/caps need to be pretty good Jul 18 23:23:32 a 741 might not be a good idea Jul 18 23:24:02 well, check out that instrumentation amp model and recommend something Jul 18 23:24:31 still need filters too Jul 18 23:24:40 I can give you part numbers but donno if it can be gotten Jul 18 23:24:52 LMV324 might work and may not be too painful to source Jul 18 23:51:13 i think the note about using a pot for R2 makes sense for tuning out variations Jul 18 23:52:44 what about Q for that filter? seems like <=10 should work Jul 18 23:54:37 how fast is that LMV324? at least 10MHz? **** ENDING LOGGING AT Wed Jul 19 03:00:03 2017