Tuesday, 2015-04-28

*** tpb has joined #timvideos00:00
*** CarlFK has joined #timvideos00:20
*** ChanServ sets mode: +v CarlFK00:20
CarlFKnerp starting, andrew and scope are sitting next to me, Ed is starting meeting00:21
*** jck has quit IRC00:23
*** jck has joined #timvideos00:24
CarlFK"we have data"00:34
CarlFKand my laptop is only 1/2 wonky.00:34
CarlFK"huh, 3 states?"  (sounds bad.)01:08
mithroI'm around now01:20
CarlFKmithro: k - we are all watching scheduled presentation now.   for another 10-20 min I am guessing01:25
CarlFKmithro: meeting wrapping up, i2c poking soon (few min)01:52
CarlFKmeeting over, scoping starting02:00
CarlFKmithro: http://www.digilentinc.com/Data/Products/ATLYS/Atlys_rm_V2.pdf02:08
CarlFKcan you find the i2c pins we should be using?02:08
CarlFKthere was some guessing02:08
cfeltonCarlFK: are you probing for SI or function?02:16
CarlFKcfelton: um.. what?02:16
CarlFK"it is the right size for a data bit, but it is going the wrong way"02:19
cfeltononly JA (the small connector) has the I2C signals going directly to a header02:19
cfeltonyou can look at the HDMI OUT (J2) on JP7 and JP602:20
cfeltonthese are buffered, you will not be able to tell if it is a signal integrity (SI) issue02:21
cfeltonbut you can see if they are toggling, functionally, as expected02:21
CarlFKcf from the scope: https://www.dropbox.com/lightbox/home/Public/atlys/OAKTON02:29
cfeltonstill loading02:36
CarlFKchanged hdmi source.  "seems to look better" but he is ... dialing it in02:37
cfeltonthe link is not working for me?02:39
CarlFKah, hold..02:40
CarlFKcfelton: https://www.dropbox.com/sh/80q7dtrlrc32ydw/AACNAh_VRuZjP-75yI3556Mpa?dl=002:41
mithroCarlFK: I sent you instructions with the jumpers labelled02:43
CarlFKah right, I was about to print that and the cat started bleeding :p02:44
mithroWhat is with those little bumps?02:44
mithroCarlFK: see page 13 of the manual02:45
*** hyades has joined #timvideos02:47
CarlFKmithro: bumps?02:49
cfeltonon the falling edges, when it is supposed to be driving it to ground (i assume)02:50
cfeltonthe scope screen shots don't contain any of the bit-errors that occurred in the CSV?02:51
mithroCarlFK: there should only be two voltage levels, 0 and 1 - there seems to be like a 0.1V bumps02:52
CarlFK"oh, this is not ground.  that's a problem."02:52
cfeltonthe master in this case is failing to drive the sink the signal all the way to ground but I don't think that is the case02:53
cfeltonyou can see that in the first byte, scope decodes it ok, below the threshold02:54
*** Niharika has joined #timvideos02:55
CarlFK"different ground, still see the tri-state"  (the bump)02:55
cfeltonthere is a bit error in byte 10 but the scope only shows address byte, byte[0] then skips up to byte[16]02:55
cfeltonthey should share a common ground, ground signals in the cable02:56
cfeltonsometimes the open-collector (or emulated) is week and won't pull it all the way to ground (picture an extra resistor in the master)02:57
cfeltonmithro: there is a csv with the complete caputre in the dropbox link, tek0000I2C.csv02:58
mithroThe first link doesn't work for me, says "no such folder"02:59
CarlFKyeah, that's for when I am logged in.  ignore that one02:59
CarlFKcfelton: https://www.dropbox.com/sh/80q7dtrlrc32ydw/AACNAh_VRuZjP-75yI3556Mpa?dl=0   <- should have 2 png and a csv03:00
cfeltonCarlFK: yup, the first png shows the address and the frist tx byte (byte[0])03:01
cfeltonthe second png shows bytes 16-2603:01
CarlFKa passer by says "those bumps are small enough nothing should see them as data"03:02
cfeltonagreed, showed be ok03:02
cfeltonbumps mainly occur on the ACKs (when the master drives to 0 end of a TX byte)03:03
cfeltonneed a waveform capture of a corrupted byte :)03:03
cfeltoncan you get a zoom in of one of the data bytes?03:05
CarlFKyes.  zooming...03:05
CarlFKtrying to find a corrupted byte03:09
CarlFKimage coming03:12
CarlFK2 new images are up03:15
CarlFKis that zoomed in enough?03:17
cfeltonwonder why it sends those single byte requests?03:18
cfeltonseems odd03:18
CarlFKcan we pack up tonight?03:18
CarlFKcfelton: mithro ping - we are hungry :)03:22
cfeltonyeah pack it up03:22
*** Niharika has quit IRC03:43
*** CarlFK has quit IRC03:47
*** CarlFK has joined #timvideos04:18
*** ChanServ sets mode: +v CarlFK04:19
*** Niharika has joined #timvideos05:08
*** diadara has joined #timvideos06:05
*** diadara is now known as darthlungi06:08
*** darthlungi is now known as diadara06:24
MaZderMindthose bumps look like ACKs: the sender pulls sda high, the receiver pulls it low. when the sender does not detect sda being pulled low on the 9th edge, the transmission was unsiccessful06:24
MaZderMindor could it be that rhe fpga is not pulling sda to gnd but instrad leavinimg it Z and rely on the pulldown? that would also explain those bumps06:27
MaZderMindi think my scope can trigger on spi issues06:27
MaZderMindi' try this evening (+10h from now)06:28
mithroMaZderMind: if you can reproduce the output, that would be *super* useful06:29
mithroMaZderMind: as cfelton / CarlFK's friend pointed out - those bumps should still be registered as a zero06:29
mithroMaZderMind: do you have a HDMI EDID breakout cable handy? It would be interesting to see the output on the other side of the buffer06:30
mithroMaZderMind: btw - Did I ever tell you that I have a design for your EDID hack for HDMI which has a significantly cheaper BOM, it just needs someone to finish it06:31
*** diadara is now known as darthlungi06:32
*** darthlungi is now known as diadara06:36
*** diadara has quit IRC06:44
*** diadara has joined #timvideos06:44
*** micolous_ has quit IRC06:45
*** micolous has joined #timvideos06:45
*** diadara has quit IRC06:46
*** diadara has joined #timvideos06:46
MaZderMindmithro: no but I can pull an HDMI Cable apart07:49
MaZderMindmithro: which EDID-Hack are you talking about?07:49
*** diadara has quit IRC09:55
*** Niharika|mobile has joined #timvideos09:56
*** Niharika has quit IRC09:57
cfeltonMaZderMind: it looks like the FPGA is pulling to gnd fine but the buffer (only visibility we had) didn't pull quite to gnd, should be benign10:42
*** Niharika has joined #timvideos11:40
*** Niharika has quit IRC11:45
*** Niharika has joined #timvideos12:18
mithroMaZderMind: https://github.com/datenwolf/vgaedidinjector12:59
tpbTitle: datenwolf/vgaedidinjector · GitHub (at github.com)12:59
mithroMaZderMind: my version https://github.com/mithro/timvideos-edid-injector13:00
tpbTitle: mithro/timvideos-edid-injector · GitHub (at github.com)13:00
*** Niharika|mobile has quit IRC13:38
*** 20WAA2QA9 has quit IRC14:02
*** scrollback has joined #timvideos14:05
*** scrollback has quit IRC14:12
*** 21WAB0NI2 has joined #timvideos14:15
*** 5EXAA3F4Z has joined #timvideos14:17
*** hyades has quit IRC16:19
*** hyades has joined #timvideos16:47
*** CarlFK has quit IRC17:34
*** CarlFK has joined #timvideos17:42
*** ChanServ sets mode: +v CarlFK17:42
*** Niharika has quit IRC19:29
*** hyades has quit IRC22:49

Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!