Wednesday, 2016-08-10

*** tpb has joined #timvideos00:00
*** cr1901 has joined #timvideos00:10
*** puck has quit IRC01:35
*** puck has joined #timvideos01:39
xfxfthaytan: i think audiorate / queue might have fixed it \o/02:19
xfxfi bet after i've said that it starts losing sync again02:19
xfxfthanks for your help03:10
*** sb0 has joined #timvideos03:18
*** rohitksingh_work has joined #timvideos03:50
mithrohey rohitksingh_work03:52
rohitksingh_workmithro: hi!03:52
mithrorohitksingh_work: RattusRattus had some questions about design decisions around the vmodvga03:53
rohitksingh_workmithro / RattusRattus: sure!04:03
mithroRattusRattus: You seem to understand power supplies better then me, looking at - there is FB5, what is its primary purpose?04:19
*** rohitksingh_wor1 has joined #timvideos04:21
*** rohitksingh_work has quit IRC04:23
mithrocr1901: you around at all, or already in bed?04:48
cr1901mithro: Now is not a good time:
mithrocr1901: Well, the frequency counter seems to be working pretty well - so thats some good news04:50
cr1901That's wonderful! My next plan was to replace the minispartan's clock PLL with DCMs. Then debug the EDID problems. Then work on data islands04:51
mithrocr1901: The data island stuff I think we should discuss before you do any work on04:56
mithrocr1901: As I think I mentioned I have a bunch of thoughts on rewriting the HDMI subsystem04:56
cr1901mithro: I don't remember tbh, but that's okay. Not prepared to have this convo tonight, unfortunately :(04:56
mithrocr1901: Yeah thats fine04:57
mithroCarlFK: ping?05:00
CarlFKmithro: here05:00
mithroCarlFK: so there were a couple of revisions I wanted you to check carefully - I'm assuming you didn't do anything about that?05:01
CarlFKmithro: not yet - it's near the top of my list... I did fix my  tester to deal with the new (or old?) version strings05:02
CarlFKtrying to get the log file posted... can't find it05:02
mithroCarlFK: Let me send you and Ryan an email with the revisions / versions05:14
tpbTitle: Ubuntu Pastebin (at
mithroCarlFK: git pull the prebuilt repo05:35
mithroCarlFK / xfxf: you have email05:40
*** danielki has joined #timvideos06:03
mithroCarlFK: does that make sense06:10
CarlFKmithro: the email?  I think so.06:11
CarlFKmithro: more test results
tpbTitle: Ubuntu Pastebin (at
mithroCarlFK: great, the version shouldn't have "-dirty" in it from now on06:20
CarlFKim off to bed.. I'll try to get to this tomorow. assuming 100 other things don't get dumped on me again...06:22
*** Bertl_oO is now known as Bertl_zZ06:40
*** paradisaeidae_ has joined #timvideos06:45
mithroRattusRattus: poke me when you are around07:23
*** paradisaeidae_ has quit IRC07:23
mithroxfxf: BTW - Did you see the following script -> ./scripts/download-prebuilt.sh07:36
*** ssk1328 has joined #timvideos07:52
*** miselin has quit IRC08:07
*** miselin has joined #timvideos08:08
*** miselin has joined #timvideos08:08
RattusRattusmithro: sorry I got nobbled by my wife to review some of her work last night....  fortunatly I am ahead in %dayjob right now  (finished today's tasks already and now waiting on managment) so I will do the schematic updates now.08:23
RattusRattuslooking at your HDMI interface FB5 is there to both snub conducted emmissions on the HDMI cable from entering the unit, and stop noise going the other way.08:25
RattusRattusthis is fairly standard practice.08:25
RattusRattusgrrr  /me also hates that the GND net doesn't have a symbol only a net name (pet hate)08:27
*** guest_ has joined #timvideos08:28
* RattusRattus likes to see all power net connections with a symbol not a net name because a well drwan schematic can be used to guide board layout (i.e. where the symbol is placed is where I expect the connection to the power plane on the board.... (the impeadance of vias is such that their placement can have a large effect on EMC performance)08:29
RattusRattusbut in this case I suspect it is because the person drawing the schematic hasnt set altium up to prevent poer nets being global and they want the domain to be restricted to just this sheet.08:31
RattusRattus(handy for inheritance of sheets in multiple designs)   anyway I am waffling......08:31
* RattusRattus is supprised that there isn't a bulk decoupling capacitor next to FB5. I would certainly have placed one there08:32
mithroRattusRattus: Do you have much experience with FPGA power supply designs?08:34
RattusRattusa bit.  I have used a few FPGAs in my past (just not written any Verilog / VHDL)08:39
*** sb0 has quit IRC10:53
*** cr1901 has quit IRC11:56
*** cr1901 has joined #timvideos12:11
RattusRattusjust spotted somthing on the power supplies I should have seen earlier :-(12:20
RattusRattusanalogue system is going to be very noisy.  the rails VD and PVD have big power loops back to the 12V rail.12:22
RattusRattusAll of the return current from these two supplies is forced through the ADC :-(12:22
RattusRattusyou should have a star point at the power supplies.12:23
RattusRattusbut looking at the ADC it doesn't seporate analoge and digital supplies :-(  so perhaps better to just use a single ground (perhaps with 'trenches' to guide eddy currents12:24
*** sb0 has joined #timvideos12:39
*** Bertl_zZ is now known as Bertl12:45
*** guest_ has quit IRC13:05
*** rohitksingh_wor1 has quit IRC13:09
RattusRattusmithro: ping13:32
mithroRattusRattus: pong13:32
RattusRattusI have updated schematics and added a couple of parts to the libuary.13:32
mithroRattusRattus: okay13:32
RattusRattusI guess I could do with walking through upload process please13:32
mithroRattusRattus: so, you have a local git checkout?13:34
RattusRattuswell I grabbed as a zip...13:35
RattusRattusnot having done git before13:35
mithroI actually have no idea what grabbing as a zip gets you - let me take a look13:36
mithroRattusRattus: first, "apt-get install git gitk"13:36
RattusRattus'doze here (work)13:36
mithroRattusRattus: well,
mithroRattusRattus: looks like when you download as a zip you just get a "snapshot" of the repo and none of the git stuff13:38
mithroRattusRattus: so, you'll need to clone the repo, copy your changes over top of it, commit and then push13:41
RattusRattusok.  I guess that is just the VGA connector schematic and libuary that I have changed (although the timestamps indicate all files have been touched in the board directory)13:42
RattusRattusso I have a MING64 'terminal' shell window...13:42
mithroRattusRattus: okay13:45
mithroRattusRattus: go to
tpbTitle: GitHub - timvideos/HDMI2USB-TOFE-VGA: A VGA capture board for use with TOFE-8x compatible boards like the Numato Opsis. (at
mithroRattusRattus: click the "fork" button up in the corner13:46
RattusRattusone second creating github account....13:47
mithroRattusRattus: okay13:47
RattusRattusyay fork now works :-)13:48
RattusRattusI guess I clone at this point?13:50
mithroRattusRattus: yeah, clone down to your local disk13:50
mithrogit clone
RattusRattusok, then overwrite into my local repository the files I have changed....13:52
RattusRattusfirst find local repository on this box :-)13:52
*** rohitksingh has joined #timvideos13:54
mithroRattusRattus: yeah - you might need "unix2dos" or "dos2unix" tools to get the files back into the right line endings13:55
mithroRattusRattus: "git add <files>; git commit" then you can do a "git push" to send them up to github13:55
RattusRattusyep commit in unix line format has already been selected13:55
RattusRattusstill trying to find where the mount point was created for the ming64 shell13:56
mithroRattusRattus: I don't know much about ming64 / windows stuff13:56
RattusRattusok /me is reading . searching13:58
mithroRattusRattus: there is a thing called "TortoiseGit" which gives you everything in the rightclick menu14:00
RattusRattusyes but half the problem here appears to by out IT deptment locking shit down....14:01
RattusRattusworking on it14:01
RattusRattusfound files (and have write access now)14:06
RattusRattusright updating local...14:06
RattusRattusok looks good in local copy14:08
*** danielki has quit IRC14:09
mithroRattusRattus: what is your github user?14:11
mithroRattusRattus: once you have them in your own repo, you then send a "pull request" using the github ui14:12
RattusRattusso I don't do a git add for each file I have changed?14:13
RattusRattusthen a commit?14:13
mithroRattusRattus: you do a "git add" for each file/change you want to add to this commit14:13
mithroRattusRattus: then you do a "git commit" to save that into git14:13
mithroRattusRattus: then a "git push" to send it to the remote server14:13
RattusRattusok doing tht bit now14:13
mithroRattusRattus: so if you wanted to seperate it out into multiple commits, you can just "git add a; git commit; git add b; git commit"14:14
RattusRattusthe git push?14:19
mithroRattusRattus: yeah14:20
mithrogit push sends it to the remote server14:21
mithrowhich in this case is github14:21
mithroRattusRattus: yay, I see you have one commit that I don't now!14:22
RattusRattusok that should be the libuary,  it should have also been the schematic but that didn't get updated (my bad)14:23
mithroRattusRattus: thats an expensive ESD array you have there -- USD1.70 :P14:24
mithroRattusRattus: you also only added it to the -cache.lib - did you create separate parts in a library?14:25
CarlFKmithro: im up.  I'll be trying to look at images soon.. whats your ET Bed?14:28
mithroCarlFK: soonish14:29
mithroCarlFK: less then a couple of hours14:29
CarlFKk, I can get to something in an hour14:29
CarlFKbowl of cereal first14:29
RattusRattusJust checking wich lib I added into...14:35
*** ssk1328 has quit IRC14:35
mithroRattusRattus: it's easy to accidently add it to the system libraries on Windows as they aren't read only like on Linux14:37
RattusRattusno I had added it to a project libuary.... just hadn't pushed it :-(14:38
RattusRattusmy problem was that I had older files than the pull because I did the edit from a snapshot.  that should go away now14:39
RattusRattushopfully thats what you wanted.14:39
mithroRattusRattus: I'll find out14:40
RattusRattusI can look for a much cheeper ESD protection for the composite inputs...14:40
RattusRattusanyway I must go - I am due at the hospital shortly14:40
mithroRattusRattus: okay, the next step would be sending a pull request, but github lets me do that for you.14:41
RattusRattusmust go (running late already)14:41
mithroRattusRattus: thanks! I'll take a look shortly14:42
mithroCarlFK: rapidly fading here15:15
CarlFKmithro: loading firmware...15:28
CarlFKmithro: is this expected?   git describe: v0.0.2-61-ga02ee1a-dirty15:29
CarlFK(first of the 3 you listed)15:29
mithroCarlFK: hrm?15:30
CarlFKmithro: mainly the -dirty - I think that went away in a later version, just making sure15:30
mithroCarlFK: yes, I didn't fix that until after those revisions15:31
CarlFKmithro: um.. test pattern is very .. not good.15:34
CarlFKlets see if I can screen shot this...15:34
mithroI'm heading to bed15:37
CarlFKk - I'll have soemthing when you wake up15:38
mithroCan you also check the latest version?15:46
CarlFKand yes, ill check latest version15:55
*** ssk1328 has joined #timvideos17:18
*** sb0 has quit IRC17:32
*** rohitksingh1 has joined #timvideos19:20
*** rohitksingh has quit IRC19:21
*** rohitksingh has joined #timvideos19:27
*** rohitksingh1 has quit IRC19:32
*** rohitksingh has quit IRC19:49
*** Bertl is now known as Bertl_oO20:43
*** ssk1328 has quit IRC22:36
mithroCarlFK: morning?22:56
mithroCarlFK: Do your BlackMagic cards accept HDMI?22:56
CarlFKmithro: hey - just getting back to testing atlys22:56
mithroCarlFK: I was wondering if maybe you could extend your script so that you have an Atlys connected to your BlackMagic card and do a capture of the pattern output on each revision22:57
mithroCarlFK: and also do a capture the pattern via the USB22:57
CarlFKah, interesting..22:57
mithroCarlFK: Would be even more awesome if you switched to a plugged in device and did a capture of that22:57
mithrotumbleweed: Hope you get better soon. Also hope my packaging jobe doesn't make you sick again :P22:57
CarlFKmithro: few 'problems' with your vision22:58
mithroCarlFK: IE It would be really good if you could help me understand which firmware versions have what problems22:58
CarlFKbut I like the vission22:58
mithroCarlFK: Only a few? :P22:58
CarlFKmainly how to keep the results organized22:59
CarlFKI suspect we will end up with 100 pngs and not really sure what we are looking at22:59
CarlFKalso, capturing an image always seems to have issues.  like timing or something23:01
tumbleweedmithro: it's not sick so much as deeply unmotivated for work, and trying to find ways back in to motivation23:01
CarlFKall of this is solvable, just takes time and effort23:02
mithroCarlFK: keep them organised via the git revision number23:02
mithroCarlFK: yeah, the BlackMagic cards are also quite sensitive23:02
mithroCarlFK: an alternative is to point a webcam at monitor23:02
mithroCarlFK: Ideally I'd like it to all be automated and run automatically when I commit23:02
mithroCarlFK: But for now, if we can get something you can "wire up" and then run the script for a couple of hours and collect some results, that would be good23:03
mithroCarlFK: specially with the prebuilt stuff making it quick to get images for each revision too23:03
CarlFKmithro: I can barely get the 3 things you emailed me about done23:03
mithroCarlFK: Yeah - that is why you "wiring it up and leaving it running" type thing would be good :-P23:04
CarlFKI am still swamped with grinding out videos from the 3 shows (or 4?  I lost track)23:04
CarlFKmithro: can you add a version string to the test pattern ?23:07
mithroCarlFK: Possibly - but if the video output is all corrupted like your test pattern, then it isn't going to be readable23:08
mithroCarlFK: which is why I suggest using what ever your script reads from git or the board23:09
CarlFKmithro: yeah.. some are just a little23:09
mithroCarlFK: knowing which versions are corrupted and good is what we are trying to do :)23:09
CarlFKmithro: I don't expect to have anything useful for days.. maybe like 1023:11
CarlFKin the mean time, being able to see the ver on the screen may help23:11
mithroOkay, it is likely that the next ~5-6 days is the most time I'll get to work on fixing major issues in the next ~3 months and I'm more likely to be testing on Opsis boards.23:15
mithroCarlFK: also can't really go back and update those revisions23:17
CarlFKgood point.23:17
CarlFKbut for future revs..23:17
CarlFKmithro: here is another flavor of problem
CarlFKwatch at most 5 seconds23:19
CarlFKAtlys, whatever rev I have been using for the last few months23:19
CarlFKnot sure how useful that is, but it was handy23:19
CarlFKmithro: another thing.. on the rev I am testing, if I do "m 9" over and over, each time the noise changes23:21
CarlFKit is still about the same amount, just moves around23:22
xfxfare you talking about the picture going all funny?23:22
xfxfmithro and i did a bunch of tests the other day23:22
xfxfturns out input0 on a few of my boards does that23:22
xfxfand only with particular cables and laptops23:22
xfxfyou're making art23:23
mithroCarlFK: Can you try a difference Atlys board and make sure it happens on that too?23:26
mithroCarlFK: I'll also try and reproduce it here later today23:26
CarlFKmithro: I knew you would ask that... and yes, I have a 2nd one handy23:26
CarlFKmithro: new thing..  I am not sure I have ever seen it hang before, but m 9  and the prompt hasn't come back and I think the stream stopped23:28
tpbTitle: Ubuntu Pastebin (at
CarlFKyeah, gst isn't counting frames read from v4l23:28
mithroCarlFK: is that the terminal connected / video won't stream bug or?23:28
CarlFKin that case, close the term, start the stream, start the term , all is well23:29
CarlFKthis is: streaming, enter commands (see paste) and suddenly both term and stream stop23:30
CarlFKhmm maybe host problem...23:31
CarlFK[158253.240760] uvcvideo: Found UVC 1.00 device - Numato Opsis Board (1d50:60b7)23:31
CarlFK[158253.245917] uvcvideo: UVC non compliance - GET_DEF(PROBE) not supported. Enabling workaround.23:31
CarlFK[158253.247056] cdc_acm 1-1:1.2: ttyACM0: USB ACM device23:31
CarlFK[185008.044753] ------------[ cut here ]------------23:31
CarlFK[185008.044787] WARNING: CPU: 2 PID: 1090 at /build/linux-dcxD3m/linux-4.4.0/ubuntu/i915/intel_pm.c:3586 skl_update_other_pipe_wm+0x16c/0x180 [i915_bpo]()23:31
tpbTitle: Ubuntu Pastebin (at
CarlFKmithro: want me to poke at anything before I power cycle everything ?23:35
mithroCarlFK: no - sorry we are packing here at the moment23:43
CarlFKmithro: no prob23:44
CarlFKmithro: you aren't able to do much for a few hours right ?  (I need to get out to the store, either now or in a few hours)23:51
*** cr1901 has quit IRC23:57

Generated by 2.13.1 by Marius Gedminas - find it at!