| 05:39.07 | *** join/#edev purl (~purl@rikers.org) |
| 05:39.08 | *** topic/#edev is Discussion of embedded systems (Hardware/Software/Linux/Commercial Device Repurposing) : Busybox and Buildroot : http://www.elinux.org : Logs available at ibot.rikers.org |
| 05:48.28 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw2-50de26-12.dhcp.inet.fi) |
| 07:23.30 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 07:37.59 | *** join/#edev NightMonkey (~NightrMon@pdpc/supporter/professional/nightmonkey) |
| 08:02.56 | *** join/#edev ffio_ (~fire@unaffiliated/security) |
| 08:51.18 | *** join/#edev Zygo (foobar@startkeylogger.hungrycats.org) |
| 08:54.21 | *** join/#edev nullpuppy (~dustin@freematrix/staff/nullpuppy) |
| 09:01.04 | *** join/#edev nullpuppy (~dustin@freematrix/staff/nullpuppy) |
| 09:15.10 | *** join/#edev Anaphaxeton (~george@unaffiliated/anaphaxeton) |
| 09:39.25 | *** join/#edev cahbtexhuk (~cahbtexhu@82-68-209-214.dsl.in-addr.zen.co.uk) |
| 11:48.58 | *** join/#edev g1powermac (~g1powerma@rrcs-74-87-150-6.west.biz.rr.com) |
| 11:48.59 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 13:02.58 | *** join/#edev groglogic (~groglogic@ip-64-134-157-206.public.wayport.net) |
| 13:59.16 | *** join/#edev g1powermac (~g1powerma@rrcs-173-197-150-13.west.biz.rr.com) |
| 13:59.17 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 14:18.19 | *** join/#edev damkerngt (~Administr@27.130.221.15) |
| 14:30.31 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 14:47.17 | *** join/#edev dgerlach (~a0274052@192.91.66.186) |
| 15:14.30 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 15:14.55 | *** join/#edev CosmicPe- (~nobody@199.106.109.189) |
| 15:16.45 | *** join/#edev old-papa-work (~wl@65-117-102-210.dia.static.qwest.net) |
| 15:21.00 | *** join/#edev CosmicPenguin (~nobody@znc.codeaurora.org) |
| 15:21.00 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 15:22.06 | *** join/#edev prpplague (~danders@adsl-68-88-78-118.dsl.rcsntx.swbell.net) |
| 15:40.21 | *** join/#edev Anaphaxeton (~george@unaffiliated/anaphaxeton) |
| 16:02.36 | *** join/#edev _ffio_ (~fire@unaffiliated/security) |
| 16:24.57 | *** join/#edev NightMonkey (~NightrMon@pdpc/supporter/professional/nightmonkey) |
| 16:29.28 | *** mode/#edev [+o prpplague] by ChanServ |
| 16:44.10 | *** join/#edev Crofton (~balister@dadsl-245036.rol.raiffeisen.net) |
| 17:32.47 | *** join/#edev groglogic (~groglogic@63.225.113.81) |
| 17:40.16 | *** join/#edev groglogic (~groglogic@63.225.113.81) |
| 18:17.58 | *** join/#edev g1powermac (~g1powerma@rrcs-74-87-150-6.west.biz.rr.com) |
| 18:17.58 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 18:43.50 | *** join/#edev dgerlach (~a0274052@192.91.66.186) |
| 18:48.52 | *** join/#edev bsdfox_ (~Bob@unaffiliated/bsdfox) |
| 19:18.31 | landley | If anybody would like to try http://lkml.indiana.edu/hypermail/linux/kernel/1307.2/00695.html and acked-by it... |
| 19:19.13 | prpplague | looks |
| 19:19.46 | kergoth | interesting |
| 19:38.00 | *** join/#edev NightMonkey (~NightrMon@pdpc/supporter/professional/nightmonkey) |
| 20:53.47 | *** part/#edev dgerlach (~a0274052@192.91.66.186) |
| 20:54.42 | *** join/#edev das_plague (~danders@adsl-68-88-78-118.dsl.rcsntx.swbell.net) |
| 21:38.20 | *** join/#edev cahbtexhuk (~cahbtexhu@cpc4-lewi16-2-0-cust814.2-4.cable.virginmedia.com) |
| 21:47.12 | cahbtexhuk | hi everybody |
| 21:47.16 | cahbtexhuk | I need some advices |
| 21:50.38 | cahbtexhuk | I need a tiny linux machine for managing distributed home automation system |
| 21:51.29 | cahbtexhuk | 20-40 devices on serial bus (rs485) and ethernet |
| 21:52.24 | cahbtexhuk | basically, just collect some data from time to time, log it to machine and upload it later somewhere |
| 21:52.50 | cahbtexhuk | via small application |
| 21:53.01 | cahbtexhuk | maybe a web interface for the whole thing |
| 21:54.34 | cahbtexhuk | raspi and other beaglebones aren't really what might be usefull, because they need additional hardware stacked on top |
| 21:54.58 | cahbtexhuk | and PSU and stuff |
| 21:55.19 | cahbtexhuk | plus they dont really fit inside DIN case |
| 21:56.25 | cahbtexhuk | we thought about making our own units to do the trick |
| 21:56.36 | MonMotha | you might look at PC104 and derived formfactors |
| 21:56.57 | cahbtexhuk | we thought about making our own device |
| 21:57.07 | MonMotha | well, that's always an (often expensive) option |
| 21:57.12 | MonMotha | you'll get exactly what you want, in theory |
| 21:57.19 | cahbtexhuk | like, I want to do it myself, to gain some experience |
| 21:57.33 | MonMotha | if you do want to go that route, there are people in here who can provide hints or (commercially) do turnkey design |
| 21:57.54 | cahbtexhuk | plus we have the budget and plans to do such devices ourselves and sell them afterwards |
| 21:57.58 | MonMotha | are you looking to start with a system-on-module type thing or do you want to do a ground-up design? |
| 21:58.43 | cahbtexhuk | what's a ground-up design? |
| 21:58.55 | cahbtexhuk | I'm sorry, english isn't my native language ): |
| 21:59.13 | MonMotha | as in doing a fully custom board starting to carry all the parts (SoC, RAM, storage, IO, etc.) |
| 21:59.59 | MonMotha | you can get so-called "system on module" where basically the hard part, which is mostly integrating the SoC with RAM+flash and getting it to boot, is already done for you. You just drop a socket down on your custom carrier board and hook the IO up as you please. |
| 22:00.18 | MonMotha | For low volume, this is usually a good idea, but the pricing on such modules is not attractive for high volume use |
| 22:00.39 | cahbtexhuk | yeah, i've already noticed |
| 22:00.56 | cahbtexhuk | cheapest I've found was about 60£ |
| 22:01.48 | *** join/#edev groglogic (~groglogic@ip-64-134-157-206.public.wayport.net) |
| 22:03.10 | MonMotha | depending on your volume, that can range for a total steal to really expensive |
| 22:03.55 | MonMotha | NRE on integrating a Linux-capable SoC into a new design can easily hit $50,000 |
| 22:04.01 | MonMotha | (quite easily well more) |
| 22:04.33 | cahbtexhuk | we need about 200 devices, then - around 1k/mo |
| 22:04.52 | MonMotha | 200 initially then shipping 1000/mo? |
| 22:05.00 | cahbtexhuk | +/- |
| 22:05.29 | cahbtexhuk | problem is - I have no idea, what power is needed for running described above |
| 22:05.44 | cahbtexhuk | back at university I've played with SBC based on AT91SAM9260 |
| 22:06.13 | MonMotha | there are all sorts of embedded/industrial targeted ARM SoCs |
| 22:06.30 | MonMotha | I've used the SAM9260 (and SAM9263, which was...interesting) |
| 22:06.59 | MonMotha | the somewhat modern equivalent to the ARM926EJ-S would be a Cortex-A5 |
| 22:07.04 | MonMotha | Atmel has a part with said core |
| 22:07.54 | MonMotha | Linux implies a certain minimal feature set: you need enough RAM to run the kernel, enough storage to hold it (compressed or more for XIP), and the CPU needs an MMU |
| 22:08.30 | cahbtexhuk | yeah, I have basic understanding of what should be there |
| 22:09.23 | cahbtexhuk | the catch is - call it a "school project" or whatever |
| 22:09.41 | MonMotha | the big thing is the RAM requirement: most of these MCUs use DDR2 RAM |
| 22:09.48 | cahbtexhuk | I have about 500£ budget, my own tools and my free time |
| 22:10.13 | cahbtexhuk | if I succeed developing a system - I'm good and I get the cookie |
| 22:10.30 | cahbtexhuk | if I fail - okay, 500£ well spent anyway |
| 22:10.46 | cahbtexhuk | so I'm a bit limited with technology |
| 22:10.54 | MonMotha | routing a DDR2 memory bus is not for the uninitiated |
| 22:11.03 | cahbtexhuk | and I'm looking for something, that can be hand soldered |
| 22:11.50 | MonMotha | oh...that puts you firmly in module territory |
| 22:11.55 | MonMotha | most of these CPUs are BGAs |
| 22:12.18 | MonMotha | if you're lucky, you can find a giant 0.5mm QFP, which probably doesn't meet your definition of hand solderability (though it's quite doable) |
| 22:13.05 | cahbtexhuk | I'm quite good at soldering, so 0.5 is not a challenge... |
| 22:13.08 | cahbtexhuk | ...anymore |
| 22:13.09 | cahbtexhuk | :D |
| 22:13.47 | MonMotha | ah, well, then you may have some options - not a lot though |
| 22:13.57 | MonMotha | DDR2 RAM is available in leaded packages, so you're OK there |
| 22:14.08 | MonMotha | the PCB design is the big problem, though |
| 22:14.11 | cahbtexhuk | apart from routing |
| 22:14.14 | cahbtexhuk | yes |
| 22:14.15 | MonMotha | DDR2 is quite fussy |
| 22:14.44 | MonMotha | The good news is that the design rules are quite well understood and documented by JEDEC. If you follow said rules, you'll probably have good luck with it working. |
| 22:14.56 | cahbtexhuk | I'll probably need to do the wavy tracks to keep the length the same, right?? |
| 22:15.15 | MonMotha | they also need to have controlled impedance and proper terminationj |
| 22:15.38 | MonMotha | for point-to-point (one chip), it's not terrible, and you can probably avoid doing multidrop for a system this smal |
| 22:15.58 | MonMotha | your volume is kinda right on the line between module and custom design. You might try and push some of the module vendors to see if they can give you a price break |
| 22:18.17 | cahbtexhuk | ok, I'll keep that in mind |
| 22:18.43 | cahbtexhuk | thx a lot, I now have some food for thought |
| 22:20.14 | cahbtexhuk | btw, do you have any experience with FreeRTOS/OpenRTOS? |
| 22:22.27 | MonMotha | yes, I've used FreeRTOS a fair bit |
| 22:22.43 | MonMotha | it's a pretty nice embedded RTOS, though minimalistic (which is good and bad) |
| 22:22.56 | MonMotha | you can of course run FreeRTOS on much smaller systems than Linux |
| 22:23.41 | MonMotha | my only real complaint with FreeRTOS is its over-reliance on heap allocation |
| 22:33.44 | cahbtexhuk | my complaint is that you need to plan reaaaally carefully, what you add in task and how much stack you allocate for each task |
| 22:34.05 | cahbtexhuk | I have a strange bug and I don't know, hwo to deal with it |
| 22:34.41 | MonMotha | well, that's true of any stack switching RTOS |
| 22:35.15 | MonMotha | there also exist single-stack RTOSes, but they can only preempt one way (toward higher priority), and all tasks have to have unique priority |
| 22:35.18 | MonMotha | so there's no round-robin |
| 22:35.40 | MonMotha | IIRC, FreeRTOS has such a mode that you can configure, but I don't know if it actually only uses a single stack |
| 22:36.44 | cahbtexhuk | I have a problem with spi flash |
| 22:37.28 | cahbtexhuk | I'm trying to send FW image and i'm buffering it on flash |
| 22:37.43 | cahbtexhuk | doesn't really matter what I do, but anyway |
| 22:38.01 | cahbtexhuk | i'm writing data to flash in 256 byte pages |
| 22:39.02 | cahbtexhuk | damn, I think I just realised, what might be the problem |
| 22:41.08 | cahbtexhuk | no, that's not the case, something different |
| 22:43.52 | cahbtexhuk | okay, so, when I try to write a page to Fash sometimes the page gets corrupted half way through and, say, starting from byte wahtever till the end of the page all bytes are 0xff |
| 22:44.52 | *** join/#edev kyou (~slashbeas@unaffiliated/slashbeast) |
| 22:45.05 | *** join/#edev g1powermac_ (~g1powerma@rrcs-74-87-150-6.west.biz.rr.com) |
| 22:46.14 | cahbtexhuk | same for reading |
| 22:46.48 | cahbtexhuk | well, that's the only way, how I can find out, if it wen't wrong |
| 22:47.28 | cahbtexhuk | I've found a workaround - I disable interrupts before I issue a write command and reenable them after it's done |
| 22:47.47 | cahbtexhuk | all portENTER_CRITICAL macros don't really help |
| 22:48.23 | cahbtexhuk | but after disabling-reenabling interrupts board becomes slow to respong |
| 22:48.44 | *** join/#edev das_plague (~danders@adsl-68-88-78-118.dsl.rcsntx.swbell.net) |
| 22:49.32 | cahbtexhuk | and funny enough, I can send image to update with normal speedm and the second time will take forever to upload it |
| 22:49.36 | *** join/#edev _ffio_ (~fire@unaffiliated/security) |
| 22:50.13 | cahbtexhuk | and ping grows from 1-4ms to 600-800ms ): |
| 23:14.00 | *** join/#edev XandriX (xandrix@gateway/shell/anapnea.net/x-ppucpvnhslrasbdl) |