**** BEGIN LOGGING AT Mon Aug 04 02:59:59 2014
Aug 04 05:19:17 <praveendath92> vvu: ping
Aug 04 05:19:35 <praveendath92> av500: https://www.youtube.com/watch?v=ukQJ9EKEcm4
Aug 04 05:27:43 <Abhishek_> praveendath92: is that related to synchronization?
Aug 04 05:38:09 <av500> praveendath92: is that realtime?
Aug 04 05:38:24 <av500> screen updates seem to happen fast
Aug 04 05:53:13 <praveendath92> av500: Yes. Realtime.
Aug 04 05:54:04 <av500> so apart from the glitches, it should work nicely?
Aug 04 05:54:27 <praveendath92> av500: It will work nice. Glitches are expected.
Aug 04 05:54:31 <praveendath92> I mean I know the reason.
Aug 04 05:54:59 <praveendath92> I was encoding the page addresses into the data sent to device.
Aug 04 05:55:32 <praveendath92> So, it is being drawn from top, till bottom and that is reapting.
Aug 04 05:56:37 <praveendath92> Abhishek_: Only the updated parts of a frame are sent to device. I wasn't using the coordinates of the data. Hence the flickering and all.
Aug 04 05:56:51 <praveendath92> Once I do that, it should work good.
Aug 04 12:33:51 * karki makes it to round 2 of ebay
Aug 04 12:34:19 <karki> panto : my driver is now modular :)
Aug 04 13:35:38 <alexanderhiam> karki: no meeting today?
Aug 04 13:35:54 <karki> I don't know :/
Aug 04 13:36:02 <alexanderhiam> hmm
Aug 04 13:36:56 <karki> alexanderhiam : I've barely code communication with jason in past 1.5 weeks
Aug 04 13:37:11 <karki> and no email saying we were supposed to meet
Aug 04 13:39:02 <alexanderhiam> karki: ok, well I've done a read through your soft pwm code, I can at least give you some feedback on that
Aug 04 13:39:14 <karki> :)
Aug 04 13:40:22 <alexanderhiam> currently your resolution is 10%, it should really be 1% resolution if you're setting it by percent, or it could be 1 bit resolution if you're setting it by an 8-bit value
Aug 04 13:42:08 <karki> alexanderhiam : what do you mean by 10% resolution? (I'm confused with the terminology here)
Aug 04 13:43:16 <alexanderhiam> your high time is set in 10ths of the total period
Aug 04 13:43:40 <karki> umm no
Aug 04 13:43:49 <karki> it is split into 100
Aug 04 13:44:24 <karki> hi time can be between 0/100th to 100/100th of  the time period
Aug 04 13:44:42 <karki> it was 10% initially
Aug 04 13:44:52 <karki> maybe I forgot to update my comments
Aug 04 13:44:56 <karki> but it is 100 parts
Aug 04 13:45:05 <alexanderhiam> ah, nevermid then, the description in pru1_firmware.h implied 10
Aug 04 13:46:26 <alexanderhiam> yup, RESOLUTION certainly is 100, I misread that
Aug 04 13:47:45 <alexanderhiam> it would also be nice if the frequency was configurable somehow, e.g. if you want to drive a servo motor you want it at ~50Hz
Aug 04 13:51:50 <karki> alexanderhiam : that should be configurable with small modifications :)
Aug 04 13:52:15 <karki> that is essentially what chris wanted right? the TONE inst.
Aug 04 13:53:16 <alexanderhiam> I think the TONE instruction is a square wave generator on a single pin, where as the PWM frequency would be the frequency of all the PWM outputs
Aug 04 13:54:01 <alexanderhiam> what's the lowest frequency you could run at with 1 timer overflow / time unit?
Aug 04 13:55:31 <karki> lowest frequency would be around 1-4hz I guess
Aug 04 13:55:37 <karki> not sure I'll check
Aug 04 13:55:57 <karki> I think I'm currently running stuff at 1KHz
Aug 04 13:56:04 <karki> *5Khz
Aug 04 13:57:17 <alexanderhiam> I think 5 or 10kHz would be a fine maximum frequency, and 1-10Hz would certainly be a fine minimum
Aug 04 13:58:05 <alexanderhiam> though the 1% resolution will be a bit low at lower frequencies. Does botspeak use percentage?
Aug 04 13:58:25 <alexanderhiam> looks like it does
Aug 04 14:03:18 <alexanderhiam> karki: also, how fast do you think you could drive the ARM GPIO pins for PWM?
Aug 04 14:03:37 <karki> Isin't ARM GPIO not fit for that?
Aug 04 14:03:53 <karki> I don't know, but I thought it would be indeterminate
Aug 04 14:04:57 <karki> alexanderhiam : is the frequency guaranteed - like wont there be varying bus latency delays?
Aug 04 14:05:10 <alexanderhiam> might not be good for it, I'm just wondering for future expandability.
Aug 04 14:05:31 <karki> I'm really not sure with the performance on the ARM side!
Aug 04 14:05:48 <karki> worth discussing though!
Aug 04 14:06:07 <alexanderhiam> I'm not sure how guaranteed ARM memory access is from the PRU
Aug 04 14:06:31 <alexanderhiam> just toggling a pin gives hardly any jitter though
Aug 04 14:07:12 <karki> alexanderhiam : Is there any PyBBIO + android thing you are thinking about?
Aug 04 14:07:24 <panto> everything that gets routed out of the PRU incurs unpredictable latency
Aug 04 14:07:29 <karki> I have to do a mini project in android
Aug 04 14:07:37 <panto> I doubt it's ever going to be up to the 1ms ranges though
Aug 04 14:08:17 <karki> panto : so PWM via GPIO is feasible as a future option?
Aug 04 14:08:57 <panto> sure
Aug 04 14:09:14 <panto> jitter is only but an engineering parameter
Aug 04 14:11:24 <alexanderhiam> karki: you mean porting PyBBIO to android? I assume all the drivers are different, that would probably be a headache
Aug 04 14:16:48 <Abhishek_> karki: The GPIO writes are "posted" and there's an internal queue
Aug 04 14:17:05 <Abhishek_> karki: Congrats on Round 2 :)
Aug 04 14:17:15 <karki> alexanderhiam : no, not really porting to android
Aug 04 14:17:38 <karki> just remember you having a mobile interface too....
Aug 04 14:18:17 <Abhishek_> karki: so one off writes would be no latency, but it would gradually settle to a fixed latency when it's being continously hit, the latency will be of the order of tens of ns
Aug 04 14:18:46 <karki> Abhishek_ : that isn't a problem :) ns is fine!
Aug 04 14:19:00 <panto_> it all depends whether there's other traffic on the interconnect fabric
Aug 04 14:19:25 <panto_> so while it would be normally a few ns, there _could_ be spikes in the usec range
Aug 04 14:19:45 <panto_> in fact you have to ask a TI hardware engineer to answer correctly
Aug 04 14:22:12 <alexanderhiam> karki: I'm not following the android thing... what are you thinking?
Aug 04 14:29:22 <Abhishek_> karki: you want to install android on your BBB or, hack a cheap android phone for some purpose?
Aug 04 14:32:44 <karki> Abhishek_ : just for fun :) want to try android on the BBB ;)
Aug 04 14:32:58 <Abhishek_> cool
Aug 04 14:33:42 <karki> alexanderhiam : I had in mind like an android API to use the BBB - send commands and stuff.
Aug 04 14:33:53 <karki> e.g. I want to build an home automation app
Aug 04 14:34:03 <karki> androif -> BBB over wifi
Aug 04 14:34:17 <DiegoTc> karki: you could use Zigbee for that
Aug 04 14:34:23 <karki> and pyBBIO handles all the IO stuff
Aug 04 14:34:23 <DiegoTc> it's more secure
Aug 04 14:34:52 <karki> DiegoTc : well the Zigbee is in the pipeline for PyBBIO.
Aug 04 14:34:59 <DiegoTc> karki: could connect the xBee series 2 with the BBB
Aug 04 14:35:07 <DiegoTc> and it will work
Aug 04 14:35:08 <alexanderhiam> karki: gotcha. Yeah, having a RESTful API for PyBBIO is on my list of things to do at some point
Aug 04 14:35:29 <DiegoTc> karki: I did it with arduino and it work excellent!
Aug 04 14:35:54 <karki> DiegoTc : but how do I get an android phone on board?
Aug 04 14:36:05 <karki> thats why I thought WiFi :)
Aug 04 14:36:12 <alexanderhiam> or BT
Aug 04 14:36:53 <Abhishek_> karki: Android phone, Use a BT module like HC05 and pair it with an Arduino
Aug 04 14:37:27 <Abhishek_> then pass IO cammands around
Aug 04 14:37:30 <karki> yes, but BT modules are low range
Aug 04 14:37:44 <karki> WiFi is more universal I feel
Aug 04 14:37:56 <Abhishek_> external antennas are always welcome :)
Aug 04 14:38:03 <karki> Abhishek_ : thats what I do normally
Aug 04 14:38:15 <DiegoTc> well I did this way, I have an app that I send the commands  to the XBee Coordinator, via RESTAPI, I did it this way because it work anywhere, I was at college and could turn off lights :)
Aug 04 14:38:15 <Abhishek_> I see
Aug 04 14:38:44 <Abhishek_> interesting
Aug 04 14:38:53 <DiegoTc> I thought of doing wifi, but there were some security issues
Aug 04 14:39:34 <alexanderhiam> yeah, beaglebone as XBee coordinator with wifi/ethernet user acces seems like a good setup for home automation stuff
Aug 04 14:44:36 <Abhishek_> although an Android phone + Arduino clone will get one a co-ordinating display and GSM data capabilities as well
Aug 04 14:44:46 <Abhishek_> more cost-effective
Aug 04 14:46:31 <alexanderhiam> the beaglebone could serve up a real nice web gui to the user though, and have cron jobs to turn on/off lights, make sure doors are locked, etc.
Aug 04 14:49:08 <Abhishek_> with a little manipulation it might be even possible to get Android phones do it as well IMO
Aug 04 14:49:48 <Abhishek_> I've seen an Android app that streams the cam over a web server and comes with a web interface
Aug 04 15:20:09 <karki> Can anyone help me with an idea for an android app that involves Hardware? (I have a few s/w based ones)
Aug 04 16:45:08 <mranostay> morning
Aug 04 16:45:23 <mranostay> Abhishek_: what is the CPU usage for the web app btw?
Aug 04 17:20:40 * rseethamraju is frustrated with gstreamer!!
Aug 04 17:50:53 <Abhishek_> mranostay: the CPU usage isn't very significant
Aug 04 17:52:01 <mranostay> Abhishek_: networking i/o probably gets you though?
Aug 04 17:52:59 <Abhishek_> Yeah, since we are quite limited to 1-2k of samples
Aug 04 17:54:27 <Abhishek_> And the fact that the logic is in the web page, the app on the BBB is just a static file server with an extra websocket link
Aug 04 17:55:44 <Abhishek_> I would be more inclined, given my current experience with Node.JS and allies to completely rewrite the WebSocket backend that gives me more control than what I currently have
Aug 04 17:58:48 <Abhishek_> I don't know if it is due to me being a JS noob or the fact that these libs have not been optimized for this kind of data
Aug 04 18:00:48 <Abhishek_> Hi jkridner
Aug 04 18:00:54 <jkridner> hi
Aug 04 18:01:44 <jkridner> rewrite the websocket backend?
Aug 04 18:02:16 <jkridner> you measured throughput with other TCP/IP-based protocols and see that socket.io is the bottleneck?
Aug 04 18:03:04 <Abhishek_> I did measure the link performance with iperf and found that maximum possible throughput that can be expected is 80Mbps
Aug 04 18:03:17 <jkridner> k.
Aug 04 18:03:18 <Abhishek_> I told you this a week or two ago
Aug 04 18:03:22 <jkridner> sure
Aug 04 18:03:49 <jkridner> but, are you somehow feeding data from /dev/beaglelogic?
Aug 04 18:04:19 <jkridner> perhaps a binary protocol would be better? are you currently using socket.io or something else?
Aug 04 18:04:25 <Abhishek_> and, when I'm trying to pipe the sigrok-cli output through socket.io, then it does not come as one big packet
Aug 04 18:04:54 <Abhishek_> socket.io buffer support serializes a buffer into [255 100 23 45 .....] so is grossly inefficient
Aug 04 18:05:26 <jkridner> right.
Aug 04 18:06:27 <jkridner> maybe switch to nginx for serving the websocket?
Aug 04 18:06:46 <jkridner> or http://binaryjs.com/
Aug 04 18:06:54 <Abhishek_> jkridner: there was a KickStarter called Red Pitaya
Aug 04 18:07:10 * jkridner vaguely remembers that
Aug 04 18:07:23 <Abhishek_> they use jQuery flot and a nginx-based backend
Aug 04 18:07:32 <jkridner> oh yes.
Aug 04 18:08:41 <Abhishek_> I'll look into BinaryJS
Aug 04 18:09:39 <Abhishek_> My internet connection is very laggy, lot of latency and missed IRC messages since I left home
Aug 04 21:02:27 <debian_> hi
Aug 04 21:57:52 <Abhishek_> jkridner: Able to render output from sigrok-cli by converting the ascii string to WaveDrom format
Aug 04 23:04:37 <Abhishek_> jkridner mranostay: http://s15.postimg.org/jqfvy4wcp/Screenshot_from_2014_08_05_04_32_50.png
Aug 04 23:23:30 <mranostay> ah cool
Aug 04 23:25:15 <ds2> have you tried using a decoder with it yet?
Aug 04 23:26:13 <Abhishek_> ds2: the backend at the moment is just piping sigrok-cli output over to the desktop and converting it into WaveDrom markup
Aug 04 23:27:11 <Abhishek_> atm sigrok-cli does not support annotated (sample index) decoded data
Aug 04 23:49:04 <Abhishek_> I could roll a forked sigrok-cli to work around it and get the web interface to do some protocol analysis, but I am not sure how effective it could be and how to configure all of it
Aug 04 23:49:54 <ds2> ah
Aug 04 23:50:24 <ds2> time is running short, probally shouldn't get to creative within the time frame of GSoC
Aug 04 23:51:19 <Abhishek_> I have the same opinion
Aug 04 23:51:49 <Abhishek_> I am probably going to clean it all up at this state and get it into the system image
Aug 04 23:52:24 <Abhishek_> Install this as a service, systemctl stuff...
Aug 05 01:59:49 <rseethamraju> alexanderhiam: ping!
Aug 05 02:01:52 <rseethamraju> right now it streams, records and takes snapshots.
Aug 05 02:02:20 <rseethamraju> streaming it can start and stop and start again
Aug 05 02:02:41 <rseethamraju> snapshots also it can take any number whenever required
Aug 05 02:03:11 <rseethamraju> only recordng I haven’t figured how to stop and start recording in a different file
Aug 05 02:03:49 <rseethamraju> I mean I know how to do it. searching for the right python equivalent
Aug 05 02:11:57 <rseethamraju> adn I just pushed it onto the ‘cam’ branch
**** ENDING LOGGING AT Tue Aug 05 03:00:00 2014