| 00:08.44 | *** join/#edev ico2_ (n=ico2@modem-1293.chimpanzee.dialup.pol.co.uk) |
| 00:43.41 | *** join/#edev prpplague (n=dave@70.249.150.251) |
| 00:44.41 | prpplague | landley: how detailed do you want the specs? |
| 00:44.59 | prpplague | landley: i assume you just want the basics from the software side |
| 00:45.10 | landley | prpplague: pretty much. |
| 00:45.13 | landley | Lemme see... |
| 00:45.23 | landley | I want to take a picture of the nail board and put little arrows pointing to the components. |
| 00:45.36 | landley | I want to say which Linux /dev nodes and which drivers attach to each device. |
| 00:45.52 | landley | Model numbers are nice too, if I can link to a manufacturer page or something. |
| 00:46.05 | prpplague | landley: just a note, i'm totally worn out tonight so if i seem slow and/or act like a complete git, give me a little slack :( |
| 00:46.11 | landley | I'm not really looking for a spec sheet, though. More "these CONFIG_ symbols in the kernel correspond to this available hardware". |
| 00:46.16 | landley | No problem. |
| 00:46.20 | landley | Doesn't have to be tonight. |
| 00:46.33 | landley | I probably won't have a chance to _do_ anything with the info for a couple days, anyway. |
| 00:46.52 | prpplague | ok i'm getting it organized |
| 00:47.03 | prpplague | i have a top board profile printed |
| 00:47.05 | landley | One of the things I was sort of getting stuck on when I was making a hardware list is that the hardware is grouped strangely. |
| 00:47.23 | landley | The nail board presents a serial port to the outside world. |
| 00:47.45 | prpplague | you mean via the usb port? |
| 00:47.46 | landley | That serial port is sort of on the hammer chip, and sort of wired through the USB hub. |
| 00:47.52 | landley | Yup. |
| 00:47.56 | prpplague | yea, that portion is confusing |
| 00:48.04 | prpplague | i have block diagram for that |
| 00:48.05 | landley | Also, technically it's _two_ serial ports. |
| 00:48.28 | landley | The host system sees a /dev node, and the ARM linux also sees a /dev node on its end two. |
| 00:48.51 | landley | And they're wired together via a microscopic null modem cable inside the hammer. |
| 00:48.56 | prpplague | right that indeed is confusing if you don't understand it |
| 00:49.07 | landley | I'm trying to explain this without actually getting into that level of detail, which is a contradiction in terms. |
| 00:49.24 | landley | So there's a certain amount of jostling, tweaking, and gentle hand waving. :) |
| 00:49.28 | prpplague | landley: you know i think i'll get our inhouse art/media guy to draw up something |
| 00:49.36 | landley | I'm trying not to be _inaccurate_, but in some areas I'm oversimplifying. |
| 00:49.42 | landley | Cool. |
| 00:50.51 | landley | If you "view source" on http://landley.net/thyme.html you'll notice I have a few "img" tags, each of which has an ALT describing what the picture should be. |
| 00:51.00 | landley | I have none of these pictures. |
| 00:55.53 | landley | Awaiting nail board's arrival so I can take a few. |
| 00:55.54 | prpplague | indeed, its coming together , just not as fast as i wanted |
| 00:55.54 | prpplague | one sec gonna post the top board profile |
| 00:55.54 | landley | No shortage of things to do at this end. |
| 00:57.06 | prpplague | http://elinux.org/Nail_Board |
| 00:57.14 | prpplague | landley: ok the board profile has been updated |
| 00:57.51 | prpplague | landley: i'll start updating the spec |
| 00:58.45 | landley | Er, what changed? |
| 01:01.36 | hmmhesays | any of you guys use uclibc buildroot in here? |
| 01:01.50 | prpplague | landley: the board profile, it has the correct placement of all the components and headers |
| 01:02.00 | prpplague | hmmhesays: in a general way yes |
| 01:02.07 | landley | Ah, ok. |
| 01:02.38 | *** join/#edev chouimat|busy (n=dieu@r2351064.cidc.net) |
| 01:02.40 | landley | hmmhesays: in general, the channel for that is #uclibc. |
| 01:02.54 | landley | But feel free to ask questions here... |
| 01:02.57 | prpplague | landley: actually they send them here |
| 01:03.04 | landley | Not surprised. :) |
| 01:03.13 | prpplague | landley: vapier and the other don't like discussing buildroot there, only uclibc |
| 01:03.17 | landley | #uclibc has been mostly dead since Erik wandered off to do other things... |
| 01:03.23 | prpplague | indeed |
| 01:03.26 | landley | how about busybox? :) |
| 01:04.14 | hmmhesays | well I'm the documentation for cross compiling without integrating into buildroot is a mess |
| 01:04.59 | hmmhesays | I've found the gcc version I need to use, but can't figure out where to install my libraries after I compile (granted I am a noob at this) |
| 01:06.06 | landley | hmmhesays: I don't use buildroot. I made my own build system based on bash scripts instead of makefiles. |
| 01:06.12 | landley | http://landley.net/code/firmware |
| 01:06.19 | landley | Much easier to modify, I think. (But I'm biased.) |
| 01:06.36 | landley | If nothing else, much easier to figure out what happens in what order, and to insert or remove things at a specific point... |
| 01:06.46 | ds2 | Sh scripts? why not perl? ;) |
| 01:06.52 | ds2 | or elisp... |
| 01:07.01 | landley | Because other people can read shell scripts I write? |
| 01:07.05 | landley | perl is a write-only language. |
| 01:07.09 | landley | (And evil, but that's a separate issue.) |
| 01:07.22 | landley | And both of the people who know how to program in lithp are busy elsewhere. |
| 01:07.25 | ds2 | dare I ask about python build systems like scons? |
| 01:07.30 | hmmhesays | you cross compile with using bash scripts eh? |
| 01:07.34 | landley | Lots of Irritating Superfluous Parentheses. |
| 01:07.44 | landley | hmmhesays: I create a cross compiler with one script, and use it with another, yes. |
| 01:07.48 | landley | There are like six scripts total. |
| 01:07.59 | ds2 | for a script base cross compiler generation, there is always cross-tools |
| 01:08.04 | landley | Top level script calls the others in sequence, but you can skip that one and do it yourself if you like. |
| 01:08.10 | hmmhesays | hmm create a cross compiler? |
| 01:08.23 | landley | Yup. Build it from source. |
| 01:08.33 | landley | ./download.sh downloads the source packages you need. |
| 01:08.36 | hmmhesays | can you point me at some documentation for that perhaps? |
| 01:08.55 | landley | ./host-tools.sh builds a few things you might not have on your host system (you can usually skip that one). |
| 01:08.58 | hmmhesays | I may be a noob but I catch on quickly |
| 01:09.16 | landley | Part of the reason I wrote it as bash scripts (and added lots of comments) is that they _are_ documentation. |
| 01:09.23 | landley | "What commands do I run to do $X"... "here's a script". |
| 01:09.50 | prpplague | http://elinux.org/Nail_Board |
| 01:10.01 | prpplague | some basic specs, slowly adding |
| 01:10.03 | landley | http://landley.net/hg/firmware/file/tip/cross-compiler.sh |
| 01:10.14 | ds2 | actually, if you have a script to rewrite all the configure scripts to cross compile properly.... |
| 01:10.59 | landley | If you want more docs, though: http://landley.net/code/firmware/about.html http://landley.net/code/firmware/design.html http://landley.net/code/firmware/new_platform.html http://landley.net/code/firmware/downloads/README |
| 01:11.02 | hmmhesays | so you build gcc |
| 01:11.07 | landley | And binutils. |
| 01:11.10 | landley | And uClibc |
| 01:11.17 | landley | And I grab headers from the Linux kernel. |
| 01:11.24 | landley | You need those four packages to create a working cross compiler. |
| 01:11.34 | landley | Enough to build "hello world", anyway. |
| 01:11.51 | landley | (Which is enough to build a surprisingly large number of packages, actually.) |
| 01:12.07 | landley | To get a complete native build environment running on your target, you need to add busybox, make, and bash. |
| 01:12.19 | landley | (And you only need bash because the shell in the version of busybox I'm using isn't good enough.) |
| 01:12.33 | landley | I'm slowly replacing busybox with toybox (a different implementation). |
| 01:12.36 | landley | But that's a side issue. |
| 01:12.48 | landley | From those seven packages, you can bootstrap the whole of Linux From Scratch. |
| 01:12.59 | landley | Meaning you can build anything you want. (Gentoo, debian, etc.) |
| 01:13.07 | landley | If you know how to build one of those distros from source, that is. :) |
| 01:13.18 | landley | Never trivial building somebody else's package management system from source... |
| 01:13.33 | ds2 | bah distros... start from a root disk and go from there |
| 01:13.34 | landley | Even gentoo, you ask them about it and they give you blank looks. |
| 01:13.46 | landley | "You build distro from source under _gentoo_..." |
| 01:13.47 | landley | Sigh. |
| 01:13.57 | landley | This gives you a root disk with a version of gcc that can compile and run executables. |
| 01:14.08 | landley | ./cross-compiler.sh builds a cross compiler. |
| 01:14.24 | landley | ./mini-native.sh builds the minimal native environment for your target (those seven packages) using the cross compiler to do so. |
| 01:14.56 | landley | And then ./package-mini-native.sh creates an ext2 image out of it. (This needs User Mode Linux built by the host-tools.sh; if you skipped that step because UML is a _pain_ to build, you can always make your own ext2 image.) |
| 01:15.11 | landley | And in the build directory it creates "run-i686.sh" and "run-armv4l.sh" or whatever target you built. |
| 01:15.19 | landley | That's a script that invokes qemu to boot your new system up to a shell prompt. |
| 01:15.26 | landley | Feeding it the kernel and ext2 image you built. |
| 01:15.46 | landley | I got it working on i686, i586, armv4l, armv5l, ppc, and mips. |
| 01:16.02 | landley | I got sparc mostly working but there's a bug in uClibc stdio somewhere and I just don't care enough to track it down. |
| 01:16.11 | landley | m68k is half finished, I don't think that's checked in yet. |
| 01:16.47 | landley | now that qemu 0.9.1 is out I need to do armv6l and armv7l (and maybe armv4b and such too, big endian). |
| 01:16.57 | landley | Oh, and I did mipseb as well. Big and little endian mipses. |
| 01:17.15 | landley | And I need to do an EABI arm system. The ones I've done so far are oabi... |
| 01:17.32 | landley | (armv4l is actually "armv4, little endian, soft float, oabi". |
| 01:17.52 | landley | armv5l is armv5, little endian, vfp floating point, oabi. |
| 01:17.57 | landley | It should probably be armv5, little endian, vfp, eabi. |
| 01:18.06 | landley | I can tweak it next release, it's the debugging that's time consuming. |
| 01:20.34 | prpplague | landley: we've got an armv5 board in the works for ya |
| 01:21.00 | ds2 | are there other samsung ARM w/MMU boards besides the hammer? |
| 01:27.51 | landley | prpplague: cool. |
| 01:27.56 | landley | Is hammer v5 or v4? |
| 01:28.09 | landley | ds2: I believe so, yes. |
| 01:28.10 | prpplague | v4 |
| 01:28.15 | prpplague | ds2: tons |
| 01:28.27 | landley | Yay v4. |
| 01:28.33 | ds2 | prpplague: any significantly cheaper then the hammer and been hacked to run Linux? |
| 01:28.38 | landley | It's nice to have a test environment that _won't_ run v5 code. |
| 01:28.54 | landley | My main gripe about qemu is that I can't tell it "pretend to be a 386". It has to pretend to be an i686. |
| 01:29.12 | landley | So if I build an i586 distro, I'm not sure it'll actually run on hardware that doesn't implement i686 instructions. |
| 01:29.12 | ds2 | i want some familiarity with a more open source friendly processor now that PXA is being strangled |
| 01:29.26 | prpplague | ds2: there are only a couple, primary problem with them are that although they are cheaper, they require alot of work to get up and running |
| 01:29.33 | landley | Hard to prove it didn't leak newer instructions into the binaries in uClibc inline assembly or the kernel or just plain gcc/binutils. |
| 01:29.46 | landley | I _think_ it didn't, but it's nice to be able to actually _test_ this. |
| 01:29.59 | ds2 | landley: handcode a SIGILL handler and have it send you mail? =) |
| 01:30.09 | prpplague | ds2: when intel sold its xscale stuff to marvell, nearly everyone jumped ship to use the samsung SoC's |
| 01:30.19 | ds2 | prpplague: lots as in more then what the webpals required? |
| 01:30.30 | prpplague | ds2: ?? |
| 01:30.30 | landley | (Learned that with soft float. I thought I had working userspace soft float for a while, then figured out how to switch floating point emulation off in the kernel and found another couple things I needed to fix. :) |
| 01:30.33 | ds2 | prpplague: *nod* hence the search |
| 01:30.43 | landley | ds2: nah, a crash or hang is plenty. |
| 01:30.49 | prpplague | ds2: not sure i understand your question |
| 01:30.53 | landley | If I get to a shell prompt, usually it's working. :) |
| 01:30.56 | ds2 | prpplague: getting the webapls to run Linux tool some effort (cutting a few traces along with a tricky reflash) |
| 01:31.11 | ds2 | prpplague: do those other boards require more effort then that? |
| 01:31.25 | prpplague | ds2: sorry you miss understood, i was refering to the hardware side |
| 01:31.41 | prpplague | ds2: linux is easy on the samsung soc;s |
| 01:31.55 | prpplague | ds2: we've been banging away on linux support for literally years |
| 01:32.02 | ds2 | prpplague: I assume you meant boards in existing products that can be hacked to run Linux? |
| 01:32.23 | prpplague | ds2: hehe no, i thought you were asking about dev board |
| 01:32.50 | ds2 | ah we completely missed each other's points |
| 01:33.25 | prpplague | ds2: hehe yea, i think on a commercial end |
| 01:33.44 | ds2 | prpplague: and I am asking about a cheap trainer ;) |
| 01:34.09 | ds2 | you might get that FB driver.... |
| 01:34.24 | landley | ds2: We're working on making the nail board cheap. :) |
| 01:34.34 | ds2 | landley: sub $20? =) |
| 01:34.38 | landley | Er, no. |
| 01:34.42 | landley | Maybe $80. |
| 01:34.55 | landley | sub $20 you're talking a production run of a million units. |
| 01:35.10 | ds2 | hence the interest in prehacked toys |
| 01:35.18 | ds2 | where they do make a million + units =) |
| 01:35.27 | landley | Very few sub-$20 devices of _any_ kind are capable of running Linux. |
| 01:35.45 | landley | Cell phones run Linux, but the cheapest of those is $100 if the phone companies aren't paying for it for you. |
| 01:35.47 | ds2 | bankruptcy helps it |
| 01:36.03 | landley | Linksys routers are around $50, and the new ones are at the low end of potentially running LInux. |
| 01:36.12 | landley | Ah, _used_ stuff you can get for free. :) |
| 01:36.19 | landley | Take an old 386 off of a library's hands... |
| 01:36.50 | landley | I got a free tuxscreen a few years ago, after they went out of business. |
| 01:36.51 | ds2 | juice box comes close... if it only had an MMU |
| 01:36.56 | landley | Gave it to my friend mark. |
| 01:36.57 | prpplague | indeed |
| 01:37.32 | prpplague | ds2: you could try a neo |
| 01:37.40 | prpplague | ds2: but for that cost you could get a hammer kit |
| 01:37.43 | ds2 | prpplague: cheaper to buy a hammer board |
| 01:37.48 | prpplague | ds2: or you could try a used h1940 |
| 01:37.58 | prpplague | ds2: but not really that hackable |
| 01:38.01 | ds2 | as in an HP compaq? |
| 01:38.16 | ds2 | ipaq I mean |
| 01:38.56 | landley | to run linux, you need a cpu, ~2 megs storage, and 6 megs dram (with mmu, 4 megs without). |
| 01:39.19 | ds2 | 4M w/MMU is doable |
| 01:39.33 | ds2 | just NFS root it so no storage either ;) |
| 01:39.35 | landley | Possibly you could strip with-mmu linux to run in 4 megs, but the quick test I ran only had 500k free with 6 megs. |
| 01:39.48 | landley | NFS is evil. |
| 01:39.52 | ds2 | but are you running a ramdisk root? |
| 01:39.55 | landley | Not a File system. |
| 01:40.04 | landley | ds2: nah, ext2 image on a virtual IDE drive. |
| 01:40.08 | landley | qemu. :) |
| 01:40.19 | ds2 | hmmm must be bloat |
| 01:40.27 | ds2 | I have run Linux on a 4M 386SX before |
| 01:40.35 | landley | 2.6? |
| 01:40.42 | landley | And once again, qemu forgot how to reboot. Sigh. |
| 01:41.18 | ds2 | no, 1.0 or 0.99 |
| 01:41.20 | wmat | landley: someone at work today pointed out the uClibc cake pics on your uClibc page; we had a good chuckle |
| 01:41.42 | landley | wmat: did you think I was joking when I said I'd sent him a cake? |
| 01:41.45 | landley | Actual, physical cake. |
| 01:41.55 | wmat | heh |
| 01:41.56 | landley | Twice. |
| 01:42.13 | wmat | a customer sent us brownies today, which was how it come up |
| 01:42.14 | landley | ds2: those were much smaller releases. Didn't have kobjects, for one thing. |
| 01:42.20 | prpplague | ds2: thats kinda the point of the hammer |
| 01:42.35 | prpplague | ds2: there aren't any cheap dev platforms for the hottest soc in the market right now |
| 01:42.38 | landley | ds2: It is possible to strip it back down. |
| 01:42.41 | prpplague | ds2: thats where the hammer comes in |
| 01:42.44 | landley | That's what the -tiny tree is doing. |
| 01:42.51 | prpplague | wmat: hey |
| 01:43.00 | landley | Making lots of the added functionality (like /proc and /sysfs) configurable. |
| 01:43.01 | prpplague | wmat: hammer today? |
| 01:43.25 | landley | There's been some nice activity on the -tiny list in the past few days, Ingo's reviewing some of those patches for inclusion. |
| 01:43.55 | landley | prpplague: much nicer Nail_Board page. Thanks. |
| 01:44.15 | ds2 | prpplague: donno about market but for open source, yeah |
| 01:44.32 | ds2 | landley: the -tiny tree needs a lot of work :( |
| 01:45.03 | landley | ds2: I know, but Michael Opendacker is doing work on it. |
| 01:45.20 | landley | he's the new maintainer. |
| 01:45.23 | ds2 | landley: thought it was Tim's thing? |
| 01:45.31 | landley | Nah, he was just holding it for a while. |
| 01:45.33 | prpplague | ds2: even for market |
| 01:45.38 | landley | Matt Mackall started it but Mercurial ate his life. |
| 01:45.50 | landley | (The way Fabrice Bellard started tinycc before qemu ate his life.) |
| 01:45.59 | landley | Tim Bird has his hands full running CELF. |
| 01:46.00 | prpplague | ds2: i dare you to find a device that you can do proof-of-concept on a s3c2410 board less than $200 |
| 01:46.11 | landley | Michael became maintainer... /rummages... |
| 01:46.15 | hmmhesays | I was under the impression the buildroot was supposed to provide me with a toolchain for cross compiling, but when I compile something using bfin-linux-uclibc-gcc it sure seem to just compile for my native arch |
| 01:46.16 | ds2 | prpplague: not sure if I believe that... the PXA's + OMAP's got to be doing higher volume |
| 01:46.36 | ds2 | landley: is this after Mingo renewed interest? |
| 01:46.44 | prpplague | ds2: pxa might have been true if intel hadn't sold it to marvell |
| 01:46.56 | landley | http://lwn.net/Articles/251573/ |
| 01:47.03 | ds2 | prpplague: seems to still be a lot of legacy PXA270 users |
| 01:47.09 | prpplague | ds2: problem with both pxa and omap now is the volume needed for min qty |
| 01:47.10 | landley | ds2: nah, Mingo's interest started... tuesday, I think. |
| 01:47.18 | landley | The new maintainership happened in October. |
| 01:47.21 | prpplague | ds2: its driving most mid sized businesses out of that area |
| 01:47.29 | landley | http://elinux.org/Linux_Tiny_Patch_Ideas |
| 01:47.35 | prpplague | ds2: 5k to 10k is just too small for marvell and TI |
| 01:47.37 | ds2 | landley: eh? Mingo responded to the cedev list months ago |
| 01:47.47 | ds2 | prpplague: ah... |
| 01:47.51 | landley | http://elinux.org/Linux_Tiny |
| 01:48.06 | prpplague | ds2: samsung is makes small qty available |
| 01:48.12 | ds2 | n/m it was Morton who I was confusing it with |
| 01:48.13 | prpplague | ds2: which marvell and TI do not |
| 01:48.21 | landley | That last link's the right one. :) |
| 01:48.43 | ds2 | prpplague: don't have have to tell me that... I am buried in there but I can't comment for the usual reasons :/ |
| 01:48.49 | prpplague | landley: i;ve spoken to the art guy, we are going to draw up some stuff tomorrow |
| 01:48.54 | landley | Cool. |
| 01:49.08 | landley | I might get an update of thyme.html out tonight, dunno yet. |
| 01:49.22 | landley | (It's irresistable to use M.C. hammer puns when referring to these boards.) |
| 01:49.30 | landley | By the way, terminology question: |
| 01:49.32 | landley | hammer is a boad. |
| 01:49.33 | landley | board |
| 01:49.36 | landley | nail is a board. |
| 01:49.47 | landley | So the nail board contains a hammer board which contains an arm chip? |
| 01:49.53 | landley | Seems like there's an intermediate term missing. |
| 01:50.04 | landley | carrier? |
| 01:50.07 | landley | destroyer? |
| 01:50.10 | landley | light tactical device? |
| 01:50.49 | prpplague | yea, thats my bad |
| 01:51.47 | landley | Just wondering. |
| 01:51.59 | landley | I'd like to unambiguously describe it. What would you suggest? |
| 01:52.07 | landley | I believe I've been using "hammer module"... |
| 01:53.22 | prpplague | landley: well, techincally we are refering to it as "hammer cpu board" |
| 01:53.32 | prpplague | landley: and it should be "nail carrier board" |
| 01:53.58 | landley | Except that I can't quite either of those down to a single word other than the name. |
| 01:54.06 | prpplague | landley: indeed |
| 01:54.07 | landley | Referring to hammer as "the cpu" is incorrect. |
| 01:54.10 | landley | board is ambiguous. |
| 01:54.22 | SpeedEvil | The SoC? |
| 01:54.25 | prpplague | nail carrier? |
| 01:54.27 | landley | nail carrier I could sort of say, but out of context that one'll be confusing. |
| 01:54.31 | prpplague | hammer module? |
| 01:54.35 | landley | prpplague; they'll think nail carrier means the _case_. |
| 01:54.42 | landley | hammer module is what I've been using. |
| 01:54.50 | landley | nail board, hammer module, smc12345 chip. |
| 01:55.10 | landley | chip->module->board makes intuitive sense. |
| 01:55.15 | prpplague | ahh |
| 01:55.17 | prpplague | gotcha |
| 01:55.26 | prpplague | i'll discuss it with the boss tomorrow |
| 01:55.28 | landley | "What plugs into what" isn't something they have to stop and look up. This is good. |
| 01:56.51 | prpplague | indeed |
| 01:57.55 | prpplague | shoot i left the buzzer off |
| 01:59.50 | prpplague | landley: you see on the nail board profile how their are 20 pins on each side of the cpu module socket? |
| 01:59.56 | prpplague | s/their/there |
| 02:01.04 | landley | Yup. |
| 02:01.43 | landley | What about it? |
| 02:01.51 | prpplague | landley: if you populate those with pins, on the back side you can add a whole new board the same size as the nail board |
| 02:01.56 | prpplague | landley: basically stackable |
| 02:01.58 | landley | Those are the little plug board thingies you use to hook into those pins... |
| 02:02.10 | landley | Ah. |
| 02:02.45 | landley | I thought those were the "headers". |
| 02:02.49 | landley | You can use them as expansion thingies? |
| 02:02.59 | prpplague | landley: yea |
| 02:03.04 | landley | Cool. Good to know. |
| 02:03.10 | landley | I'm not writing up that paragraph, though. :) |
| 02:03.16 | landley | A hardware person should do that. |
| 02:03.16 | prpplague | landley: we won't populate them on the base device |
| 02:03.25 | prpplague | landley: yea, i'll write that up |
| 02:03.25 | landley | An end user can do so, though? |
| 02:03.29 | landley | Is there a standard part for it? |
| 02:03.29 | prpplague | landley: yea |
| 02:03.37 | prpplague | landley: radio shack special |
| 02:03.40 | landley | Also, if you decide to make up a few example expansion boards... |
| 02:03.51 | landley | proof of concept, and also demonstration. |
| 02:03.52 | prpplague | landley: yea, already working those up |
| 02:03.56 | landley | Something to photograph and stick in the wiki. :) |
| 02:04.09 | landley | A battery backed-up realtime clock board would be a nice example. :) |
| 02:04.16 | prpplague | landley: the 2x10 header can be used with either jumpers or you can add an expansion board there as well |
| 02:04.23 | prpplague | landley: i.e. an ethernet board |
| 02:04.32 | prpplague | landley: yep |
| 02:04.36 | prpplague | landley: you can do that too |
| 02:05.00 | prpplague | landley: as soon as i get the proto's in , i'll have some photo's of the expansion board examples |
| 02:05.31 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 02:05.34 | landley | Woot. |
| 02:07.12 | prpplague | RobotGuy: greetings |
| 02:07.15 | prpplague | RobotGuy: hows the arm? |
| 02:08.12 | RobotGuy | prpplague: I have cellulitis in my left arm, but they say I caught it in time. |
| 02:08.18 | RobotGuy | prpplague: Greetings |
| 02:08.35 | RobotGuy | I just got back from the ER. |
| 02:13.50 | prpplague | RobotGuy: my ex-wife had uncontrolled diabetes, i know how it is to always have those ER visits |
| 02:14.21 | RobotGuy | prpplague: I am prone to getting cellulitis. :( |
| 02:14.31 | *** join/#edev Heet (n=lucas@a82-92-5-118.adsl.xs4all.nl) |
| 02:14.43 | RobotGuy | Have I missed anything important? |
| 02:15.05 | prpplague | landley: shoot, i can't find my bookmark for your nail board docs? |
| 02:15.33 | prpplague | landley: got it handy? |
| 02:15.45 | prpplague | RobotGuy: nada, just busy with nail board stuff |
| 02:15.52 | landley | http://landley.net/thyme.html |
| 02:16.05 | landley | It's a random snapshot, what's on my laptop is a bit different from that... |
| 02:16.25 | prpplague | landley: thanks, just talked to the boss, we'll start slow changing over to "hammer module" |
| 02:16.26 | landley | Got bogged down updating the hardware bullet points and decided to wait for updates from upstream... |
| 02:16.34 | landley | Cool. |
| 02:19.59 | RobotGuy | I must be jinxed somehow. I can't seem to consistantly get my thumb drive to mount. Not sure what I am doing wrong. |
| 02:20.51 | ds2 | anyone know if a 7805 regulator can tolorate the range of spikes present in a automotive 12V electrical system? |
| 02:21.06 | prpplague | uh no |
| 02:21.18 | RobotGuy | ds2: That info should be in the datasheet. |
| 02:21.38 | SpeedEvil | RobotGuy: no it's not |
| 02:21.49 | ds2 | RobotGuy: it would be if I had a decent model or a description of the range, and magnitude of the spikes |
| 02:21.50 | prpplague | ds2: 7805 is gonna overheat at 12v |
| 02:21.51 | SpeedEvil | RobotGuy: it doesn't contain info on automotive fault conditions |
| 02:22.11 | SpeedEvil | prpplague: that does rather depend on the heatsink. |
| 02:22.14 | ds2 | prpplague: I am not too concerned with the static implications; I am just concerned with the spikes at the moment |
| 02:22.19 | RobotGuy | SpeedEvil: Shouldn't it contain info on maximum ratings and such though? |
| 02:22.26 | SpeedEvil | ds2: no - it won't. |
| 02:22.34 | ds2 | the 7805 has thermal shutdown |
| 02:22.45 | SpeedEvil | ds2: fault condutions on a car can run from about -80V to +80V |
| 02:22.55 | ds2 | ah numbers |
| 02:23.05 | SpeedEvil | ds2: this is not normal. In general. |
| 02:23.08 | prpplague | ds2: i mean that 12v is out of the range, iirc +10v is the max of 7805, |
| 02:23.16 | SpeedEvil | prpplague: no it's no |
| 02:23.20 | ds2 | SpeedEvil: can I protect that with 2 back to back 40V zeners |
| 02:23.21 | SpeedEvil | prpplague: +30V |
| 02:23.22 | prpplague | SpeedEvil: not? |
| 02:23.24 | SpeedEvil | IIRC |
| 02:23.27 | prpplague | SpeedEvil: you sure? |
| 02:23.28 | ds2 | prpplague: no, it is +40V |
| 02:23.47 | prpplague | hmm, must be thinking of something else |
| 02:23.51 | SpeedEvil | ds2: no - the simplest way to do this is: |
| 02:24.22 | SpeedEvil | ds2: one 1N4007 diode from the battery + supply to a 1 ohm 1W resistor, to a few thousand uF capacitor. |
| 02:24.39 | SpeedEvil | ds2: then the 7805 connected off the end of that. |
| 02:25.11 | ds2 | SpeedEvil: 1 ohm + say 1000uF cap is enough of a RC time constant to protect it from a postive going spike? |
| 02:25.13 | SpeedEvil | ds2: This protects you from reverse voltage - the resistor eats any fast spikes (an inductor works too, but is harder to select. |
| 02:25.28 | SpeedEvil | ds2: yes - the spikes don't have much energy in worst case |
| 02:25.35 | ds2 | ah okay |
| 02:25.37 | prpplague | SpeedEvil: ahh nice |
| 02:25.39 | ds2 | thanks |
| 02:26.16 | ds2 | SpeedEvil: what is the min. voltage rating of the cap? |
| 02:26.28 | SpeedEvil | ds2: If the battery remains connected properly to the system, the LF voltage stays between 8 and 16V under 'normal' fault conditions. |
| 02:26.37 | ds2 | <-- no idea how spikes are handled in a electrolytic cap :( |
| 02:26.48 | SpeedEvil | ds2: 16V is the absolute minimum. 22V is proabbly more sensible. |
| 02:26.53 | SpeedEvil | ds2: they can explode. |
| 02:27.03 | SpeedEvil | ds2: though only usually at sustained overvoltages. |
| 02:27.06 | ds2 | but that's failure |
| 02:27.19 | ds2 | so short spikes are tolerated w/o cat. failure |
| 02:27.33 | SpeedEvil | ds2: yes - 22V is probably the sane minimum that you'd go for. |
| 02:27.39 | SpeedEvil | prpplague: small switching PSUs. |
| 02:27.58 | SpeedEvil | prpplague: on paper PCBs with 8 pin chips. |
| 02:28.08 | prpplague | hehe |
| 02:28.33 | ds2 | some place sells tiny switchers in TO-220 compatible packages |
| 02:28.40 | SpeedEvil | yeah - those aren't used. |
| 02:28.43 | SpeedEvil | Too expensive. |
| 02:29.15 | ds2 | prpplague: speaking of game boy car chargers; they were selling them at the dollar store back in August |
| 02:29.21 | SpeedEvil | The |
| 02:29.24 | prpplague | hmm |
| 02:29.29 | SpeedEvil | Beware of dollar store SMPSs |
| 02:29.38 | SpeedEvil | I got one powered USB hub. |
| 02:29.47 | ds2 | SpeedEvil: why? these were properly branded |
| 02:29.51 | SpeedEvil | On investigation (stuff wasn't working) I investigated the PSU. |
| 02:30.08 | SpeedEvil | It was 6V off-load, and 2-4V under load. |
| 02:30.13 | SpeedEvil | (2V hum) |
| 02:30.23 | prpplague | yikes |
| 02:30.25 | SpeedEvil | It was a SMPS, with absolutely no input capacitor. |
| 02:30.28 | ds2 | btw, 7805C datasheet claims 35V max in - http://www.national.com/opf/LM/LM7805C.html |
| 02:30.45 | ds2 | 2V hum at what frequency? |
| 02:30.51 | SpeedEvil | ds2: 100Hz |
| 02:31.02 | ds2 | wtf... SMPS at 100Hz?! |
| 02:31.16 | ds2 | sounds like a transformer + diode to me |
| 02:31.42 | SpeedEvil | ds2: no - it was running at a higher speeed. It just diddn't have an input resovoir capacitor to feed off when the voltage of the input AC dropped. |
| 02:31.57 | SpeedEvil | Adding one from an old CFL bulb made it work. |
| 02:32.00 | ds2 | oh |
| 02:32.27 | MonMotha | 100Hz? |
| 02:32.30 | MonMotha | that's insane |
| 02:32.34 | MonMotha | who would design a SMPS at 100Hz? |
| 02:32.39 | MonMotha | the passives would be huge |
| 02:32.41 | SpeedEvil | MonMotha: err - no |
| 02:32.45 | prpplague | MonMotha: ping ping |
| 02:32.49 | SpeedEvil | MonMotha: the SMPS was running at 50Khz |
| 02:32.49 | prpplague | MonMotha: you get that url? |
| 02:32.54 | MonMotha | prpplague: no... |
| 02:32.57 | ds2 | ManMotha: if you need mechanical ballast... |
| 02:33.16 | prpplague | MonMotha: http://www.efunda.com/designstandards/plastic_design/plastic_intro.cfm |
| 02:33.33 | prpplague | MonMotha: according to the IE, that has the good tips for newbies |
| 02:33.40 | MonMotha | prpplague: awesome |
| 02:33.42 | prpplague | MonMotha: has complete info |
| 02:33.43 | MonMotha | dunno how far I'll take it |
| 02:33.51 | MonMotha | I suspect we will still outsource our more complicated designs |
| 02:34.28 | prpplague | indeed, but its a good start |
| 02:34.33 | SpeedEvil | MonMotha: however, normal SMPS have an input capacitor that's charged by the line voltage - it remains at 300V+ (in the UK). This PSU did not have that. It relied on the output capacitor for its (poor) supply capability when the input voltage got below 100V or so. |
| 02:34.34 | prpplague | MonMotha: i read thru it myself |
| 02:34.53 | MonMotha | SpeedEvil: ah, yeah |
| 02:34.54 | prpplague | MonMotha: think i might do a small sla just for the hell of it |
| 02:34.54 | MonMotha | that's dumb |
| 02:35.11 | MonMotha | SpeedEvil: I've done several SMPS designs at this point |
| 02:35.12 | SpeedEvil | MonMotha: I think it'd work with a 100mA load. |
| 02:35.21 | SpeedEvil | MonMotha: just |
| 02:35.24 | MonMotha | you don't have to oversize the input cap, but you do need one when running off AC! |
| 02:35.43 | MonMotha | (well, rectified AC) |
| 02:36.00 | SpeedEvil | MonMotha: unless you don't care about ouptut waveform - lighting transformers often produce modulated AC 12V at 50Khz |
| 02:36.47 | MonMotha | yes, but if you want DC... |
| 02:36.54 | SpeedEvil | Of course. |
| 02:39.13 | landley | If somebody did want to learn all this hardware stuff, where would one start? |
| 02:39.23 | SpeedEvil | At the beginning! |
| 02:39.24 | landley | Not that I really have time to discuss that question _now_... :) |
| 02:39.29 | SpeedEvil | More seriously. |
| 02:39.33 | landley | Do re mi? |
| 02:39.44 | SpeedEvil | I'd recommend picking up a copy of 'The Art of Electronics' |
| 02:39.55 | SpeedEvil | 2nd edition. |
| 02:40.06 | landley | Ok. |
| 02:41.32 | SpeedEvil | http://books.google.co.uk/books?id=bkOMDgwFA28C&dq=the+art+of+electronics&pg=PP1&ots=F28gPHcYVr&sig=RveGqCABfPzwPFUyo-fVs7oJvts&hl=en&prev=http://www.google.co.uk/search?client=firefox-a&rls=org.mozilla:en-US:official&channel=s&hl=en&q=the+art+of+electronics&btnG=Google+Search&sa=X&oi=print&ct=title&cad=one-book-with-thumbnail |
| 02:41.55 | prpplague | landley: you will get a ton of different answers |
| 02:42.02 | SpeedEvil | It's a relatively maths-lite treatment. |
| 02:42.04 | landley | prpplague: yeah, I know. |
| 02:42.11 | landley | Hence "no time". |
| 02:42.12 | prpplague | landley: for most software people slowly moving into digital electronics is easier |
| 02:42.13 | ds2 | AoE is nice but a bit outdate on a lot of stuff |
| 02:42.17 | SpeedEvil | It starts at the most basic models, and then builds from them. |
| 02:42.22 | landley | It's like asking "ok, how do I get into this whole programming thing". |
| 02:42.22 | prpplague | landley: then move into items such as the power supply stuff |
| 02:42.26 | SpeedEvil | Yes - 1989 for V2 |
| 02:42.31 | MonMotha | it kinda depends on what specifically you want to do |
| 02:42.35 | MonMotha | but yes, at the beginning |
| 02:42.38 | landley | The way _I_ got into it is 20+ years out of date. |
| 02:42.45 | prpplague | hehe |
| 02:42.50 | SpeedEvil | However, I know of no better 'all in one' book. |
| 02:43.04 | landley | What I tell people is "learn C". |
| 02:43.04 | SpeedEvil | AoE still has _wirewrap_ |
| 02:43.24 | landley | I've noted the book. |
| 02:43.44 | ds2 | landley: learn C with the "learn" program from the old bell labs unices? |
| 02:43.52 | prpplague | landley: crossing from the software world with bitmasks and registers, doing digital electronics portion is a little easier to move into |
| 02:43.58 | landley | And keep in mind I'm talking about "what interesting wires can I poke into the holes in the hammer board", not "how do I design a microprocessor out of resistors and diods and wires and such." |
| 02:44.00 | SpeedEvil | What I'd love would be an AoE moderated wiki |
| 02:44.07 | ds2 | not sure if they still make half the op-amps referenced in AoE :/ |
| 02:44.08 | landley | ds2: no. |
| 02:44.09 | prpplague | landley: indeed |
| 02:44.33 | SpeedEvil | I mean - start off with a wikified AoE |
| 02:44.36 | landley | More "ok, so I have a USB chip and an ARM chip, how _do_ you get them to talk to each other, anyway?" |
| 02:44.37 | prpplague | landley: the base portion of the hammer's i/o are all voltage 1's and 0's |
| 02:44.41 | landley | And how would one power this from a USB port? |
| 02:44.48 | landley | You guys are doing it _for_ me, but I'm curious... |
| 02:44.59 | prpplague | landley: the power portion starts to get complicated |
| 02:45.03 | landley | I have to be somewhere in 7 minutes. |
| 02:45.10 | landley | I'm very interestedin having this conversation, just not right now. :) |
| 02:45.11 | ds2 | so you want basically the reverse of what my RF engr friend askes me... how do I make this chip do XYZ |
| 02:45.16 | prpplague | landley: np, i have to drink a beer in 7 minutes |
| 02:45.17 | MonMotha | in terms of interface, it's pretty much a matter of finding compatible IO and gluing it together |
| 02:45.17 | prpplague | hehe |
| 02:45.36 | landley | With conductive glue? |
| 02:45.37 | prpplague | landley: np, when you have time |
| 02:45.39 | MonMotha | if you don't have compatible IO, then the glue gets slightly more complicated :) |
| 02:45.41 | landley | Elmer's. |
| 02:45.50 | ds2 | for primary (incident wave?) stuff, it is pretty straightforward |
| 02:45.58 | ds2 | the reflected wave technology can drive you bonkers |
| 02:46.00 | prpplague | landley: there are plenty of hardware junkies around to tutor |
| 02:46.12 | prpplague | landley: thats why this channel is such a boon |
| 02:46.24 | prpplague | landley: we have a great mix of software and hardware for faq |
| 02:46.25 | landley | I know about edge triggered vs level triggered and I even wrote a column once on how transistors work at a chemical level, as part of a week on microprocessor fabrication. |
| 02:46.39 | SpeedEvil | ##electronics too |
| 02:47.00 | MonMotha | if you want to learn reflected wave stuff, take a look at PCI System Architecture |
| 02:47.29 | landley | Must go now. |
| 02:47.33 | prpplague | landley: later |
| 02:47.59 | ds2 | what confuses me about reflected wave is, why can't I just drop in a normal input with a threshold that is above the summed value of the 2 waves |
| 02:48.13 | ds2 | but then I just avoid PCI like the plague ;) |
| 02:49.12 | prpplague | hehe |
| 02:49.14 | MonMotha | I have to admit that reflected wave stuff can throw me for a loop |
| 02:49.19 | MonMotha | just terminate the damned bus! |
| 02:49.32 | prpplague | hehe |
| 02:49.36 | prpplague | MonMotha: me too |
| 02:49.48 | ds2 | glad I am not the only one |
| 02:50.09 | MonMotha | you'll notice that they arne't as popular these days |
| 02:50.10 | prpplague | ds2: was it you that asked about teaching last night? |
| 02:50.18 | ds2 | prpplague: yeah |
| 02:50.22 | MonMotha | most busses are incident switched and they just terminate the thing |
| 02:50.27 | ds2 | saw your response about your boss talking to the univ. |
| 02:50.49 | ds2 | people argue about power consumption from the terminators |
| 02:51.14 | prpplague | ds2: on a personal note, i'd love to teach, i think the US educational system is broken since people who have experience in their field can not teach in their field |
| 02:51.48 | ds2 | prpplague: that's in a normal Univ; JCs should be a possible path for that |
| 02:52.52 | prpplague | ds2: i'd love to give 1 person a head start on doing this kind of dev |
| 02:53.55 | prpplague | ds2: i always wonder if i hadn't grown up in a redneck town if i'd gotten a better chance at getting a head start |
| 02:54.25 | prpplague | just have to work twice as hard |
| 02:54.32 | ds2 | yep |
| 02:54.38 | ds2 | same with growing up in the big cities |
| 02:54.56 | RobotGuy | prpplague: It's the same if one doesn't have any degrees. |
| 02:55.30 | prpplague | hehe i do have almost 500 semester hours, hehe, but no degree |
| 02:56.47 | ds2 | whoa geeks.com has the EeePC |
| 02:56.53 | prpplague | oh yea? |
| 02:57.00 | prpplague | ds2: how much? |
| 02:57.13 | ds2 | 359.99 for one and 409.95 for another.. donno the specs |
| 02:57.20 | prpplague | cool |
| 02:57.31 | prpplague | ds2: wish they'd had them for xmas |
| 02:57.33 | RobotGuy | prpplague: Wasn't it a lot harder to get where you wanted to be? |
| 02:58.47 | prpplague | RobotGuy: not really, generally the only time that its and issue is when other engineers ask "soo.... what engineering degree do you have?" |
| 02:59.38 | RobotGuy | prpplague: I think there is way too much emphasis put on degrees, which is mostly done by human resources folks. |
| 02:59.45 | prpplague | RobotGuy: indeed |
| 02:59.54 | prpplague | RobotGuy: most HR's have no clue what they are hiring for |
| 03:00.03 | prpplague | RobotGuy: its not like hiring someone to do filing |
| 03:00.11 | prpplague | RobotGuy: everyone understand filling papers |
| 03:00.21 | RobotGuy | prpplague: I know, but that's who you usually have to interview with first. |
| 03:00.31 | prpplague | RobotGuy: understanding sql database isn't something people understand |
| 03:00.36 | prpplague | RobotGuy: yea |
| 03:00.43 | prpplague | RobotGuy: thats why there are lots of tricks |
| 03:00.57 | RobotGuy | prpplague: I guess I never learned the tricks. |
| 03:01.10 | prpplague | RobotGuy: i.e. if your experience is postgresql, you put SQL database |
| 03:01.17 | prpplague | RobotGuy: not postgresql |
| 03:01.45 | prpplague | RobotGuy: because they might be looking for sqlite, or mysql , or one of the hundreds of variations |
| 03:02.22 | RobotGuy | prpplague: Oh, yeah, I do know the trick of putting general stuff down in place of specific stuff. |
| 03:02.33 | prpplague | RobotGuy: thats just one example |
| 03:03.11 | prpplague | RobotGuy: hehe, i got hired for an OMAP project at TI, gave my two weeks notice, and the day before i was to start work at TI they killed the project |
| 03:03.23 | RobotGuy | prpplague: My problem was always work history, which it turned out was because of a disability I didn't know I had until 2000. |
| 03:03.30 | prpplague | RobotGuy: hehe, i had another job by the end of the 2nd day |
| 03:03.46 | prpplague | RobotGuy: ahh |
| 03:04.52 | prpplague | RobotGuy: my problem is that our family has about a 300 year old curse |
| 03:05.25 | RobotGuy | prpplague: I'm not sure I believe in curses, but anything is possible, so I try to keep an open mind. |
| 03:05.30 | prpplague | RobotGuy: hehe, for about 300 years our family has been cursed with the skills to do lots of things, but great at nothing |
| 03:06.39 | prpplague | RobotGuy: we have a long line of scientists |
| 03:07.28 | RobotGuy | prpplague: Where is the buildroot repo(repository?) you referenced earlier today? I looked but did not see any way to search for issues regarding a specific package. |
| 03:08.18 | prpplague | RobotGuy: http://buildroot.busybox.net/cgi-bin/viewcvs.cgi/trunk/buildroot/ |
| 03:08.29 | prpplague | RobotGuy: know who Anders Celsius is? |
| 03:09.12 | RobotGuy | prpplague: Hmmm, that last name does seem familiar.. |
| 03:09.31 | prpplague | hehe, celsius temperature scale? |
| 03:10.15 | RobotGuy | prpplague: Yeah, that's the guy! |
| 03:10.25 | prpplague | one of his sons changed his last name to Anders in honor of his father |
| 03:10.46 | RobotGuy | prpplague: Oh! You have quite a geneology then. |
| 03:10.54 | prpplague | indeed |
| 03:11.06 | prpplague | RobotGuy: you ever see forest gump? |
| 03:11.14 | RobotGuy | prpplague: At least twice. |
| 03:11.41 | prpplague | hehe, you know how Lt.Dan had a long history of his family in war ? we have a long history in science |
| 03:11.54 | RobotGuy | prpplague: Ah, cool. :) |
| 03:12.21 | prpplague | RobotGuy: our most famous Anders - http://buildroot.busybox.net/cgi-bin/viewcvs.cgi/trunk/buildroot/ |
| 03:12.24 | prpplague | oops |
| 03:12.26 | prpplague | wrong url |
| 03:12.35 | prpplague | http://en.wikipedia.org/wiki/William_Anders |
| 03:13.08 | RobotGuy | As to buildroot repo.. I was hoping for something searchable that would bring up issues relating to a specific package. I guess there isn't such a thing. It would be handy when I find what I think is a problem, to see if it has been resolved. |
| 03:14.05 | prpplague | RobotGuy: you ever see a commercial during xmas where you see a picture of the earth and you hear people reading from the book of genesis? |
| 03:14.25 | prpplague | RobotGuy: why isn't that usable? |
| 03:14.32 | RobotGuy | prpplague: I am not sure I have seen that. |
| 03:14.37 | prpplague | RobotGuy: i found your problem with grep in about thirty seconds |
| 03:14.55 | prpplague | RobotGuy: i'm sure you've seen the photo "earth rise" |
| 03:15.08 | RobotGuy | prpplague: How is that searchable? I must be missing something. |
| 03:15.39 | prpplague | RobotGuy: i.e. you have a problem with grep, you go to the grep makefile, check the history and bam |
| 03:16.10 | prpplague | RobotGuy: http://www.abc.net.au/science/moon/earthrise.htm |
| 03:16.42 | prpplague | RobotGuy: http://buildroot.busybox.net/cgi-bin/viewcvs.cgi/trunk/buildroot/package/grep/grep.mk?rev=20692&view=log |
| 03:16.50 | prpplague | RobotGuy: scroll down and there it is |
| 03:16.59 | prpplague | RobotGuy: where someone fixed the dependency |
| 03:20.32 | RobotGuy | prpplague: Oh, OK, I think I begin to understand how to use that now, thanks. Could be due to the fact I have not slept in nearly two days because of my arm. |
| 03:20.43 | prpplague | np |
| 03:21.21 | prpplague | RobotGuy: i learn something new every day |
| 03:21.40 | RobotGuy | I think I will get some food and head to bed early. Technical stuff and loss of sleep does not usually mix well, at least with me. |
| 03:21.49 | prpplague | hehe |
| 03:22.07 | RobotGuy | I will see you tomorrow, hopefully after getting some sleep. :) |
| 03:22.55 | prpplague | http://en.wikipedia.org/wiki/Anders_(crater) |
| 03:48.36 | *** join/#edev landley (n=landley@cpe-72-177-114-214.austin.res.rr.com) |
| 04:06.52 | hmmhesays | How do I got about figuring out what path buildroot is using for gcc? |
| 04:30.33 | *** join/#edev dijenerate (n=dijenera@72.22.151.195) |
| 05:10.59 | *** join/#edev redblue (i=star@ppp117.108-253-207.mtl.mt.videotron.ca) |
| 05:50.45 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 05:52.15 | landley | hmmhesays: feed -v to gcc maybe? |
| 05:52.19 | landley | I dont' use buildroot... |
| 07:22.46 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 07:34.50 | *** join/#edev landley_ (n=landley@cpe-72-177-114-214.austin.res.rr.com) |
| 07:45.19 | *** join/#edev harpal (n=dhaval@124.125.76.8) |
| 08:06.06 | *** join/#edev ejdav_gon (n=james@75.157.13.250) |
| 08:18.20 | *** join/#edev steliosk (n=Stelios@85.72.92.83) |
| 09:14.00 | *** join/#edev linac (n=lin@125.77.72.66) |
| 09:38.42 | *** join/#edev lyakh (n=lyakh@p57BD2F1F.dip0.t-ipconnect.de) |
| 11:52.14 | *** join/#edev __lifeless (n=patrick@90.150.114.126) |
| 11:58.36 | *** join/#edev __lifeless (n=patrick@90.150.114.126) |
| 12:01.18 | *** join/#edev emil|k (n=emil@e179235076.adsl.alicedsl.de) |
| 12:07.12 | *** join/#edev dijenerate (n=dijenera@66.54.118.16) |
| 12:29.03 | *** join/#edev steliosk (n=Stelios@85.72.92.83) |
| 13:19.30 | *** join/#edev _lifeless (n=patrick@89.20.124.62) |
| 14:14.18 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 14:23.10 | *** join/#edev sjhill (n=sjhill@real.realitydiluted.com) |
| 14:28.45 | prpplague | ~ping sumstix.com |
| 14:28.45 | purl | pong sumstix.com |
| 14:28.51 | prpplague | ~ping gumstix.com |
| 14:28.52 | purl | pong gumstix.com |
| 14:29.06 | prpplague | ~nslookup gumstix.com |
| 14:29.38 | sjhill | yeah, i get 503 Service Unavailable |
| 14:30.03 | sjhill | looks like it's being chewed.... |
| 14:31.06 | prpplague | thats what i thought |
| 14:31.11 | prpplague | sjhill: morning btw |
| 14:37.06 | *** join/#edev linac (n=lin@121.41.36.85) |
| 14:40.44 | *** join/#edev T0mW (n=Tom@24.238.86.158.res-cmts.sth.ptd.net) |
| 14:40.53 | *** join/#edev bbradley (n=bbradley@82-32-35-42.cable.ubr05.azte.blueyonder.co.uk) |
| 14:44.30 | *** join/#edev GPSFan (n=kenm@12.10.255.246) |
| 14:44.30 | *** mode/#edev [+o GPSFan] by ChanServ |
| 14:52.12 | *** join/#edev chouimat|work (n=mchouina@209.217.106.98) |
| 14:55.01 | chouimat|work | morning |
| 14:55.21 | prpplague | chouimat|work: morning |
| 14:58.04 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 15:09.53 | sjhill | prpplague: g'morning |
| 15:09.57 | sjhill | hey chouimat|work |
| 15:10.06 | chouimat|work | hi sjhill |
| 15:15.26 | *** join/#edev FireEgl (i=FireEgl@Sebastian.Atlantica.DollarDNS.Net) |
| 15:31.46 | *** join/#edev samuelr (n=samuelr@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 15:36.12 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 15:36.12 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 15:51.35 | *** join/#edev emil|k (n=emil@e179232119.adsl.alicedsl.de) |
| 15:58.07 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 16:14.47 | *** join/#edev ico2 (n=ico2@modem-3927.cougar.dialup.pol.co.uk) |
| 16:49.42 | prpplague | wmat: ping |
| 16:52.12 | wmat | prpplague: pong |
| 16:52.28 | wmat | prpplague: haven't even powered it up yet |
| 16:52.31 | prpplague | wmat: hehe |
| 16:52.38 | prpplague | wmat: just curious |
| 16:52.50 | wmat | prpplague: i'm stupid busy with crap right now |
| 16:52.58 | prpplague | wmat: np |
| 16:53.29 | prpplague | wmat: if/when you get support done, let me know, tincantools would like to do a joint PR |
| 16:53.44 | wmat | prpplague: will do |
| 16:54.22 | prpplague | jbevren: ping |
| 16:59.32 | wmat | prpplague: I also haven't moved elinux.org yet, due to being too busy |
| 17:00.17 | prpplague | wmat: np, that seems to be the current condition for everyone |
| 17:00.40 | *** join/#edev T0mW (n=Tom@24.238.86.158.res-cmts.sth.ptd.net) |
| 17:00.51 | prpplague | T0mW: hehe, i was just typing up an email to you |
| 17:00.57 | T0mW | heh |
| 17:01.20 | T0mW | prpplague: well, I could save you that stamp |
| 17:01.26 | prpplague | hehe |
| 17:01.43 | T0mW | the 4 cent USPO email postage |
| 17:01.57 | prpplague | T0mW: i need a small favor |
| 17:02.00 | prpplague | T0mW: http://www.allelectronics.com/cgi-bin/item/LCD-101/365/256_X_128_LCD_PANEL_.html |
| 17:02.22 | prpplague | T0mW: you got time in the next week or so to do a schematic up for making that work with the hammer? |
| 17:03.36 | prpplague | T0mW: should be pretty simple, a couple of 245's |
| 17:03.48 | prpplague | T0mW: and the power supplies |
| 17:04.06 | T0mW | looking |
| 17:06.24 | T0mW | ow |
| 17:06.42 | T0mW | lcd bias voltage is +30v |
| 17:07.23 | prpplague | T0mW: iirc thats misleading |
| 17:07.27 | T0mW | naw |
| 17:07.29 | prpplague | T0mW: its a negative |
| 17:07.30 | T0mW | it is real |
| 17:08.13 | T0mW | typically 15v |
| 17:08.47 | T0mW | prpplague: how many volts ya' got? |
| 17:09.02 | prpplague | T0mW: regulated +5v |
| 17:09.08 | T0mW | ok |
| 17:09.34 | T0mW | it looks like it would need a level shifter 245 |
| 17:09.38 | prpplague | T0mW: right |
| 17:09.47 | prpplague | <prpplague> T0mW: should be pretty simple, a couple of 245's |
| 17:09.51 | prpplague | hehe |
| 17:10.22 | T0mW | 74lvx245 |
| 17:10.42 | T0mW | run the 74lvx245 @ 3.3v |
| 17:10.52 | prpplague | T0mW: curious, why not a 74ahct245? |
| 17:14.51 | prpplague | T0mW: am i correct that this one uses a negative bias voltage - http://docs.bgmicro.com/pdf/lcd1030.pdf |
| 17:17.59 | prpplague | T0mW: typical for that lcd is -15v for the bias |
| 17:18.07 | prpplague | T0mW: Vdd-Vo |
| 17:18.39 | T0mW | prpplague: because your VIH is 2v on that device |
| 17:19.02 | prpplague | T0mW: ahh |
| 17:19.19 | prpplague | T0mW: yea, on the both of those lcd's they are using Vdd-Vo |
| 17:19.28 | prpplague | T0mW: which means both need -10v for the bias |
| 17:19.35 | prpplague | T0mW: correct? |
| 17:21.38 | prpplague | T0mW: sorry, looks to be -15v on one and -12 on the other (for standard temp) |
| 17:25.53 | prpplague | HA - http://www.foxnews.com/story/0,2933,324800,00.html |
| 17:25.57 | prpplague | funny funny |
| 17:30.16 | T0mW | prpplague: you want a 74LCX family |
| 17:30.33 | prpplague | T0mW: okie dokie |
| 17:30.46 | prpplague | T0mW: you hook me up with a schematic? |
| 17:31.38 | T0mW | prpplague: ah, problem is that you would have probs finding a PDIP part for that. |
| 17:32.26 | prpplague | T0mW: interesting though, we use the ahct for all of our legacy lcds |
| 17:32.35 | prpplague | r |
| 17:33.35 | T0mW | prpplague: the 'X' means level-shifting (3.3 <--> 5v) |
| 17:34.17 | prpplague | so whats the difference between a ahct and lcx? |
| 17:35.26 | T0mW | the input voltage thresholds and supply voltage |
| 17:35.52 | prpplague | hmm |
| 17:36.07 | T0mW | ahct is a 2.0 volt minimum for a HIGH, lcx is 1.5v min |
| 17:36.21 | T0mW | vcc on ahct is 5v, vcc on lcx is 3.3v |
| 17:36.55 | T0mW | lcx is 5volt tolerant, ahct is maybe ok to use with 3.3v logic. |
| 17:37.09 | prpplague | that is odd then |
| 17:37.13 | T0mW | depends on VoH spec of the cmos |
| 17:37.39 | prpplague | T0mW: all the refence schematics i've seen for 3.3v lcd signals to 5v lcd's all use ahct |
| 17:37.40 | T0mW | low speed you can get away with it, higher speeds you should see some problems. |
| 17:37.53 | T0mW | prpplague: who showed you those? |
| 17:38.11 | *** join/#edev emil|k_ (n=emil@e179232119.adsl.alicedsl.de) |
| 17:38.12 | prpplague | T0mW: just a collection of ones i've found on the net |
| 17:39.14 | prpplague | T0mW: you have some time to fix me up a schematic? |
| 17:39.54 | T0mW | prpplague: I could do that, all it would be is bit-bang |
| 17:40.19 | prpplague | T0mW: yea, since it has an onboard controller |
| 17:40.32 | prpplague | T0mW: i just need the guts of the power supply and level shifting |
| 17:40.43 | T0mW | prpplague: why use that one? |
| 17:41.04 | prpplague | T0mW: because its cheap and popular |
| 17:41.15 | T0mW | LOL |
| 17:41.40 | prpplague | T0mW: really what i'll do is make a small board that can be used with a range of 5v lcd's |
| 17:42.13 | prpplague | T0mW: hehe hammer board is in feb issue of circuit cellar's new products section |
| 17:52.27 | T0mW | prpplague: simple circuit that I used on one design to get 13..26volts for LCD: http://www.openhardware.net/LcdSupply.png |
| 17:52.33 | *** join/#edev ico2 (n=ico2@modem-3927.cougar.dialup.pol.co.uk) |
| 17:53.32 | prpplague | T0mW: yea, just a reminder, both of those lcds use negative bias voltages(Vdd-Vo) |
| 17:54.20 | prpplague | T0mW: yea, i've used the m34063 chips from ti |
| 17:56.25 | steliosk | anyone interested OE has working avr32 support in the .dev branch |
| 17:57.43 | prpplague | steliosk: seems like avr32 is picking up speed |
| 17:57.54 | prpplague | price of their dev boards is darn reasonable |
| 17:58.11 | steliosk | it would had more speed in OE if atmel decided to supported ;) |
| 17:58.17 | prpplague | but then when you are the size of atmel you can do that kind of pricing |
| 17:58.35 | steliosk | oh well its the very well known story |
| 17:58.39 | prpplague | steliosk: atmel isnt cooperating? |
| 17:58.44 | steliosk | i give you the saving machine |
| 17:58.55 | steliosk | buy the rasors |
| 17:58.58 | steliosk | ;) |
| 17:59.23 | prpplague | steliosk: ?? |
| 17:59.25 | steliosk | prpplague : its a rather long story, that makes no sense now |
| 17:59.42 | prpplague | oh |
| 17:59.47 | prpplague | s/saving/shaving |
| 17:59.50 | steliosk | yes |
| 17:59.52 | steliosk | sorry |
| 18:00.17 | steliosk | anyway code is there free as in speech and beer :) |
| 18:17.00 | GPSFan | prpplague: ping, got my lcd displaying text, very wierd, the font is reversed. each char is correct but backwards. I tried playing with the LCDCON5(1:0) swap bits but they didn't help. it is as if the font is left-right reversed. using fbcon |
| 18:37.10 | *** join/#edev steliosk (n=Stelios@athedsl-117730.home.otenet.gr) |
| 18:43.06 | *** join/#edev jdav_gone (n=james@75.157.13.250) |
| 19:00.19 | prpplague | GPSFan: pong |
| 19:00.26 | prpplague | GPSFan: i meant to tell you about that |
| 19:00.48 | GPSFan | prpplague: found the problem, the hammer puts out data msb first, the lcd wanted it lsb first, working now... |
| 19:01.09 | prpplague | GPSFan: yea |
| 19:01.15 | prpplague | GPSFan: which lcd are you using again? |
| 19:01.26 | GPSFan | lm32k10 sharp |
| 19:01.32 | prpplague | ahh yea |
| 19:01.50 | prpplague | GPSFan: i meant to tell you about msb/lsb swap the other day but forgot |
| 19:01.52 | prpplague | GPSFan: sorry about that |
| 19:02.14 | prpplague | GPSFan: hows it look? |
| 19:02.22 | prpplague | GPSFan: thats a 5v display right? |
| 19:02.29 | GPSFan | np, only took a few min to figure it our and swap the connections. |
| 19:02.32 | GPSFan | yes |
| 19:02.48 | prpplague | GPSFan: which 245 are you using? |
| 19:02.49 | GPSFan | using a 74AC244 as a buffer |
| 19:02.53 | prpplague | GPSFan: ahh |
| 19:03.36 | GPSFan | -18v bias |
| 19:03.45 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 19:03.53 | RobotGuy | Good Morning |
| 19:04.00 | *** join/#edev samuelr (n=samuelr@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 19:04.04 | prpplague | GPSFan: cool, gonna post some schematics? |
| 19:04.07 | prpplague | RobotGuy: greetings |
| 19:04.14 | prpplague | samuelr: hey there! |
| 19:04.43 | GPSFan | prpplague: I'm going to take some pix this afternoon, and try and get some sutff up on the wiki this evening. |
| 19:04.54 | prpplague | GPSFan: good deal |
| 19:05.00 | prpplague | GPSFan: wants some more lcds? |
| 19:05.31 | prpplague | GPSFan: http://docs.bgmicro.com/pdf/lcd1030.pdf |
| 19:05.38 | GPSFan | prpplague: maybe, other stuff going on over the next few weeks, I'll let you know. |
| 19:05.45 | prpplague | GPSFan: okie dokie |
| 19:08.31 | GPSFan | prpplague: btw you know that I rebuilt the hammer buildroot using a snapshot (back in oct) using uclibc 0.9.29, binutils 2.18, mfpr 2.3.0, gcc 4.1.2, fakeroot 1.7.1, and busybox 1.7.1. |
| 19:09.04 | prpplague | GPSFan: if i did know, i had forgotten, hehe |
| 19:09.33 | GPSFan | k |
| 19:14.17 | GPSFan | prpplague: that lcd uses KS0108B controllers, it will need to be a parallel gpio based interface, the internal lcd controller won't do. |
| 19:14.31 | prpplague | GPSFan: yea i know |
| 19:14.51 | prpplague | GPSFan: we'd use the same lines as the lcd controller, just in gpio mode |
| 19:15.02 | GPSFan | ah, ok I thought you might be thinking it was a real lcd iterface. np... |
| 19:15.46 | prpplague | GPSFan: naw, just that its cheap, popular, and easy to find |
| 19:17.24 | GPSFan | yeah that will be a useful interface to get working, all sorts of projects need a small display for status & such. |
| 19:17.40 | prpplague | indeed |
| 19:17.48 | prpplague | ~seen TimRiker |
| 19:17.49 | purl | timriker <n=timr@pdpc/supporter/bronze/TimRiker> was last seen on IRC in channel #infobot, 1d 2h 15m 25s ago, saying: 'infobot: status'. |
| 19:25.08 | prpplague | T0mW`gone: http://www.futurlec.com/Mini_Logic.shtml |
| 19:27.49 | RobotGuy | ~seen me |
| 19:27.51 | purl | me <n=me@079-046-128-083.dynamic.caiway.nl> was last seen on IRC in channel #norganna, 10d 9h 8m 21s ago, saying: 'maybe its allready bid on'. |
| 19:28.00 | *** join/#edev roxfan2 (n=dunno@159.208-64-87.adsl-dyn.isp.belgacom.be) |
| 19:28.47 | RobotGuy | prpplague: I almost bought something from Futerlec one time. |
| 19:29.10 | prpplague | yea i was looking at some of their miniboards |
| 19:29.18 | prpplague | those would work well with the hammer |
| 19:29.41 | RobotGuy | I read a BUNCH of bad reviews of Futerlec, so decided against ordering from them. |
| 19:29.53 | prpplague | oh |
| 19:30.11 | prpplague | RobotGuy: bad service? or bad products? |
| 19:33.23 | *** join/#edev sjhill (n=sjhill@real.realitydiluted.com) |
| 19:33.54 | *** mode/#edev [+o prpplague] by ChanServ |
| 19:36.56 | RobotGuy | prpplague: Bad service |
| 19:37.05 | prpplague | RobotGuy: ahh |
| 19:38.23 | prpplague | sjhill: greetings |
| 19:42.14 | RobotGuy | prpplague: Plus, I e-mailed them about availablility of something I was interested in and never did get a reponse. |
| 19:42.23 | RobotGuy | Hi sjhill |
| 19:42.29 | prpplague | RobotGuy: ahh ok |
| 19:51.04 | prpplague | ~seen sjhill |
| 19:51.07 | purl | sjhill is currently on #edev (17m 44s) #uclibc (17m 44s) #elinux (17m 44s), last said: 'hey chouimat|work'. |
| 19:53.06 | RobotGuy | I was interested in their STAMP modules, and I see they have some other interesting stuff now too. |
| 20:01.18 | RobotGuy | prpplague: Could you make that sdc patch relative to the top of the kernel directory? |
| 20:01.33 | prpplague | RobotGuy: sdc patch? |
| 20:01.41 | prpplague | oh you mean the udc patch? |
| 20:01.49 | RobotGuy | prpplague: Yeah, that's what I meant. :) |
| 20:02.05 | prpplague | RobotGuy: i'll be posting a update kernel patch in the next few days |
| 20:02.13 | RobotGuy | prpplague: Ah, OK. |
| 20:02.34 | RobotGuy | prpplague: I am testing the udc patch now. |
| 20:08.32 | RobotGuy | prpplague: With the udc patch, I did not get /mnt/flash mounted. |
| 20:08.54 | prpplague | the udc patch has nothing to do the with /mnt/flash |
| 20:11.20 | RobotGuy | prpplague: Without the udc patch I get /mnt/flast mounted. |
| 20:11.36 | RobotGuy | <PROTECTED> |
| 20:11.48 | prpplague | the udc patch has nothing to do the with /mnt/flash |
| 20:12.34 | prpplague | RobotGuy: the only thing the udc patch does is enable the usb slave controller |
| 20:14.01 | RobotGuy | prpplague: Interesting. |
| 20:14.32 | RobotGuy | prpplague: I am just reporting my results. :) Going back to a minimal kernel with just usb-storage and gadget ether enabled. |
| 20:14.54 | prpplague | RobotGuy: i suspect you changed an option in your kernel config |
| 20:15.10 | RobotGuy | prpplague: That's why I am going back to a 'vanilla' kernel. |
| 20:15.50 | RobotGuy | prpplague: I started this one out with hammmer-config and just enabled gadget ether. |
| 20:16.01 | prpplague | okie dokie |
| 20:16.32 | *** join/#edev ico2 (n=ico2@217.134.80.151) |
| 20:16.58 | RobotGuy | prpplague: I think it's important to find where my kernel setup fails. |
| 20:22.37 | *** join/#edev redblue (i=star@ppp189.108-253-207.mtl.mt.videotron.ca) |
| 20:23.57 | RobotGuy | prpplague: What is the correct fstab line to mount the usbfs? |
| 20:25.04 | prpplague | RobotGuy: mount -t usbfs usbfs /proc/bus/usb |
| 20:25.31 | prpplague | RobotGuy: should occur after the mounting of the procfs |
| 20:30.06 | RobotGuy | prpplague: OK, added that. |
| 20:45.08 | *** join/#edev landley (n=landley@cpe-72-177-114-214.austin.res.rr.com) |
| 20:58.26 | chouimat|work | hi landley |
| 20:59.40 | landley | Which is a strange name to have to spell, and apparently the pipe char gets interpreted by my client as some sort of command... |
| 21:00.05 | chouimat|work | :) |
| 21:14.00 | landley | So how goes stuff? |
| 21:18.45 | landley | Lunchtime... |
| 21:26.53 | RobotGuy | It's quiet today. Everybody must be working or getting into mischief. |
| 21:39.24 | RobotGuy | Mirell: Sort of working is not an option. ;) |
| 21:40.01 | Mirell | Sure it is. |
| 21:40.31 | RobotGuy | Mirell: Nope. You must be working or getting into mischief. ;) |
| 21:42.57 | *** join/#edev catastrop (n=catastr0@d14-69-178-65.try.wideopenwest.com) |
| 21:47.10 | *** join/#edev davidc__ (n=davidc__@S01060002b360aacd.vc.shawcable.net) |
| 21:57.26 | Mirell | Heh |
| 22:14.47 | *** join/#edev ico2 (n=ico2@modem-1803.cheetah.dialup.pol.co.uk) |
| 22:38.43 | agaffney | heh |
| 23:01.49 | hmmhesays | /src/.libs/libiksemel.a(libiksemel_la-stream.o): In function `handshake': |
| 23:01.49 | hmmhesays | /opt/astfin/build_stamp537/iksemel-1.3/src/stream.c:73: undefined reference to `_gnutls_global_init' |
| 23:02.05 | hmmhesays | gnutls_global_init is in gnutls.h |
| 23:08.12 | hmmhesays | root/include/gnutls/gnutls.h: int gnutls_global_init (void); |
| 23:09.31 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 23:09.31 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 23:19.57 | *** join/#edev Redhatter-Out (n=stuartl@gentoo/developer/redhatter) |
| 23:34.36 | RobotGuy | hmmhesays: This is from what? |
| 23:36.00 | RobotGuy | hmmhesays: Are you trying to build a toolchain for the stamp537 (Analog Devices) |
| 23:36.20 | *** join/#edev dijenerate_ (n=dijenera@72.51.101.205) |
| 23:38.13 | *** join/#edev dijenerate_ (n=dijenera@72.51.101.205) |
| 23:43.22 | *** join/#edev dijenerate_ (n=dijenera@72.51.101.205) |
| 23:44.59 | hmmhesays | RobotGuy: yeah I am |
| 23:53.18 | *** join/#edev dijenerate_ (n=dijenera@72.51.101.205) |
| 23:55.36 | hmmhesays | unsuccessfully it seems |
| 23:59.27 | RobotGuy | You may be in the wrong place for Analog Devices support. We mostly do ARM stuff here. |
| 23:59.32 | *** join/#edev dijenerate_ (n=dijenera@72.51.101.205) |
