**** BEGIN LOGGING AT Wed Jul 12 03:00:05 2017 Jul 12 06:04:37 hmmm Jul 12 15:31:31 ugh.... E14 CANCELLED order # 17750136 :'( Jul 12 15:33:08 why?? :( Jul 12 15:33:51 jkridner: please try with other supplier, e14 messed up earlier also Jul 12 15:34:05 what did they mess up earlier? Jul 12 15:34:13 taxes? Jul 12 15:34:30 I had to get TI India involved in the last order. :( Jul 12 15:34:42 apart from taxes they delayed the order for abt one month Jul 12 15:35:07 for tax id Jul 12 15:35:43 after all tax issue, they were unable to transfer order intra-state, Jul 12 15:36:29 ouch Jul 12 15:37:16 Abhishek_: do you have any other suggestion for supplier?? Jul 12 15:37:37 which board is it? Jul 12 15:37:55 last time they told me that they were b2b organisation Jul 12 15:38:29 exact details are available with ds2:, but mainly it is a ultrasonic transducer board Jul 12 15:38:39 from maxq Jul 12 15:39:16 Have you tried KitsNspares website for individual ordering? Jul 12 15:39:32 nope Jul 12 15:39:37 They may be able to get you something but it will be another 14 days Jul 12 15:40:20 http://shopping.kitsnspares.com/Element14/ Jul 12 15:40:27 Site is currently down for maintenance Jul 12 15:40:47 Use the exact element14 part number there and you might be able to order Jul 12 15:41:20 I would discuss with my mentors abt h/w if they can provide some sort of remote access Jul 12 15:41:45 Yeah, in the meantime until you get stuff sorted Jul 12 15:43:57 ds2: is it possible that we can bypass chip in HC-SR04 board like maxq Jul 12 15:43:58 thetransformerr[: who has it for sale and can ship right away in India? Jul 12 15:44:48 Are Digi-Key or Mouser better? Jul 12 15:45:42 last time you sent PRUDAQ, I received it without any hassles Jul 12 15:45:54 It was from which supplier?? Jul 12 15:46:09 although they took their time Jul 12 15:46:35 what if we tried to ship ds2's boards via fedex or whatever? Jul 12 15:47:06 this is good idea I think... Jul 12 15:47:40 they are pre-hacked even Jul 12 15:49:13 other option would be to make some analog circuits to interface the murata sensors Jul 12 15:49:40 too much hardware work unless you have local "expert" Jul 12 15:50:39 analog circuits also seem good idea, if it can be implemented with some sort of IC's Jul 12 15:51:26 the example ones have transistors/opamps/coils/etc Jul 12 15:51:52 better to work on software and try fedex Jul 12 15:52:20 I can find expert, but last problem I guess would be the quality of connections Jul 12 15:53:13 that stuff is breadboard-able but it's not software Jul 12 15:55:22 but breadboard can reduce the time spent in waiting for h/w Jul 12 15:56:04 for now, can you use lab equipment to test bb/prudaq? Jul 12 15:56:32 as in signal generator/scope Jul 12 15:56:35 m_w, jkridner: BeagleLogic Standalone's schematic is 80% complete. Jul 12 15:57:08 cool. definitely need to swap for OSD3358-SM. I'll e-mail Greg now. Jul 12 15:57:25 I can use signal generator, but it would require some regulation Jul 12 15:57:56 meaning what exactly? Jul 12 15:58:07 jkridner: you mentioned earlier that you'd get me some additional hardware upon request in return for the customs I paid on the initial shipment Jul 12 15:58:11 and we have 2 mins to discuss... Jul 12 15:58:28 was just wondering if I could get a microSD card or two, would make dev work a little easier Jul 12 15:58:41 one hardware blocker at a time... Jul 12 15:58:53 ee: k. can you get your mentor to add it to the list? Jul 12 15:58:55 hopefully that one is easier... Jul 12 15:59:08 Abhishek_: cool Jul 12 16:00:19 m_w, jkridner : Here's a PDF link to it - https://drive.google.com/file/d/0B7U2bJEjkNeZZlVONUstREE0VUE/view?usp=sharing Jul 12 16:00:32 * nerdboy rings the mtg gong Jul 12 16:00:43 all hands on deck Jul 12 16:00:44 hi Jul 12 16:01:00 that's one Jul 12 16:01:03 hi Jul 12 16:01:11 two Jul 12 16:01:23 hi all Jul 12 16:01:29 hello Jul 12 16:01:31 jkridner: sounds good, I'll talk to him Jul 12 16:01:33 hello :) Jul 12 16:01:35 Hi everyone Jul 12 16:01:55 are we short one? Jul 12 16:01:57 Hello !! Jul 12 16:02:25 no i think everybody is present Jul 12 16:02:31 one more plus reports? Jul 12 16:02:47 everybody is here Jul 12 16:02:51 hi all Jul 12 16:04:27 6 reports even Jul 12 16:04:40 check Jul 12 16:04:51 who's up today? Jul 12 16:04:58 me and pmezydlo Jul 12 16:05:15 can you be first maciejjo? Jul 12 16:05:19 * nerdboy hands maciejjo the talking stick Jul 12 16:05:19 yes np Jul 12 16:06:24 since last report i worked mostly on DMA transfers scheduled from PRU Jul 12 16:07:28 I created PoC application which receives transfer parameters such as buffer size and address of kernel memory fro kernel module Jul 12 16:07:59 and configures EDMA for transfer, triggers to start, and waits for completion Jul 12 16:08:39 now with that working I started refactoring the code to create a library that can be linked with pru applications Jul 12 16:09:22 so I can create custom apps, link with pru-dma lib and reuse dma functionality for different purposes Jul 12 16:10:04 i added example where I transfer some buffer from kernel to PRU and blink LEDs with patterns defined in it Jul 12 16:10:56 now I want to refactor the kernel module so it can be also more reusable, allowing other modules with specific functionality be built on top of it Jul 12 16:11:08 maciejjo: how are configuring the parameters on the PRU side ? Jul 12 16:12:23 sending a message for that ? Jul 12 16:12:30 zeekhuge: on PRU i wait on mesage from kernel and when recievied I write them to PaRAM slot which is also configured from kernel Jul 12 16:13:05 yes, I created structure that is defined both in module and PRU fw, and just pass pointer to this structure to rpmsg Jul 12 16:13:27 Ideally there should be not involve sending message, and a driver should somehow do it. Jul 12 16:13:49 like the remoteproc configures vring buffers while booting it up Jul 12 16:14:28 using resource_table Jul 12 16:14:34 like this : https://github.com/ZeekHuge/BeagleScope/blob/port_to_4.4.12-ti-r31%2B/firmware/resource_table_pru0.h#L75 Jul 12 16:14:52 so that its ready as soon as it boots up. Jul 12 16:15:03 so I would need to add custom entry to resource table Jul 12 16:15:13 yeah. Jul 12 16:15:20 with definition of dma config Jul 12 16:15:30 right ! Jul 12 16:15:57 ok, good, I will try to change it to this way Jul 12 16:16:34 and the kernel module will write the exact details while it boots up the PRUs. Jul 12 16:18:10 ok, I will read rpmsg driver to see how it is done there, and model my solution based on this Jul 12 16:18:35 ah, no .. Jul 12 16:18:50 you'll probably find it in the remoteproc driver Jul 12 16:19:08 oh ,ok Jul 12 16:19:36 got a plan for this week? Jul 12 16:20:17 and nothing blocking? (didn't see anything) Jul 12 16:20:46 yes, work on making kernel module generic, so it can be used from other modules. currently it implements the LED example and nothing else. Jul 12 16:20:56 no, nothing blocking me Jul 12 16:21:10 it is just elf symbols on the PRU binary Jul 12 16:21:17 check Jul 12 16:21:52 pmezydlo: are you ready? Jul 12 16:22:00 yeah I'm ready Jul 12 16:22:10 Welcome on my short standup Jul 12 16:22:15 you have the stick... Jul 12 16:22:22 This week I have worked on SPI ip core and linux kernel driver. Jul 12 16:22:36 My current work concerns the idea of reprogrammable hardware. Jul 12 16:22:46 It’s very interesting. I will try to briefly describe it to you. Jul 12 16:22:55 It's * Jul 12 16:23:04 In short - CPU peripherals are customized to board work environment. Jul 12 16:23:20 For example, when more PWM channels or additional SPIs, UARTs are needed, you can load a different configuration to FPGA. Jul 12 16:23:35 Ip cores are mapped in ARM memory so a kernel driver can be written for it. Jul 12 16:23:46 During gsoc I'm creating ip cores and kernel support for these cores. Jul 12 16:23:57 Until now I wrote SPI, PWM, GPIO core and SPI driver. Jul 12 16:24:06 kinda like the opencores drivers? Jul 12 16:24:15 yeah Jul 12 16:24:30 GPIO has its driver in kernel tree. Now I'm working on PWM driver. Jul 12 16:24:47 Thanks to my kernel drivers, ip cores can be used from userspace as a typical GPIO, PWM, SPI etc. Jul 12 16:25:03 Each FPGA ip core consists of a component and a top module which maps the component to memory. Jul 12 16:25:23 To add next component to the existing one, an offset needs to be added to the new component. Jul 12 16:25:41 are fpga expensive then processors or they are almost same but with differnet utility Jul 12 16:26:20 similar Jul 12 16:26:30 no, ice40 isn't expensive Jul 12 16:26:37 thetransformerr[: you can get that FPGA in a usbstick dev board for about $25-$30 Jul 12 16:27:13 It's very simple so I don't plan to change anything. Jul 12 16:28:00 In a few days I will start working on UART module Jul 12 16:28:29 I think thats all thank you for attention Jul 12 16:28:51 Do you have any questions? Jul 12 16:29:06 can what you are doing be adapted to the PRU? Jul 12 16:29:38 * nerdboy notes the existing tools/ip cores are somewhat less than accessible for most people... Jul 12 16:29:42 Hmm yes I think yes Jul 12 16:30:06 if PRU has access to gpmc memory area Jul 12 16:31:03 I will try communication with PRU soon Jul 12 16:31:34 * nerdboy trying to remember the _av500_ agenda items... Jul 12 16:31:58 who's up for reports next week? Jul 12 16:32:09 me Jul 12 16:32:10 ds2: do you have any idea what to use PRU and ice40 for? Jul 12 16:33:09 pmezydlo: for me - cameras Jul 12 16:33:13 need one more... ravikp7? indu? Jul 12 16:33:29 pmezydlo: I need the ice40 to make the bus more friendly for the PRU Jul 12 16:33:55 it's indu's turn next week Jul 12 16:34:40 technically we don't need to keep the same order, but that might be "easier" Jul 12 16:35:11 indu: you and thetransformerr[ next week? Jul 12 16:35:34 I'm ready, if one's short Jul 12 16:35:45 yes Jul 12 16:35:55 check Jul 12 16:36:29 okay, i think we covered the important bits Jul 12 16:38:29 good work, got a full compliment of students/deliverables this week Jul 12 16:38:58 just one last shipping hump to squash... Jul 12 16:39:50 * nerdboy turns it over to general discussion Jul 12 16:41:17 thetransformerr[: i think you need to stick with the software/lab equipment for now Jul 12 16:42:47 tommrow is not possible for me to reach lab but on friday, I would make test with scope Jul 12 16:43:24 and until then I would iron out any problem due to kernel file system Jul 12 16:44:38 do I need to flash bbb with some customized kernel for pru support from pasm and using pru library from TI Jul 12 16:46:30 no, the bone kernel works for uio_pruss Jul 12 16:46:45 just install the latest one Jul 12 16:46:55 okk Jul 12 16:47:27 jkridner: any ideas up for fixing spl device opening issue on Windows and OSX, i think removing one device configuration isn't going to help, as for spl transfer both Windows and OSX required rndis init through usb control transfers on the interface other than one for bulk transfer Jul 12 16:48:10 thetransformerr[: latest 4.4 can have mdp's ecap but not 4.9 Jul 12 16:48:18 ravikp7: but, didn't that extra control transfer get added? Jul 12 16:48:38 it doesn't forward-port to 4.9 without some serious refactoring Jul 12 16:48:50 ecap is not required for now.. Jul 12 16:49:03 nope Jul 12 16:49:25 jkridner: shouldn't it be required again for spl rndis device? Jul 12 16:49:52 ecap was required when we planned to bypass adc, which we are not Jul 12 16:51:31 i was testing 4.9.35-ti-rt-r44 and bone 4.4 i think Jul 12 16:51:53 ravikp7: I'd assume so, yes. Jul 12 16:53:39 seems pretty slow booting up... Jul 12 16:54:25 I would first try with default and if it fails try with another image Jul 12 16:56:00 you can't use ti kernel for uio_pruss Jul 12 16:56:36 jkridner: what makes the behavior different in Linux vs Windows / OSX for the rndis device initialization ? Kernel code ? Jul 12 16:56:42 prudaq_capture depends on it so bone kernel is a req Jul 12 16:57:54 even if you enable the old driver in ti kernel and blacklist rproc it still won't work Jul 12 16:58:13 okk then done I would try to make round robin run to its max Jul 12 17:00:26 jkridner: Can you test libusb through BBBlfs on OSX and confirm that it also isn't able to open spl device ? On Windows I can't even get the SPL transfer done through BBBlfs, it gets very slow in file transfer, eventually packets get lost and transfer fails Jul 12 17:00:43 * nerdboy pinging wormo about hangout Jul 12 17:02:18 jkridner: Just to make sure problem isn't with node-usb. On Windows, USB analyzer tool can access all configurations but node-usb can't Jul 12 17:04:36 jkridner: And one more problem, on Windows even USB analyzer can't get device descriptor for initial rndis device when BB is connected to a USB 2.0 port. Jul 12 17:05:03 *config descriptor Jul 12 17:05:43 Everything works fine on Linux. Jul 12 17:06:50 What to blame on Windows? it's Kernel? Jul 12 17:06:56 zeekhuge: regarding dma configuration in resource table, I assume I need to do some modifications to pru_rproc driver? Jul 12 17:07:33 I see eg. interrupts are handled here: https://github.com/beagleboard/linux/blob/4.9/drivers/remoteproc/pru_rproc.c#L625 Jul 12 17:08:33 maciejjo: yeah. Jul 12 17:08:44 so .. the PRU binaries are in ELF format. Jul 12 17:09:09 and .. IIRC the pru_rproc parses that Jul 12 17:09:22 to put the buffer address and all there .. Jul 12 17:09:39 you will probably have to investigate into it Jul 12 17:10:03 the best way would be to look how vring buffers get added there .. Jul 12 17:12:49 I will dig into it Jul 12 17:15:47 ravikp7: i just blame windoze period Jul 12 17:16:58 thetransformerr[: can you bring Wormo up to speed? Jul 12 17:17:14 up to speed?? Jul 12 17:17:17 * nerdboy needs to make more coffee Jul 12 17:17:22 status-wise Jul 12 17:17:43 is he present here ?? Jul 12 17:17:49 wormo: hi..... Jul 12 17:17:50 here Jul 12 17:18:30 so we are making anemometer to measure wind speed and for that we have to use ultrasonic transducers Jul 12 17:18:54 hardware is in limbo, he has some lab access, need to test what can be tested... Jul 12 17:19:08 she knows that part... ;) Jul 12 17:19:21 thetransformerr[: yup, I sent you the fake data Jul 12 17:19:40 so you don't have to start quite that far back ;) Jul 12 17:19:41 *that* wormo... Jul 12 17:19:57 okk Jul 12 17:20:03 :) Jul 12 17:21:02 so after that we are going to use some sensor for temperature measurement and Jul 12 17:21:25 also want to interface murata with PRUDAQ and PRU 0 output Jul 12 17:22:15 so we should wait for maxq board or go ahead with some home-made circuit using transistors,op-amps Jul 12 17:23:20 do you have PRUDAQ up and capturing straight from PRU 0 yet? Jul 12 17:24:16 I havent tried that yet but that should work even without interface, I guess with wire and some resistance may be Jul 12 17:25:27 The previous year project isn't quite what you want for PRU 0 yet Jul 12 17:25:53 resistance shouldn't matter unless you are doing something i am not aware of Jul 12 17:26:06 assuming you are still thinking burst of waves rather than continuous Jul 12 17:26:21 also nerdboy pointed out the frequency wasn't quite right Jul 12 17:26:45 and jumping into someone else's PRU project isn't trivial Jul 12 17:26:47 last year project had to change the flow abruptly ftom ths 1206 to prudaq, so it sort of mixed up things Jul 12 17:27:12 True Jul 12 17:27:29 I recommend getting that sorted out while waiting for your board Jul 12 17:27:36 but I have used round-robin example from PRUDAQ and edited it to take 6 sample in round robin manner Jul 12 17:28:02 thetransformerr[: do you have access to a signal generator? Jul 12 17:28:23 it seems pretty straight to me than last year project Jul 12 17:28:26 Yes I have Jul 12 17:28:39 i wonder about the sample rate and its consistancy Jul 12 17:28:41 thetransformerr[: looks can be deceiving till you're into it Jul 12 17:29:02 yes I am into it and came out today only Jul 12 17:29:08 once you have it sampling, it may not hurt to run a tiny sine wave into it and plot the result Jul 12 17:29:20 make sure it is what you expect Jul 12 17:29:34 I would do that, and then start tweaking the PRU code Jul 12 17:29:46 okk that sounds good, I would surely do that Jul 12 17:29:50 and if you know what a FT is, take that result and run it through a FT and look at the harmonic content Jul 12 17:30:21 there might be some function already present in python for ft Jul 12 17:30:26 (for this purpose, the FT doesn't have to be a FFT...anything that gets yu a plot) Jul 12 17:30:53 true about python Jul 12 17:31:29 counting numpy of course Jul 12 17:31:49 I have pushed some code today, on github branch https://github.com/thetransformerr/beagle-sonic/tree/thetransformerr-method-1/round-robin%20method Jul 12 17:32:11 please see the diffrence between original and my modification Jul 12 17:32:17 commit Jul 12 17:33:44 and for this code , I have thought that to measure TOf we will connect all tx input on channel 0 and all rx output on channel 1 Jul 12 17:33:54 the prudaq branch from last year actually works Jul 12 17:34:07 base on round-robin also afaict Jul 12 17:34:14 *based even Jul 12 17:34:49 I would try that today and tommrow, it might require some t=minr tweaks regarding dts I guess Jul 12 17:35:10 extra points for code reuse Jul 12 17:35:19 :) Jul 12 17:35:47 * nerdboy would start there before the generic example... Jul 12 17:36:14 Can I assume we all know what I want to do with round robin example Jul 12 17:36:38 I didn't get why you don't want to alternate like the code currently mentions Jul 12 17:36:49 or I just provide some overview? Jul 12 17:36:57 there's a time aspect, but you're the one doing it Jul 12 17:37:09 "connect all input on channel 0" etc Jul 12 17:37:27 no atually channel 0 has 4 inputs Jul 12 17:37:33 EE hat on - Signal quality is better if it is coming from the SMAs Jul 12 17:37:34 0,1,2,3 Jul 12 17:38:21 actually, no. there are really only 2 inputs....the rest are muxes Jul 12 17:38:36 both those code branches already switch between inputs Jul 12 17:39:00 no inputs are 8 but adc can sample two inputs at a time Jul 12 17:39:09 prudaq branch just skips the second channel to do one axis for now Jul 12 17:39:12 one from each channel Jul 12 17:39:16 different discussion.... get something working and we can go from there (that's my opinion) Jul 12 17:39:57 why do you think it shold be different from current examples? Jul 12 17:40:24 okk so the hump we are facing actually is how to interface transducer Jul 12 17:40:37 or we bypass it using generator Jul 12 17:40:44 can the prudaq directly deal with the small signal? Jul 12 17:41:03 and how many amplitude bits is needed by the alg? Jul 12 17:41:08 that's an open question Jul 12 17:41:09 nope i guess Jul 12 17:41:37 the output of rx side is somewhere between 100 and 6 mV Jul 12 17:41:38 wormo: sent me data with 10 bits I guess Jul 12 17:42:03 *according to data sheet/user guide Jul 12 17:42:14 s/is/should be/ Jul 12 17:42:20 what's VREF? Jul 12 17:42:36 1 volt i think Jul 12 17:43:03 okay so at 100mV, and VREF 1V; we have about 7 bits of data Jul 12 17:43:34 do we need an input preamp to scale that up a little? Jul 12 17:43:35 at ~10mV, we have about 3.5bits of data Jul 12 17:43:46 will the alg handle this or does it get lost in the "noise"? Jul 12 17:44:09 nerdboy: that would help or a transformer but getting the HW to the student has been a challenge Jul 12 17:44:19 i don;t have a feel for that so i would test Jul 12 17:44:33 3.5 bits could be doable, depending on how crazy the transfer function is for mangling the input pulses Jul 12 17:44:33 <-- donno the alg stuff Jul 12 17:44:56 the preamp could be like one or two transistors i think Jul 12 17:44:58 it should handle it as long as correlation finds some similarity Jul 12 17:45:00 the data you had was kinda weird Jul 12 17:45:04 but, Jul 12 17:45:18 depends on what thetransformerr[ can scrounge in the lasb Jul 12 17:45:19 I agree correlation should probably work Jul 12 17:45:39 since pulse/not-pulse was distinct Jul 12 17:45:42 sure, a reasonably quiet NPN can give us the gain we need Jul 12 17:46:13 even a 2N3904/2N2222 could probally do it...common emitter config Jul 12 17:46:27 for this one off, a no feedback config should suffice Jul 12 17:46:49 AC couple it in and out Jul 12 17:47:10 transistors are cheap and available locally also, but I cant say abt quiet transisitors Jul 12 17:47:25 Wormo: it was very wierd... but I donno a good way to track down the source Jul 12 17:47:44 thetransformerr[: do you guys use the 2N system or the weird BCxxx system? Jul 12 17:48:43 I am sorry but are they abt the code of electronic device Jul 12 17:48:52 yes Jul 12 17:49:20 transistors in the US are commonly identified by 2nxxxx....but i think europe uses something like BCxxx Jul 12 17:49:29 this is all ignoring the other series like the TIP, etc Jul 12 17:49:50 in any case, if we go down that route, we can look up specs Jul 12 17:50:03 the 2n2222/2n3904 are generic ones Jul 12 17:50:11 I dont know abt transistors code, pls wait I guess I have a transistor in my rack, Jul 12 17:50:14 so we need some "lab" tasks that are useful Jul 12 17:50:35 what would be really handy right now is a PSOC Jul 12 17:50:55 or even a MSP430 with a digitally controlled op-amp Jul 12 17:51:01 yeah they are weird bcxxx Jul 12 17:51:12 let's start with measuring the signal coming out of the prudaq tx Jul 12 17:51:20 the one I have right now has code BC547 Jul 12 17:51:24 thetransformerr[: will keep that in mind... just need to look them up Jul 12 17:52:11 OKK As nerdboy: says, the signal for tx would be coming from pwm oin Jul 12 17:52:28 the BC547 can work... it is a NPN... probally similar to the 2n2222 Jul 12 17:52:38 thetransformerr[: what does your round-robin branch have for output Jul 12 17:53:00 PWM pin p9_42 Jul 12 17:53:17 the prudaq branch says "connect transmitter to P8.42" Jul 12 17:54:08 okk i just shot about pin no, but output is from pwm Jul 12 17:54:08 so can you measure the output of both code bases? Jul 12 17:55:18 so voltage measurement is from that pin and I have set duty cyle as 50% so voltage must be half of its max capacity Jul 12 17:55:20 now I cant but on friday I would Jul 12 17:57:09 I don't follow voltage will be half of max, aren't you looking at it on a scope? Jul 12 17:57:43 if vref == 1V, please use a 1:3 attentuator before feeding it to the ADC Jul 12 17:57:53 ADCs get upset when Vin > Vref Jul 12 17:58:16 prudaq wants 0-2V input Jul 12 17:58:40 no I was talking about measurement if done with multimeter, it averages Jul 12 17:58:57 and attentuator can be?? Jul 12 17:59:34 and vref is 1.0V Jul 12 17:59:36 resistance or some other device??? Jul 12 18:00:13 resistor devider Jul 12 18:00:23 divider Jul 12 18:00:42 2 series resistors Jul 12 18:01:08 that I was saying earlier in discussion, a wire and resistance Jul 12 18:02:00 so we just separate the output fro pwm and feed it to prudaq after attentuation and tx after amplification Jul 12 18:02:43 and amplification is not to be bothered abt if we are going to use maxq Jul 12 18:02:47 don't want to steal a bit? Jul 12 18:03:45 sorry but that means what ?? Jul 12 18:03:54 quality.. Jul 12 18:04:52 feed in the PWM signal on say bit 15 of PRU0 (or one of the unused ones0 Jul 12 18:05:19 since you sample by saving r30; that automatically puts the PWM signal in your sample... one less then to unwrap Jul 12 18:05:36 all you are going to get is a noisy square wave if you run it through the ADC Jul 12 18:06:37 yeah we can do that but that requires tweaking assembly code and i was planning that tweak once we get something working Jul 12 18:06:56 Ok Jul 12 18:07:18 once we get this working, I can write code pru0 such that we do not require pwm module Jul 12 18:07:30 so tx is set Jul 12 18:07:53 and for rx I should transistors before feeding it to adc? Jul 12 18:12:43 that's kinda why i asked for some code analysis on prudaq branch Jul 12 18:13:07 it's time for a todo list with priorities perhaps nerdboy? Jul 12 18:13:13 i can't really tell what it's doing without teaching myself asm Jul 12 18:13:53 i think it's doing its own burst signal on that pin Jul 12 18:13:59 *without pwm Jul 12 18:14:11 okk I would email you asm code with comments tomm Jul 12 18:14:49 both mine modified and last year asm code Jul 12 18:15:05 Wormo: can i use that usb widget to record tx output? Jul 12 18:15:32 basically in mine for waves we are using pwm and cross correaltion for tof Jul 12 18:16:27 last year he used the registers of pru to record timestamp of burst for threshold method Jul 12 18:17:30 so as you say last year code, enables pin for some time, and performs some functions Jul 12 18:17:39 I would email you both Jul 12 18:17:54 so for now lets make to do list for this week Jul 12 18:18:17 okay, then one more todo should be a recommendation on which branch is more accurate/precise/maintainable Jul 12 18:18:22 yes nerdboy that's the kind of thing I loaned it to you for Jul 12 18:18:49 I want to be on schedule by working hard Jul 12 18:19:25 s/branch/approach/ maybe Jul 12 18:19:36 so first is to set up pwm drivers and other prudaq setup Jul 12 18:20:14 for that nerdboy: you recommended using bone kernel 4.4 Jul 12 18:20:59 if you want to use prudaq without beaglelogic device then you pretty much need bone kernel/headers Jul 12 18:21:10 secondly as ds2: said feed signal generator and find the waveform reported by adc Jul 12 18:21:36 and do some ft Jul 12 18:21:59 4.4.75-bone-rt-r18 is what i have now Jul 12 18:22:16 4.9 should also work but i dodn't test it Jul 12 18:22:26 I can clone it from... Jul 12 18:22:37 4.9-x-bone-blah that is... Jul 12 18:22:46 it's already in the repos Jul 12 18:23:12 apt-cache search linux-image to find the latest version number Jul 12 18:23:19 any link if you have, or I would find it easily? Jul 12 18:23:47 on your bb use apt-get to install Jul 12 18:24:16 okk Jul 12 18:24:19 building from source i would use bb-kernel build scripts Jul 12 18:24:57 all of the available kernels are already built/packaged and in the rcn repo Jul 12 18:25:15 all f the common ones anyway Jul 12 18:25:50 there are also scripts in /opt/ for installing kernels and such Jul 12 18:25:59 and then thirdly we test the tx output from pwm using wire and resistor divider Jul 12 18:26:10 they just use apt-get anyway Jul 12 18:27:05 so I should able to use them after a search and apt-get Jul 12 18:27:26 tx out -> voltage divider with max 2V -> (delay?) -> rx in Jul 12 18:27:39 is that what we're talking about? Jul 12 18:28:07 yes from me Jul 12 18:28:24 how good is the signal generator? Jul 12 18:29:09 can you trigger it and add delayed/small fake data signal on top of pwm output? Jul 12 18:31:14 i never thought abt this but may be its possible, I was thinking of using rx transducer with transitor\ Jul 12 18:32:08 this is the exact generator in lab http://www.amazon.in/Scientech-Function-Generator-Frequency-Counter/dp/B014WD6JBE Jul 12 18:32:43 Scientech 4060 is an extremely versatile Function Generator. It has a wide frequency range with LCD display for frequency and amplitude. The output amplitude can be varied upto 30 Vpp (open circuit) and can be attenuated to -60 dB. All outputs are short-circuit proof. It has a switchable 50 W / 600 W output. TTL / CMOS out is available for Digital Lab experiments. And also it has a 50 MHz Frequency Counter Jul 12 18:32:44 in-built with Function Generator. Additional features indicating its high quality include constant amplitude flatness and low distortion factor of signals. It is a new laboratory standard offering low cost solution to all your needs. Jul 12 18:34:17 FM input on the pic, so it probably can produce modulation on top of a base signal Jul 12 18:34:18 SSLError: [Errno bad ca_certs: '/usr/local/lib/python2.7/dist-packages/requests/cacert.pem'] [('system library', 'fopen', 'No such file or directory'), ('BIO routines', 'BIO_new_file', 'no such file'), ('x509 certificate routines', 'X509_load_cert_crl_file', 'system lib')] (file "/usr/local/lib/python2.7/dist-packages/requests/adapters.py", line 447, in send) Jul 12 18:35:09 * thetransformerr[ posted a file: 1mhz-function-generator-with-50-mhz-frequency-counter.pdf (225KB) Jul 12 18:35:09 SSLError: [Errno bad ca_certs: '/usr/local/lib/python2.7/dist-packages/requests/cacert.pem'] [('system library', 'fopen', 'No such file or directory'), ('BIO routines', 'BIO_new_file', 'no such file'), ('x509 certificate routines', 'X509_load_cert_crl_file', 'system lib')] (file "/usr/local/lib/python2.7/dist-packages/requests/adapters.py", line 447, in send) Jul 12 18:35:25 datasheet of generator, I will dig out further Jul 12 18:35:25 SSLError: [Errno bad ca_certs: '/usr/local/lib/python2.7/dist-packages/requests/cacert.pem'] [('system library', 'fopen', 'No such file or directory'), ('BIO routines', 'BIO_new_file', 'no such file'), ('x509 certificate routines', 'X509_load_cert_crl_file', 'system lib')] (file "/usr/local/lib/python2.7/dist-packages/requests/adapters.py", line 447, in send) Jul 12 18:37:16 gotta run... will check back later Jul 12 18:37:16 SSLError: [Errno bad ca_certs: '/usr/local/lib/python2.7/dist-packages/requests/cacert.pem'] [('system library', 'fopen', 'No such file or directory'), ('BIO routines', 'BIO_new_file', 'no such file'), ('x509 certificate routines', 'X509_load_cert_crl_file', 'system lib')] (file "/usr/local/lib/python2.7/dist-packages/requests/adapters.py", line 447, in send) Jul 12 18:37:59