kristianpaul | oh wow http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&No=662&from=enbanner | 00:28 |
---|---|---|
kristianpaul | "The significance lies in the fact that the entire system is completely open-source, compounding the influence immensely." | 00:28 |
Action: pabs3 thought all FPGA were proprietary | 00:32 | |
kristianpaul | of course they are but remenber, is "open source" | 00:33 |
wpwrak | kristianpaul: bah. i thought this bombastic comment was about M1 | 00:57 |
kristianpaul | wpwrak: i dont get it well then.. | 00:57 |
kristianpaul | wpwrak: ah | 00:58 |
kristianpaul | wpwrak: well | 00:58 |
kristianpaul | sorry ;) | 00:58 |
kristianpaul | now i lost, you meant the link i posted right? | 00:58 |
kristianpaul | but yes sorry bring all those not very copyleft links latelly, i was reading some news.. then.. well ;) | 00:59 |
wpwrak | ;-) | 01:07 |
kristianpaul | so.. | 01:27 |
kristianpaul | hmm need push a branch i created in a read-only git repo of my own.. | 01:30 |
wpwrak | maybe you can just edit .git/config | 01:32 |
wpwrak | else, there's always git diff ... >~/file; cd ../good-branch; patch -p1 -s <~/file :-) | 01:33 |
kristianpaul | hehe you just read my mind | 01:34 |
kristianpaul | but found it i need add a remote branch | 01:34 |
kristianpaul | and i wanted to avoid that patching foo | 01:35 |
GitHub13 | [milkymist] kristianpaul pushed 1 new commit to gnss-sdr: https://github.com/kristianpaul/milkymist/commit/13ca0cdd2c7a383e71ab2d72f006de3cdf0ba853 | 01:40 |
GitHub13 | [milkymist/gnss-sdr] namuru gnsssdr initial import - Cristian Paul Peñaranda Rojas | 01:40 |
roh | milled pcb today (mechanics, outline and a hole) | 01:46 |
roh | http://yamato.hyte.de/tmp/pcb_dummy_test_fr4_front.jpg http://yamato.hyte.de/tmp/pcb_dummy_test_fr4_back.jpg | 01:47 |
roh | just taped down, milled in one round, 1mm diameter endmill | 01:47 |
kristianpaul | roh: gta04 super cool acrylic case? :-) | 01:48 |
roh | kristianpaul: nah. a pcb mechanical prototype for an actual customer | 01:48 |
wolfspraul | cool, same idea as I had recently? | 01:48 |
roh | wolfspraul: well.. doublesided tape is not really news depending on what | 01:49 |
wolfspraul | sure | 01:49 |
kristianpaul | you're using tape as a case shell? | 01:49 |
wolfspraul | btw, speaking about doublesided tape. in Germany I always thought this was a relatively exotic 'tool' | 01:49 |
wolfspraul | but since I live in China, #1 I find double-sided tape in pretty much every small store, corner store, grocery store - EVERYWHERE | 01:50 |
wolfspraul | and... | 01:50 |
wolfspraul | I know that doublesided tape can be used to: | 01:50 |
wolfspraul | SOLVE EVERY PROBLEM! | 01:50 |
wolfspraul | :-) | 01:50 |
kristianpaul | "It is designed to stick two surfaces together, often in a way which is _not visible_ in the end product" | 01:51 |
wolfspraul | you name it, I solve it | 01:51 |
wolfspraul | doublesided tape can solve *every* problem | 01:51 |
kristianpaul | hmm | 01:51 |
wolfspraul | it's magic | 01:51 |
wpwrak | worst case, it can be used to shut up anyone complaining about the solution ;-) | 01:52 |
wolfspraul | I have one (actually several) rolls in my household | 01:52 |
wolfspraul | on my desk | 01:52 |
wolfspraul | everywhere | 01:52 |
kristianpaul | is not too much? | 01:52 |
wolfspraul | it's magic | 01:52 |
wolfspraul | whenever I can, I buy another roll 'just in case' | 01:52 |
Action: kristianpaul imagines wolfspraul doing a doublesided ritual | 01:52 | |
wpwrak | roh: did you put some victim material between the PCB and the mill's table ? or are things precise enough that the tape's thickness was enough ? | 01:52 |
kristianpaul | how durable is it? i mean it solves the problem for let said a week untill ;) | 01:53 |
kristianpaul | hehe | 01:53 |
wpwrak | wolfspraul: are you sure you're not confusing it with toilet paper ? | 01:53 |
wpwrak | of course, double-sided toilet paper ... eww :) | 01:53 |
roh | wpwrak: wood | 01:54 |
kristianpaul | if replaces all glue uses well.. a good start | 01:54 |
wpwrak | wood makes a great victim :) | 01:54 |
wolfspraul | kristianpaul: ok, back to being serious | 01:54 |
kristianpaul | sure :) | 01:54 |
wolfspraul | I think it should be reserved for special cases | 01:54 |
roh | milled about 0.1 to 0.2 into the wood | 01:54 |
wolfspraul | because as with every glue, it's nasty in the long run | 01:54 |
wpwrak | roh: my mill doesn't have any good mounting system. you basically fix the entire table. so i made disposable tables of FR2, tape the things to them, then let the mill overshoot a bit | 01:56 |
wpwrak | not great (bending), but on a good day, i get < 0.1 mm Z tolerance | 01:57 |
wpwrak | (FR2 tables) of course, i can reuse them many times before they get too badly scarred | 01:57 |
wpwrak | wolfspraul: you can wind it around things. since it's double-sided, you don't even need to think about which side goes where. it'll just stick to whatever :) | 01:59 |
cladamw | wpwrak, one git question I don't know --> http://dpaste.com/718835/ sorry that interrupt your chats. | 02:00 |
wpwrak | try to git pull | 02:01 |
wpwrak | hopefully, this shouldn't cause any merge conflict | 02:01 |
wpwrak | then you can git push | 02:01 |
cladamw | still to 'git pull' since I've changed a file ? | 02:01 |
wpwrak | you get this because i've committed changes since the last time you pulled, so git needs to synchronize again | 02:01 |
cladamw | oh..okay | 02:01 |
wpwrak | the git pull won't override your changes | 02:02 |
cladamw | aha...good. | 02:02 |
wpwrak | if there's a conflict, it will tell you. but i don't think there will be any issues | 02:02 |
wpwrak | we should move those "official" things out of wernermisc. that's meant to be my personal playground. and its orderliness reflects that ;-) | 02:03 |
cladamw | http://dpaste.com/718836/ after 'git pull' :-) | 02:03 |
wpwrak | hmm, i think git pull origin master should do the trick | 02:05 |
wpwrak | i hope you're not in "detached head" mode ... | 02:06 |
cladamw | type 'git pull original master' ? | 02:06 |
kristianpaul | what was tha famous way to reload fpga bitstream from flash using urjtag? | 02:06 |
wpwrak | yeah | 02:06 |
wpwrak | kristianpaul: make -C wernermisc/m1/jtag-boot boot | 02:07 |
qi-bot | [commit] Adam Wang: changed J24 to p/n: TSW-105-26-H-D-009, see: (master) http://qi-hw.com/p/wernermisc/458b90d | 02:07 |
qi-bot | [commit] Adam Wang: C145 -> 1uF/0402, C234 -> 10uF/0805, R62 -> 0603 (master) http://qi-hw.com/p/wernermisc/d5f6e03 | 02:07 |
qi-bot | [commit] Adam Wang: Merge branch 'master' of projects.qi-hardware.com:wernermisc (master) http://qi-hw.com/p/wernermisc/e228723 | 02:07 |
cladamw | wpwrak, nice , thanks. :-) | 02:08 |
wpwrak | congratulations ! :) these are among the scarier moments of life with git | 02:11 |
cladamw | oah...poor adam , learn git while needs, but you're involved life with git. :-) | 02:13 |
kristianpaul | wise words | 02:13 |
wpwrak | it has its dark sides :) but the good thing is that pretty much all the questions you can possibly have have already been answered. and google is good at finding them (and the answers) | 02:14 |
cladamw | wpwrak, ha...yes, i should have googled first instead of direct asked you. Sorry. :) | 02:18 |
wpwrak | naw, no problem :) i don't mind showing off my recently acquired knowledge and/or google skills ;-) | 02:19 |
kristianpaul | hmm i need update some threshold not getiing a lock.. | 02:32 |
Action: kristianpaul manually counts capacitors from CvPCB window.. | 03:37 | |
Action: kristianpaul quits and open a console | 03:39 | |
kristianpaul | wpwrak: does boom/bom provides summary for components? ie i need x capacitors y resistor n diodes.. | 03:41 |
kristianpaul | i dunno if is general usefull, but i'm about to make a qoute to a electronic shop here for passive smd stuff | 03:42 |
wolfspraul | you will have a very hard time competing with digikey I think | 03:45 |
wolfspraul | but definitely try and share your experiences | 03:45 |
kristianpaul | well digikey order will cost me 100usd shipping! :( | 03:45 |
kristianpaul | or was 50usd.. | 03:45 |
wolfspraul | how about a forwarding service? | 03:46 |
kristianpaul | anyway i just need passive components i have the other ICs already here | 03:46 |
wolfspraul | yes but that's the problem, if things are too cheap they become expensive again :-) | 03:46 |
kristianpaul | wolfspraul: perhaps.. | 03:46 |
kristianpaul | dont remenber me that.. : | 03:46 |
wolfspraul | or the other way round - if you can deal with your cheapest/smallest problems most efficiently, the rest is a piece of cake | 03:48 |
kristianpaul | i hope.. | 03:48 |
kristianpaul | lets see, one day i will order from digikey, perhaps when making my own maxim + spartan6 board ;) not now.. | 03:49 |
wolfspraul | I think there may be good forwarding services in florida you may want to check out | 03:49 |
kristianpaul | Sure i know then ! | 03:49 |
kristianpaul | I got my avnet board that way | 03:49 |
kristianpaul | ok i'll qoute on digikey too :-) | 03:50 |
kristianpaul | -_- http://www.digikey.com.co/Web%20Export/hp/es/free_shipping_banner_co.gif | 04:00 |
Action: pabs3 notes there isn't any NN teardowns on iFixit, interesting comments on http://hardware.slashdot.org/story/12/03/20/0115228/ifixits-kyle-wiens-on-the-war-on-diy-electronics | 04:01 | |
kristianpaul | argh, i dont have SAW filer.. | 04:07 |
kristianpaul | and digikey dont said too much about TA1575IG .. | 04:07 |
kristianpaul | ergh, i have it was confusing it with xtal phew ;) | 04:10 |
kristianpaul | wpwrak: any suguestions for buying resistor and capacitors at digikey? | 04:37 |
kristianpaul | i'm getting 300 results just by writing 20K SMD 0603 | 04:38 |
kristianpaul | okay size and dimension.. i need take care about that | 04:39 |
kristianpaul | wow from 0.02 to 0.6... i need find diffences here too | 04:44 |
kristianpaul | MQO is 1 btw | 04:45 |
Action: kristianpaul choosed P20KGCT-ND for now | 04:46 | |
kristianpaul | 445-4839-1-ND 1.5usd :-| | 05:12 |
DocScrutinizer | all about packaging | 06:51 |
DocScrutinizer | I'd not worry too much about it, as long as you get enough for little money. Does it matter if you pay USD1.5 for 100 or USD2.5 for 1000, when you need 35? | 06:54 |
qi-bot | [commit] Xiangfu: m1/patches/rtems/: Milkymist driver: add mic-boost support (master) http://qi-hw.com/p/wernermisc/901a483 | 09:21 |
kyak | xiangfu: the workaround on missing libncursesw probably breaks the UTF-8 support | 09:59 |
kyak | why is libncursesw missing? because there are some bugs in Kconfig | 09:59 |
xiangfu | kyak, yes. we better fix that :) it's mark as module. and if I install the libncursesw from package repo, it will give other error | 10:00 |
xiangfu | we needs fix all those in next release. so next release will be in 1 or 2 month :-) | 10:01 |
kyak | i still wasn't able to dig to the root of this issue with libncursesw | 10:02 |
kyak | even with the help of jow | 10:02 |
kyak | packages that depend on it get deselected or libncursesw itself suddenly dissapears from the menu.. | 10:03 |
kyak | the number of qi packages failed to build is & too big | 10:03 |
whitequark | DocScrutinizer: some Android stuff was integrated into Linux 3.3 | 10:04 |
whitequark | through not wakelocks yet, these (or, as I understand it, some more sane variant of them) are scheduled for 3.4 | 10:04 |
kyak | hm wait a minute | 10:04 |
kyak | some packages are there in packages, but still listed as not built | 10:05 |
kyak | i think it's some problem with parsing the build output log | 10:05 |
kyak | see fbterm | 10:05 |
kyak | qt4 is not there.. | 10:06 |
kyak | i think i mentioned the bug and workaround to mirko... | 10:06 |
xiangfu | kyak, fbterm. yes. | 10:07 |
kyak | dunno if it's a good idea to release without qt | 10:07 |
kyak | all qt packages are therefore failed | 10:08 |
xiangfu | I want make progress on release. so I just release one image without all package that failed. | 10:08 |
xiangfu | then we start working on this one. fix bugs. | 10:08 |
xiangfu | :-) | 10:08 |
wolfspraul | kyak: thanks for noticing those things! | 10:18 |
wolfspraul | yep we gotta go back and bring things back up, it's tough | 10:19 |
wolfspraul | we are still lacking a strong test plan and regression testing, and regular, at least -testing images, say monthly or bimonthly? | 10:19 |
xiangfu | kyak, fbterm didn't find it under package. | 10:20 |
wolfspraul | so the only way for xiangfu to break out of it after 4 months was to drastically cut down on packages or features that just don't build for this or that reason, so at least we have a basis to work on | 10:20 |
wolfspraul | and in fact, as far as I'm concerned, I would go all the way back to boot to console if that is what is needed to come out with most recent ben images :-)( | 10:20 |
wolfspraul | :-) | 10:20 |
wolfspraul | if someone doesn't like that, they can either take the old images, or help fix the latest | 10:21 |
wolfspraul | xiangfu: we should leave the old package binaries around as well | 10:21 |
wolfspraul | say someone installs the 2011-11-xx image, what package repo does it point to? | 10:21 |
wolfspraul | is that repo still there? | 10:21 |
xiangfu | kyak, I will test the libncursesw in next daily build. | 10:21 |
wolfspraul | we should keep the entire old images + package repo alive, especially if we cut things in more recent builds | 10:21 |
xiangfu | wolfspraul, package repo. every release have one. and back well. | 10:22 |
wolfspraul | so the old package repo is still there? | 10:22 |
wolfspraul | and will stay there? | 10:22 |
xiangfu | here: http://downloads.qi-hardware.com/software/packages/NanoNote/Ben/ | 10:22 |
xiangfu | yes. people can easy switch to the one he like by edit /etc/opkg.conf | 10:22 |
wolfspraul | ok perfect | 10:22 |
wolfspraul | so then nothing is lost | 10:22 |
wolfspraul | if someone notices a missing or broken feature/package in latest, they can either use an older image and repo, or help fix latest | 10:23 |
kyak | xiangfu: http://downloads.qi-hardware.com/software/packages/NanoNote/Ben/2012-03-18/ - fbterm package is there | 10:24 |
kyak | oh! | 10:24 |
xiangfu | hmm... | 10:24 |
kyak | that's jfbterm :) | 10:24 |
kyak | browser search hilightning covered the "j" :) | 10:25 |
wolfspraul | kyak: so bottom line. I apologize if our build process looks rough and we seemingly trample over things done in the past. none of these things will be forgotten, it's just a matter of how we resurrect them into the latest images and sw versions... | 10:25 |
wolfspraul | and it's great to see you back and even alerting us of things that broke - phew :-) let's get them back! | 10:25 |
kyak | wolfspraul: yeah, probably it's a good thing to release anyway, perhaps it will motivate people to fix things.. THough i don't lack for motivvation, i lack for time | 10:26 |
xiangfu | kyak, yes there is something wrong on 'menuconfig' | 10:33 |
kyak | xiangfu: http://en.qi-hardware.com/irclogs/qi-hardware_2012-01-22.log.html#t10:07 | 10:43 |
kyak | however, the patch is gone :) | 10:43 |
kyak | oh no, i reported it: https://dev.openwrt.org/ticket/10834 | 10:43 |
kyak | i don't feel it's a good idea to override qt4 though | 10:44 |
kyak | xiangfu: (menuconfig) - you'll find plenty of this in logs.. | 10:44 |
kyak | i know you can't be angry on inanimate objects, but i got really angry with Kconfig | 10:45 |
whitequark | yeesh, kconfig is nontrivial | 11:36 |
whitequark | it becomes easy once you damage your brain in exactly the same way as kconfig dev did | 11:36 |
C-Keen | TRISTATE | 11:37 |
DocScrutinizer51 | whitequark: (android stuff in linux) not something I start to cheer about | 12:00 |
Action: wpwrak vaguely remembers the first time he ran into wakelocks. they've been trying to get them into mainline for a mighty long time :) | 12:29 | |
whitequark | seriously, never buy anything from tuxbrain. he forgot to put one atben into package and does not respond to email anymore. | 12:40 |
wpwrak | he really seems to be coming apart :-( | 12:50 |
wolfspraul | whitequark: is that all that was missing? | 12:50 |
wolfspraul | maybe I will just send you a 'free' (you already paid) atben from my side, we have to hold together... | 12:50 |
wolfspraul | david from tuxbrain is a good guy, but retail/shop business is just crazy tough. maybe he is too good a guy to make it there. | 12:51 |
whitequark | wolfspraul: yes, otherwise it was fine (except for 2-month delay, but I can at least understand that) | 13:14 |
whitequark | I'd be very grateful if you did | 13:15 |
wolfspraul | you already paid | 13:16 |
wolfspraul | I'll email you | 13:16 |
whitequark | wolfspraul: ok, thank you | 13:43 |
DocScrutinizer | hi $channel | 21:25 |
DocScrutinizer | wpwrak: nice doc about tuned line termination :-) also nice (for me) that I was right about optimum design is termination (50R) on *both* ends | 21:27 |
DocScrutinizer | also the mystery about impedance of *input* still the same, they just mentioned 15R driver output impedance | 21:28 |
DocScrutinizer | maybe the 3pF parasitc load of input that they postulated in their virtual schematics IIRC, that's the effective load impedance at input? | 21:30 |
DocScrutinizer | makes little to no sense to me | 21:30 |
wpwrak | i've seen input capacitances in that range mentioned a few times | 21:36 |
wpwrak | the FETs and clamps (if present) should have little else on terms of impedance, no ? | 21:36 |
DocScrutinizer | wpwrak: anyway see figure 2 in http://download.micron.com/pdf/technotes/TN4606.pdf p2, that's what I'd expect. for a real circuit with a chip input pin they seem to imply this input also has 18R impedance (fig 5) | 21:38 |
wpwrak | ah, you mean including the termination | 21:40 |
DocScrutinizer | nope | 21:40 |
DocScrutinizer | in fig5 the RX 18R is labeled "Receiver" | 21:41 |
DocScrutinizer | I read that as "inside chip" | 21:41 |
wpwrak | there are two "receivers" | 21:41 |
DocScrutinizer | fig7 is completely different in that respect | 21:41 |
wpwrak | i think "receiver" is actually source. and "receiver (2)" is the actual receiver. | 21:42 |
wpwrak | fig. 5 is bidirectional. and confusing ;-) | 21:42 |
wpwrak | yes, fig 7 is cool. just terminate in the middle ;-) | 21:43 |
DocScrutinizer | wpwrak: whatever fig5 is meant to be, the schematics are clearly more like fig2 rather than fig7, regarding terminating impedance on both ends | 21:44 |
DocScrutinizer | whatever that 18R in fig2 right end means, it's there | 21:45 |
DocScrutinizer | their simulation net in fig7 is completely messed up regarding that | 21:45 |
DocScrutinizer | err s/fig2/fig5 | 21:46 |
wpwrak | they admit that: "From the above discussion this is not an optimal configuration." | 21:47 |
wpwrak | (page 4. right column) | 21:47 |
wpwrak | they may be concerned that people trying to achieve a perfect but complicated termination would do more damage than good :) | 21:48 |
DocScrutinizer | they picked that compromise solution for sake of BOM | 21:58 |
DocScrutinizer | as they state somewhere | 21:58 |
DocScrutinizer | this doesn't put any sense into fig7 for me | 21:58 |
DocScrutinizer | as this sim net assumes termination impedance of infinite + 5pF | 21:59 |
DocScrutinizer | or zero, which wouldn't make any more sense | 22:00 |
DocScrutinizer | actually the sim net is garbage, they neither placed any proper source symbol in there, nor a proper sink that in any way siggests what impedance (4) shall have | 22:01 |
DocScrutinizer | *usually* you assume impedance of zero for sources and infinite for sinks, in such cases | 22:01 |
DocScrutinizer | zero for surce is in line with 15R at left side, which models the actual driver impedance | 22:02 |
DocScrutinizer | I don't see anything at right side modelling the sink impedance (unless we assume the chip actually has near to infinite resistive component, and 5pF capacitive component) | 22:03 |
DocScrutinizer | actually this might match the real chip's input characteristics | 22:06 |
DocScrutinizer | for me the 7nH and 5pF on right side fig7 are similar to left side, and indeed part of the chip pin sim | 22:07 |
DocScrutinizer | looking into fig9* you see the effect of this mismatching: overshots | 22:09 |
DocScrutinizer | wpwrak: also note please that this is an optimization for DQ line, aka bidir | 22:12 |
wpwrak | yes. for unidirectional, they use more traditional termination | 22:13 |
DocScrutinizer | :nod: | 22:13 |
wpwrak | which makes me wonder ... do we ? checking .. | 22:13 |
DocScrutinizer | at least for unidir you should stick with fig2 | 22:13 |
DocScrutinizer | which directly leads to a series R for source, and a parallel R to GND for sink | 22:14 |
DocScrutinizer | also note that even a transmission line with massive mismatch at one end still can work reasonably well when other end is optimally matched | 22:16 |
wpwrak | micron actually recommend to do the same for bidir and unidir | 22:16 |
DocScrutinizer | there will be reflections (aka overshots) but they get killed after one travel along the line | 22:16 |
wpwrak | yeah | 22:16 |
wpwrak | transmission lines are fun :) | 22:17 |
DocScrutinizer | and actually I think that's what they do with their center termination | 22:17 |
wpwrak | yup. sounds like it's that | 22:17 |
DocScrutinizer | anyway, doesn't that give you an invincible desire to do your own spice tests? | 22:20 |
wpwrak | so totally not ;-)) | 22:21 |
DocScrutinizer | hehe | 22:21 |
DocScrutinizer | they offer ready made spice data ;-) | 22:22 |
DocScrutinizer | >> Micron provides both IBIS and Spice models to help in this area (see ww.micron.com/models). | 22:23 |
wpwrak | feel free ;-) | 22:24 |
DocScrutinizer | I'm actually tempted, alas it's again almost midnight here | 22:24 |
DocScrutinizer | I start to hate this making_money | 22:24 |
wpwrak | having a dayjob kinda sucks ;-) | 22:25 |
wpwrak | and it's not the money. it's the day :) | 22:25 |
DocScrutinizer | there's a design glitch in this big simulation | 22:25 |
DocScrutinizer | maybe the parameter "MONEY" been a totally useless change to the original setup | 22:26 |
wpwrak | mmh ? what original setup ? | 22:27 |
Action: DocScrutinizer wanders off, searching for the RED PILL | 22:27 | |
wpwrak | ah, the hour of the tranquilizer :) | 22:27 |
DocScrutinizer | nah, RED PILL isn't tranquilizer, rather a POP on stack | 22:28 |
wpwrak | return to sleep mode :) | 22:29 |
DocScrutinizer | but we all learnt from Matrix, that it's a SIM in a SIM in a SIM (btw a concept that's as old as "Welt am Draht" at least) | 22:30 |
wpwrak | a film that would have greatly benefited from some cutting :) | 22:31 |
DocScrutinizer | so maybe not going to take RED PILL will be the more comfortable alternative / the better SIM, despite this MONEY thing | 22:31 |
wpwrak | heh :) | 22:32 |
DocScrutinizer | tw as cool as those delay elements in FPGA may be, they need nasty config - I strongly suggest to try and keep traces same length | 22:34 |
DocScrutinizer | .s/tw/btw | 22:34 |
Action: whitequark has read 10 lines of backlog | 22:35 | |
DocScrutinizer | whitequark: then yu missed all the good stuff ;-) | 22:36 |
wpwrak | ah, you've dropped off the cc again | 22:36 |
whitequark | wpwrak: cc? | 22:36 |
wpwrak | i forwarded you the missing mails | 22:36 |
wpwrak | no, DocScrutinizer | 22:36 |
DocScrutinizer | umm? | 22:36 |
DocScrutinizer | aah | 22:36 |
DocScrutinizer | 22:37 | |
wpwrak | whitequark: he prefers if we copy him explicitly on our mails instead of simply subscribing to the list. so every once in a while, the cc: is lost and the thread goes on without him | 22:37 |
wpwrak | whitequark: then he gets confused when we discuss in a context that has already changed. until someone detects the problem and forwards him what seem to be the missing mails | 22:37 |
whitequark | hmm, context | 22:38 |
wpwrak | they're of course unthreaded then, possibly increasing the confusion :) | 22:38 |
whitequark | DocScrutinizer: looks like your SP gets corrupted sometimes | 22:39 |
whitequark | and you suddenly find yourself in a middle of a wrong stack frame | 22:39 |
whitequark | err... recompile yourself with -fstack-smashing-protector ?.. | 22:39 |
DocScrutinizer | no increased confusion here :-D | 22:41 |
DocScrutinizer | yep, a slightly caring layout should result in a routing that "just works" | 22:42 |
DocScrutinizer | except for those "differential lines"... o.O - we got any of those? | 22:43 |
wpwrak | yes, the clock is differential | 22:43 |
DocScrutinizer | :nod: | 22:43 |
DocScrutinizer | so you may consider 2-laxered design for those traces | 22:44 |
DocScrutinizer | 2-layered* | 22:44 |
wpwrak | hmm. just parallel should do the trick. otherwise you need to consider the vias as well | 22:48 |
DocScrutinizer | you can easily make via length identical | 22:49 |
wpwrak | but they're on the same side. so one would go down and then up again | 22:50 |
wpwrak | unless you have a pcb with buried components. that would be fun ;-) | 22:50 |
DocScrutinizer | 2/6 via + trace + 4/6 via, vs n/6 via + trace + (6-n/6) via | 22:51 |
DocScrutinizer | down and then up, what's wrong about it? | 22:51 |
wpwrak | but why ? | 22:52 |
wpwrak | just seems to make things more complicated | 22:52 |
DocScrutinizer | to do sth about your concerns regarding via length | 22:52 |
wpwrak | you mean trace | 22:52 |
DocScrutinizer | nope, via | 22:53 |
wpwrak | now you lost me :) | 22:53 |
DocScrutinizer | >>>... otherwise you need to consider the vias as well | 22:53 |
wpwrak | simplest design: everything on the surface | 22:53 |
wpwrak | next to simplest: both dive out of the FPGA, then come back up together somewhere | 22:54 |
DocScrutinizer | as long as your traces are straight line, yes | 22:54 |
DocScrutinizer | have some bend in them and you run into length issues | 22:54 |
wpwrak | as long as they are exact copies of each other | 22:54 |
DocScrutinizer | and 0.5mm i actually a bit demanding | 22:55 |
wpwrak | yes. different radii would be trouble. | 22:55 |
wpwrak | they're probably exaggerating ;-) | 22:56 |
wpwrak | just consider the propagation speed of the signal | 22:56 |
wpwrak | units 0.5mm/c ps -> 1.667 | 22:57 |
DocScrutinizer | c on PCB is not what you might think it is | 22:58 |
wpwrak | even if we assume 66% of c, that's merely 2.5 ps | 22:58 |
DocScrutinizer | I frown on that unit, mil | 22:58 |
DocScrutinizer | maybe they meant millimeter? | 22:59 |
wpwrak | i doubt it :) | 22:59 |
DocScrutinizer | hah, http://openbookproject.net/electricCircuits/AC/AC_14.html | 23:12 |
DocScrutinizer | nice | 23:12 |
DocScrutinizer | TL for dummies | 23:12 |
DocScrutinizer | ;-D | 23:12 |
DocScrutinizer | >>> Velocity factor is a fractional value relating a transmission line's propagation speed to the speed of light in a vacuum. Values range between 0.66 and 0.80 for typical two-wire lines and coaxial cables | 23:13 |
wpwrak | that sounds like "chess in hyperbolic n-spaces for dummies" :) | 23:14 |
wpwrak | see. 2/3 was already a pessimistic estimate. | 23:14 |
DocScrutinizer | no really, it's nice | 23:14 |
wpwrak | i think it should be a bit faster on the pcb (if i interpret wikipedia right :) | 23:15 |
DocScrutinizer | even with extremely slower velocity we get not even in pall park where 0.5mm matters | 23:15 |
wpwrak | yes, seems to me that way,too | 23:18 |
DocScrutinizer | DANG | 23:19 |
DocScrutinizer | 0.1ns results in ~0.025mm if my math doesn't suck | 23:20 |
wpwrak | i think it does :) | 23:21 |
wpwrak | units 0.1ns*0.66c mm -> 19.786302 | 23:21 |
DocScrutinizer | yeah, sth is wrong with my math | 23:23 |
DocScrutinizer | my feeling tells | 23:23 |
DocScrutinizer | 1ns = 1GHz | 23:24 |
DocScrutinizer | GHz wavelength is not sub-mm | 23:24 |
wpwrak | only for very slow light :) | 23:24 |
wpwrak | tardions, not tachyons | 23:24 |
DocScrutinizer | meh, time for zzZZzz | 23:25 |
wpwrak | sweet suspend ! :) | 23:25 |
DocScrutinizer | :-) | 23:25 |
DocScrutinizer | one last thing, I think I heard they built some sort of TL for light, that was actually in range of mm/s | 23:27 |
DocScrutinizer | also c inside sun is astonishingly low | 23:30 |
DocScrutinizer | needs several 1000 years to reach surface from center | 23:30 |
wpwrak | must be a huge traffic jam for those poor photons | 23:30 |
DocScrutinizer | indeed, it is | 23:30 |
DocScrutinizer | luckily they can crowd up really dense | 23:31 |
DocScrutinizer | ;-D | 23:31 |
DocScrutinizer | afaik they won't ever collide with their own kind | 23:32 |
wpwrak | can't have the same kind of particle at the same place at the same time | 23:34 |
wpwrak | that's why you never read of, say, rolls royces colliding | 23:34 |
DocScrutinizer | I think they never are in one place at one time, since Heissenberg | 23:36 |
DocScrutinizer | but they must have *some* mass. I guess e/c^2 | 23:36 |
DocScrutinizer | so too many of them crowded up in one spot may result in a black hole | 23:37 |
DocScrutinizer | is there any theoretical upper limit for frequency of electromagnetic waves? | 23:41 |
DocScrutinizer | IOW what's that radiation that starts where gamma spectrum ends? | 23:44 |
wpwrak | one period per Planck length ? | 23:45 |
wpwrak | and only relativistic mass. m0 = 0. or they couldn't travel at c. | 23:45 |
DocScrutinizer | sure, as I said, e/c^2 | 23:47 |
DocScrutinizer | planck length, hmmm | 23:49 |
DocScrutinizer | Planck-Kreisfrequenz 1,85487 · 10^43 s1 | 23:55 |
wpwrak | future generations will have a heck of a time getting their layout constraints right :) | 23:57 |
--- Wed Mar 21 2012 | 00:00 |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!