#milkymist IRC log for Friday, 2012-03-30

qi-botThe firmware (using branch) build was successful, checkout the VERSIONS for detail, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120330-0022/00:02
wpwrakdrum roll. fanfares ! :)00:08
whitequarkthe build wasn't FAILED!00:55
whitequarkor hasn't been00:55
whitequarkenglish. never understood it correctly.00:55
kristianpaullayout, *g*01:02
kristianpaulwpwrak: easy to pull but that requires a bit resolucion and color depth..01:15
kristianpaulalso i dont get it well, does he really control the media or there is some body behing pulling the plug and he just syncs...01:15
Action: kristianpaul just learn a equation to calcultate metastabilly probabillity to happen01:20
kristianpaulnow i feel okay when my M1 early board sudendly stop working just after a month or so ;)01:21
Action: kristianpaul try found manual for fpga_edline01:59
kristianpaullekernel: do you have documentation about the memory frames and commands that are part of bitstream?02:01
diginethow fast can MilkyMist go to on a the Virtex-4 as referenced on the site as the main supported system?05:18
diginetas in, how many megahertz05:18
cladamw(M1r4 placement) It's under processing now. Today I went to house to explain all m1r4's changes and requirements firstly. So a work of creating footprint got to start and later import to pcb file will be processed. One week later the rough 'placement' will be sent back to me.07:58
cladamw(M1r4 KiCad version) schematic symbols/libraries are under creations, see http://projects.qi-hardware.com/index.php/p/kicad-libs/08:01
rejonWho cool, a milkymist commercial! http://www.youtube.com/watch?v=mtTptScsqt809:35
rejonfrom spencer/bonewolf09:35
wpwrakrejon: it's from india. why is there no mass dance scene ??12:15
wpwrakcladamw: how did they layout guys react to the changes ? was there anything they were less enthusiastic about ?12:16
cladamwno less enthusiastic now. house guys like to earn money though. so far now they just need to spend time to create footprint then do a work of placement, then i feedback to them once I receive placement file. :-)12:20
cladamwstep1 -> create all footprint, step2 -> i check all footprints when i receive the placement, after step2, we need to check if we're happy on those placements.12:22
wpwrakdid they comment on the difficulty they expect to have with routing the new design ?12:22
wpwrakyes, that's the implementation process. that assumes they find a good way to implement it. so they're certain of that ?12:22
wpwrakor will they just go ahead, see how things go, and complain if there's a problem ?12:23
wpwrak(which may be a good approach, considering that they're quite fast anyway)12:23
cladamwthey now don't comment on the difficulty on the new changes, they can feel the 'difficulty' only when they import footprints inside pcb board size, then will see.12:24
kristianpaulbut never comment or at some point they...12:27
cladamwi tried to ask the same questions as you just asked me here. But the answers I got was they will feel if/how difficulty under fpga. They basically are conservatism.12:27
wpwrakokay, that's reasonable. let's see how things go then.12:29
cladamwthe communication between irc open communities or close om internal engineering status are quite differences.12:29
cladamwBUT, all i can do were to ask and try to convince. Else I can't do too much.12:29
cladamwjust give them sometimes to create footprints then generate dxf file to see.12:31
kristianpaulcladamw: but if something theroically may go wrong with new layout can you sue then? :)12:31
kristianpaulyup lets wait12:32
cladamwkristianpaul, i could tell you that your concerns are almost happened here(eastern) if things are close. Nevermind12:32
cladamwthe facts already have been happened/proved that even you didn't follow layout rules suggested from vendor's datasheet, your board still could be failed. things becomes chance problems/probabilities.12:34
cladamwbut wpwrak always could be correct that if we encounter all such layout rules in one day by ourselves then boards will get stronger as possible.12:35
cladamwnow the tasks for milkymist or qi's projects, we should move forward into new stage: Build layout rules for supporting Qi projects ans else.12:37
cladamwwpwrak, have you heard gta02 they have had rules for it ? :-)12:37
kristianpaulourselves, oh sure12:38
cladamwwpwrak, sorry that I am not kidding, i should say gta04.12:38
wpwrakcladamw: i think the three golden rules were. 1) don't look for bugs. 2) don't mention bugs. 3) if you hear someone talk about bugs, don't listen.12:39
wpwrakcladamw: (gta04) ah, don't know about that one. maybe :)12:39
wpwrakcladamw: (gta02) but i also have to say that the team got better. by the time you joined, it was already very different from the situation we had at the beginning12:40
cladamwwpwrak, hey...all rules came from vendor, also came from their own internal improvement 'bug' in success. :-)12:40
cladamwwpwrak, when i back to see m1's sch while I build KiCad libraries, there's a lot ways to improve now.12:43
cladamwfor examples, parts with SOT23-3 or 23-5, those pins's name and number had better to use/sync the same orientation from top side view on chip body.12:45
wpwrakyes :-)12:45
wpwrakbut be careful. sometimes the vendors change the orientation, too. very frustrating, when that happens12:46
wpwrakand i think i have sot23-like parts already. lemme see ...12:46
cladamwfrom time to time, i feel it as important since people can identify chip body directly and know quickly know which pin is which one.12:46
cladamwso today  i changed back few parts to be liked this way.12:47
wpwrakyeah, sot-323.fpd12:47
wpwrakthis one should be generalized12:47
wpwrakerr, footprint, of course12:47
wpwrak(just having my morning caffeine. brain not fully up yet, sorry :)12:48
wpwrak(identify things) yes, i think that's good12:48
cladamwyes, but I meant in those schematic symbols should be corresponding/visibly. :-)12:49
cladamw(caffeine) ;-O12:50
wpwrakagreed. there are only a small number of cases where a symbol that doesn't indicate the package makes sense. mainly things like resistors and such12:51
wpwrakeven diodes sometimes come in packages with 3 pins. (i like those - very easy to check the polarity ;-)12:51
cladamwwith that way, when we ask for an review on symbols or modules, they both easier to visually check when forth and back surf in datasheet pin assignments. :-)12:51
wpwrakby the way, a few projects on qi-hw have their local library extensions. that's for things that haven't been merged into the main library yet:12:53
cladamwwpwrak, yes, m1 have diode with 3 pins with one pin un-used, so review adds difficulty to read. :( I should have edited in this way.12:53
cladamwwpwrak, oah...yeah12:53
cladamweven in previous avt2.12:53
cladamwi copied avt2's libraries then edited /saved into another for m1. ;-)12:54
cladamwthat way is fast. :-)12:54
wpwrakben-wpan/components/, wernermisc/labsw/components/, f32xbase/c2usb/components/12:54
wpwrak(avt2) hehe :)12:54
wpwrakthere are even more footprints (all with fped) in the various projects12:55
cladamwavt has SOT23-3 and SOT23-5 even 23-6.12:55
cladamwyeah..so later we need to merge them. :-)12:55
wpwrakjonand: ah, gnucap says it also supports IBIS. last release from 2006, but the lists seem reasonably alive. http://www.gnu.org/software/gnucap/12:57
qi-botThe firmware build was successful, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120330-1446/14:27
jonandwpwrak: even older than eispice then14:30
jonandwpwrak: but that doesn't mean that it is useless14:30
GitHub140[milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/8b-TVA14:34
GitHub140[milkymist-ng/master] asmicon/refresher: fix refresh sequence done signal - Sebastien Bourdeauducq14:34
wpwrakjonand: sometimes, a project just stops making releases and everyone just runs things from the repository. that may be the case there. what's more worrying is if all activity (also mailing list of commits) stops. of course, that could mean that the project simply relocated, and maybe changed name to cover its tracks.14:43
wpwrakalways a bit of detective work :)14:43
GitHub133[migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/bb864c65dc3c4df4f8e3bc84db8c8b80de6b236d14:48
GitHub133[migen/master] sim: proxy - Sebastien Bourdeauducq14:48
GitHub127[milkymist-ng] sbourdeauducq pushed 2 new commits to master: http://git.io/8Ew4sA14:49
GitHub127[milkymist-ng/master] tb/asmicon: refresher test - Sebastien Bourdeauducq14:49
GitHub127[milkymist-ng/master] tb: remove obsolete norflash test bench - Sebastien Bourdeauducq14:49
qi-botThe firmware (using branch) build was successful, checkout the VERSIONS for detail, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120330-1627/16:07
AlarmI want the user manual of Flickernoise DMX?17:10
GitHub72[migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/94b02aa8ed355707f108ee813ea48c9e47b5e13320:24
GitHub72[migen/master] bus/asmicon: initiator - Sebastien Bourdeauducq20:24
--- Sat Mar 31 201200:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!