#milkymist IRC log for Tuesday, 2011-08-23

wpwrakwe've had boards with trouble on NOR DQ5 and DQ7. that would be W4 and Y7, relatively far away00:02
wpwraklekernel_: one test to prove the DQ8-PROGRAM_B contamination hypothesis would be to pull PROGRAM_B high through ~100 Ohm. that should visibly affect DQ8 if the two are linked00:04
wpwrakand the battle continues ! :)00:45
wpwrakaw: i described the funny things o PROGRAM_B to joerg and he thinks this looks like ESD damage (or similar). he also thinks regular xray wouldn't show us anything. you'd need xray combined with a very good microscope to really identify such an issue.00:47
wpwrakaw: so there's probably no use in taking the boards to be xrayed. that is, unless we find something else.00:48
awwpwrak, hmm....all good info to be considered. ;-) so a reset circuit would be likely as ESD damage or other circuits?00:49
wpwrakaw: now, yesterday, we reached the "comfortable" conclusion that some of the I/O pads of the FPGA are damaged/fused. "comfortable", because this can explain pretty much anything. but i think we can easily verify whether there's really a "connection" from DQ8 to PROGRAM_B00:50
awwpwrak, btw, I'm keep on reworks, maybe needs another 1.5 hours can be done this morning firstly. Once I start to test, I'll go for 0x3c.00:51
wpwrakaw: maybe the board was improperly handled at some point in time, apparently when some of the rework was done. as i understand it, both the fab and you had done part of the rework, so it may be difficult to track down the origin of the problem.00:51
awwpwrak, hmm..okay..a word of "comfortable" is good to explain that "connection".00:52
wpwrakaw: before 0x3c, please have another look at 0x77. there's one test that should tell us if our analysis so far is correct: monitor TP36/PROGRAM_B and NOR pin 34, as before. then reproduce the "pulses" on PROGRAM_B.00:52
awwpwrak, i see. but don't forget your idea on removing D16 to checking the thinking of feedback of system transfer.00:53
awwpwrak, alright, I'll ping you once start to test though. ;-)00:53
wpwrakaw:  and then, connect TP36 through 100 Ohm to 3V3. this should pull up PROGRAM_B. if DQ8 the signal that manipulates PROGRAM_B, then DQ8 should also be pulled up a bit00:53
awwpwrak, okay00:54
wpwrakaw: (d16) removing d16 may make the problem disappear without solving it. so if we go to the point of removing components, i'd first remove the reset IC, then see if the problem still happens (i expect that it will)00:55
wpwrakaw: only then remove d16 and see what happens. if the problem goes away, we may not have learned anything, because it may then simply happen too quickly to notice (and setting up the equipment to find the very first occurrence may be hard). only if it doesn't go away.00:56
wpwraktataa, my M1 is in brazil now (near sao paulo). getting close ;-)01:08
kristianpaulwpwrak: :-)01:53
wpwrakM1 has reached buenos aires ! now customs will sit on it for a day ...03:15
awwpwrak, sorry that I've not tested yet, just were interrupted from other tasks. the "available" of fix2 boards are all became "fix2b" ver. I'll start to test after lunch.04:26
awroh, hi, you sent me totally 21 pcs of case top cover (with m1 logo) in second shipment, right?05:59
wolfspraulaw: should be 20 sets (packaged separately in plastic bags), plus some spare parts including an additional top acrylic, yes06:02
awmm...spare though...thanks.06:03
wpwrakaw: ready for poking 0x77 ?06:07
awwpwrak, not yet06:10
wpwraktime to crawl to bed ...08:14
errordeveloperhello hello10:01
errordeveloperlekernel_: did you know of this thing http://orcc.sourceforge.net/10:02
errordeveloperlooks like yet another HDL-related use of LLVM10:04
larsc_errordeveloper: the llvm and hdl parts of that project seem to be orthogonal10:18
errordeveloperok ...10:27
errordeveloperlekernel_: hm, anyhow .. that was a random thing I bumpend into10:29
errordeveloperlekernel_: have you done any work with qemu ?10:30
errordeveloperI'm trying to find some docs on how to code virtual devices for it and possibly co-simulate say a simple device in HDL and it's driver10:31
errordeveloperlarsc_: I just noticed that it was your reply, not lekernel_'s .. :)11:11
errordeveloperlarsc_: how are you doing anyway ?11:11
lekernel_hmm, interesting12:05
lekernel_except it's eclipse-based12:06
lekernel_Orcc requires the Eclipse environment, which runs on the Java platform. I12:08
lekernel_why didn't they simply write a stand-alone command-line compiler? ...this whole focus on Eclipse looks like a warning sign to me12:09
lekernel_first, Eclipse (and Java) are bloated and poorly designed, second, they are irrelevant to writing an efficient code generator12:10
lekernel_but well... maybe it's simply poorly documented and there's good stuff in it12:13
Fallenouis Eclipse more bloated than ISE ?12:14
lekernel_ah, it's academic... that helps explain the eclipse/java crap12:14
lekernel_the bloat/functionality ratio of eclipse is much higher than that of ise12:15
lekernel_it seems serious, maybe the authors simply have read too much Sun propaganda12:21
Fallenouis it a C program that looks like a java program ? :p12:29
Fallenouyes it is12:29
Fallenounice :)12:29
Fallenoufunny the two imports does not make an error12:31
larsci recently read somebody arguing that linux kernel developers should switch away from comandline gcc and vi/emacs and use a proper tool like eclipse.12:42
larscand that was not academi12:44
lekernel_Eclipse as a text editor/IDE isn't too bad (except for the massive bloat ofc), but I really don't understand people building complete programming tools _depending_ on it12:47
azonenbergExcept for the massive bloat? Lol13:06
azonenbergAnd i dont usually do command line editing13:06
azonenbergi normally use an X11 based editor like geany (with IDE features disabled) or gedit13:06
azonenbergand then manually compile in command line gcc13:06
wpwrak /nick richelieu13:37
wpwrakazonenberg: so, an X11-based editor ? i wonder if that's enough to make the heresy charge stick ...13:38
azonenbergI do use nano on a routine basis13:38
azonenbergbut i like my mouse tyvm13:38
azonenbergfor editing13:38
larsci prefer vim, because i don't have to switch between mouse and keyboard13:52
wpwrakmouse, n.: computer peripheral to set the input focus to the xterm of the user's choice13:56
errordeveloperlekernel_: I'm not an eclipse supporter either ...13:58
larscwpwrak: i though that's what 'ctrl+a n' is for13:59
wpwraklarsc: ;-)14:01
azonenbergwpwrak: The big thing i like about X based editors is the ability to respond to a mouse scroll wheel14:03
azonenbergas well as higher resolution than any of my framebuffer modes support14:03
wpwrakazonenberg: oh, i like the ability to have multiple xterms in view at the same time. i have a triple-head setup. usually have ~6-12 xterms visible at any given moment. each head (screen) has a 4x4 array of virtual screens. e.g., xchat shares its space with a few xterms with mutt reading my inbox, plus a few xterms for very short-liven activities (kinda post-its). the next virtual screen has 6 large xterms with mutts reading various threa14:07
wpwrakds of the qi-hw lists. another has a web browser. etc. that's the left screen. the middle screen has the editing/development i do. e.g, right now, six xterms, one kicad, and fped overlaid (it's a bit crowded).14:07
Action: kristianpaul likes a lot his byobu + vim14:10
kristianpaulbut yeah with a larger screen ratpoison will fit very well :-)14:13
aw_rc3: 5 boards - 0x30, 0x33, 0x35, 0x37, 0x38 with fix2b are happy from previous "available" with fix2 version.14:20
wpwrakaw_: any boards that regressed after fix2b ?14:21
aw_since the notes may hard to realize/understand easily, you can just see the http://en.qi-hardware.com/wiki/Milkymist_One_run_3_schedule#Test_Results14:21
aw_I collected those PASSs and FAILs with two catagory , so you can see it.14:22
aw_wpwrak, so far no regressed happened yet. ;-)14:22
aw_I'll follow MAC address number to test, so you can know it.14:23
wpwrakah, nice summary14:24
aw_wpwrak, tomorrow morning let's go for a little experiments on 0x77 or 0x3c firstly but not whole day even not half day.14:24
aw_and sorry that other tasks to be done so that didn't go through that failed board.14:25
wpwrakalright, i'll be around. your morning is good. i can't stay up all night for it anyway. not with fedex coming the next day :)14:25
aw_and i gotta go sleep.14:26
aw_wpwrak, yup14:26
aw_wpwrak, you should get fedex one probably by tomorrow. Good luck.14:27
errordeveloperhm ... has anyone watched this: http://free-electrons.com/pub/video/2011/elc/elc-2011-williams-fpga-soc-co-simulation-x450p.webm15:17
errordevelopertitle is supposed to be self-explanatory :)15:19
errordeveloperlekernel_: that's follow-up to my earlier qemu question15:21
errordeveloperwell, I think I found the answer ;)15:21
wpwrakM1 cleared customs ! tomorrow it'll be mine :)21:21
kristianpaulwpwrak: :_D22:26
kristianpaulwpwrak: You'll have more colored/animated parties after M1 arrival ;-)22:28
wpwraki'm counting on it ;-) i'll probably have to get a projector and paint the terrace walls a nice white. the old paint is dirty and fissured anyway.22:29
--- Wed Aug 24 201100:00

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