| 00:31.56 | prpplague | g1powermac bills |
| 00:32.07 | prpplague | g1powermac i think we have a winner for the SoC on the Didj |
| 00:34.09 | g1powermac | ahh |
| 00:34.15 | g1powermac | oh, thats cool |
| 01:48.17 | *** join/#edev dijenerate (n=dijenera@65.48.164.209) |
| 01:58.04 | MonMotha | prpplague: figured something out, did you? |
| 02:10.23 | prpplague | MonMotha i can't claim the price on it |
| 02:10.36 | prpplague | MonMotha a buddy pointed out that i was looking for a 393mhz device |
| 02:10.45 | prpplague | MonMotha doesn't mean that is the max speed of the device |
| 02:11.03 | prpplague | MonMotha so i widended the range |
| 02:11.20 | prpplague | MonMotha found something that matches |
| 02:11.46 | prpplague | MonMotha http://www.mesdigital.com/english/Products/product_pollux.asp |
| 02:12.06 | prpplague | MonMotha right number of pins, right size, right capabilities |
| 02:12.45 | prpplague | i had all the pieces, but i missed the fact that it was down clocked |
| 02:12.59 | prpplague | hangs his head in shame |
| 02:18.10 | MonMotha | prpplague: indeed, that seems to fit |
| 02:18.27 | MonMotha | isn't 533MHz AWFULLY fast for an ARM926EJ-S? |
| 02:18.39 | MonMotha | actully, they don't claim it's a -S |
| 02:18.44 | MonMotha | which may help a little |
| 02:18.53 | prpplague | yea |
| 02:18.53 | MonMotha | still seems REALLY fast for ARM926EJ |
| 02:19.09 | MonMotha | I'd expect MAYBE 433, probably more like 350 |
| 02:19.25 | prpplague | that was my expectation as well |
| 02:19.46 | MonMotha | I wonder if the temp range on that is overly restrictive |
| 02:19.48 | prpplague | but there seems to be several companies that have tweak 926 for those upper speeds |
| 02:20.05 | MonMotha | most of the 926s I've seen are spec'd for industrial or even automotive tmep |
| 02:20.10 | prpplague | MonMotha well i lied on their registration and downloaded a datasheet |
| 02:20.27 | MonMotha | they claim 70C, but who knows |
| 02:21.09 | prpplague | the pinouts look to match |
| 02:21.23 | MonMotha | well, if it is that, you should be able to find JTAG |
| 02:21.37 | MonMotha | is the board all through-hole vias or does it have blind vias under the BGA? |
| 02:21.53 | prpplague | lots of blind vias, but there are some pads |
| 02:22.00 | MonMotha | ah, that makes it harder |
| 02:22.06 | prpplague | which the traces seem to match up with the datasheet |
| 02:22.22 | MonMotha | I guess they make this thing in high enough volume that the cost of the blind vias isn't terrible |
| 02:23.31 | MonMotha | I've wanted to do a design with blind/buried for a while, but I haven't been able to justify the cost |
| 02:24.36 | prpplague | indeed, our next handheld will be omap3xx based, which is going to demand blind vias |
| 02:24.48 | MonMotha | when you get down to 0.5mm, it's pretty much required |
| 02:24.59 | MonMotha | it's eithre that or microvias, at which point the "blind" part tends to be included |
| 02:25.17 | MonMotha | 0.5mm pitch may require blind microvias anyway. That's TIGHT pitch |
| 02:25.45 | prpplague | indeed, i don't envy the layout guy |
| 02:26.43 | MonMotha | heh, I find that stuff interesting |
| 02:26.54 | MonMotha | it's hard to be replaced by a computer in that job, too |
| 02:27.03 | MonMotha | the problem is NP-complete |
| 02:27.13 | MonMotha | (reduces to 3-SAT) |
| 02:27.54 | MonMotha | Modern layout packages are pretty good. They've moved the focus away from "autorouting" and more to "interactive" routing |
| 02:29.07 | *** join/#edev shaz (n=shaz@202.179.138.35) |
| 02:30.42 | MonMotha | prpplague: even at 0.65mm, blind vias are required for 8mil PTH w/ 5mil annular ring, 3/3 trace/space |
| 02:31.11 | MonMotha | though if you can get your PCB maker down to a 4mil ring, it works. |
| 02:31.21 | MonMotha | (that breaks IPC specs, though) |
| 02:31.33 | prpplague | MonMotha one of the best layouts i've ever seen was done by a woman who had no training and had no electronics experience |
| 02:31.46 | prpplague | MonMotha she would do fantastic work |
| 02:35.10 | MonMotha | you don't need a EE background to do a good chunk of layout work |
| 02:35.28 | MonMotha | it can help when you get into the signal integrity, EMC, and controlled impedance/RF stuff, but you can pick that up as you go, too |
| 02:35.40 | MonMotha | I do both layout and circuit stuff |
| 02:36.05 | MonMotha | there is something to be said for having the EE who did the circuit design also do the layout. Tends to result in better layout tradeoffs being made. |
| 02:36.20 | MonMotha | also reduces "I can't do that" ECO turnaround time |
| 02:44.31 | prpplague | heads to bed |
| 02:44.37 | prpplague | MonMotha later bud, i'm beat |
| 02:51.19 | *** join/#edev pea (n=pea@adsl-75-58-4-163.dsl.dytnoh.sbcglobal.net) |
| 03:13.31 | *** join/#edev dijenerate (n=dijenera@65.48.164.209) |
| 03:20.54 | *** join/#edev pea (n=pea@adsl-75-58-4-163.dsl.dytnoh.sbcglobal.net) |
| 03:30.50 | *** join/#edev DrZimmerman (n=theo@77-56-242-58.dclient.hispeed.ch) |
| 04:04.27 | *** join/#edev dijenerate (n=dijenera@65.48.164.209) |
| 04:35.05 | *** join/#edev dijenerate (n=dijenera@65.48.164.209) |
| 04:48.47 | *** join/#edev mbuf (n=user@61.16.248.242) |
| 05:06.55 | *** join/#edev edlikestoboogie (n=Edlikest@pool-72-89-123-175.nycmny.fios.verizon.net) |
| 05:08.37 | edlikestoboogie | Can an arm7 decode mp3s while running a webserver (w/ tcp/ip stack) and http upload? |
| 05:10.08 | MonMotha | that's probably pushin git |
| 05:10.32 | MonMotha | though I'd suspect RAM would be as much of an issue as CPU time on most ARM7s |
| 05:10.46 | MonMotha | unless you have one of the rare ARM7s with external memory and have hooked up lots |
| 05:11.27 | edlikestoboogie | can't you get ARM7s with as much ram as your heart desire? |
| 05:11.28 | edlikestoboogie | s |
| 05:11.35 | edlikestoboogie | it should still be cheap, no? |
| 05:12.43 | MonMotha | edlikestoboogie: the SoCs tend to be limited to 256 or sometimes 512k |
| 05:12.54 | MonMotha | and they start getting a little pricey at the 512k point, like $15-20 |
| 05:13.08 | MonMotha | even 256k of RAM is kinda rare. Most are 64k or less |
| 05:13.25 | MonMotha | I'm thinking ARM7TDMI. Most people are using ARM9 for the big stuff these days. |
| 05:13.29 | *** join/#edev lifeeth (n=Praneeth@unaffiliated/lifeeth) |
| 05:13.32 | MonMotha | ARM926EJ is especially popular |
| 05:13.42 | edlikestoboogie | how much are those? |
| 05:14.17 | edlikestoboogie | I heard ARM9e's DSP instructions will help in mp3 decoding |
| 05:14.24 | MonMotha | yes, they certainly can |
| 05:14.30 | edlikestoboogie | but i figured it would be too expensive |
| 05:14.46 | MonMotha | the bare MCU (with maybe 32k of SRAM) is usualy in the $15-20 range, but you'll need external RAM and storage then |
| 05:15.04 | MonMotha | there are some showing up with ~512k of flash and 64-128k of SRAM |
| 05:15.09 | MonMotha | that MIGHT be enough |
| 05:15.17 | MonMotha | it'll be tight on the RAM still, I suspect |
| 05:15.31 | MonMotha | TCP isn't lightweight, and there's a memory footprint/speed tradeoff for the MP3 decode |
| 05:15.46 | MonMotha | certainly don't expect to run a "real" OS like Linux |
| 05:15.55 | lifeeth | edlikestoboogie, how about an atngw? |
| 05:15.58 | MonMotha | you'll be stuck using something like LWIP and FreeRTOS |
| 05:16.05 | edlikestoboogie | what is that lifeeth? |
| 05:16.12 | MonMotha | an AVR32 might be good for this |
| 05:16.16 | lifeeth | It is a avr32 based board... |
| 05:16.17 | edlikestoboogie | yes.. freeRTOS is what I have been looking at |
| 05:16.21 | MonMotha | heh, there you go |
| 05:16.29 | lifeeth | Comes with uclinux on board... |
| 05:16.36 | edlikestoboogie | why would it suit my needs lifeeth? |
| 05:16.39 | MonMotha | AVR32 is targeted to multimedia applications with networking |
| 05:16.39 | edlikestoboogie | and is it cheap? |
| 05:16.49 | lifeeth | and 2 network cards + SD card for the memory |
| 05:16.50 | edlikestoboogie | ahh.. |
| 05:16.55 | lifeeth | it is cheap... |
| 05:17.20 | lifeeth | ~95$ for a new board...with everything loaded |
| 05:18.20 | edlikestoboogie | yes but what microcontroller does it have? |
| 05:18.37 | edlikestoboogie | I need to run a webserver with http upload |
| 05:18.38 | lifeeth | AVR32 based on...Google for the exact number... |
| 05:18.38 | edlikestoboogie | of mp3s |
| 05:18.42 | MonMotha | AVR32 AP7 series |
| 05:18.52 | MonMotha | AP7000, looks like, which I think is the big one |
| 05:18.56 | lifeeth | I have one lying around...IT comes with an web server on board |
| 05:19.19 | MonMotha | you can run full blown Linux on it no problem. Looks like the board has 32MB of RAM, 8MB of Flash + some dataflash and a SD slot |
| 05:19.32 | edlikestoboogie | I was speaking of cheap, as in production wise.. I am doing a homework on embedded systems |
| 05:19.33 | lifeeth | Yeah...It is good... |
| 05:19.43 | lifeeth | Ah...The processor is 9$ |
| 05:19.50 | lifeeth | So you can calculate the rest |
| 05:19.57 | MonMotha | I think that proc is available in QFP, too, so cheaper PCB |
| 05:20.00 | MonMotha | and cheaper assembly |
| 05:20.14 | MonMotha | if not, I think it might be a 1mm BGA, at least |
| 05:20.43 | lifeeth | They are selling it for 95$....So it must cost them ~50$ to make I guess |
| 05:21.03 | MonMotha | hum, looks like 1mm BGA only. That's standard design rules at least, just have to pay to get the BGA soldered down at assembly |
| 05:21.48 | edlikestoboogie | doesn't look cheaper than an arm7 |
| 05:22.19 | MonMotha | no, but unlike an ARM7TDMI, this has a good chance of working for you :) |
| 05:22.32 | MonMotha | trying to do all that on an ARM7TDMI is likely to result in hair loss |
| 05:22.51 | edlikestoboogie | but if I could prove my point that it is doable in class |
| 05:22.55 | edlikestoboogie | that would get me an A |
| 05:23.01 | *** join/#edev Blu3 (i=david@reserved21-vlan999.bcrtfl01.us.wh.verio.net) |
| 05:23.01 | MonMotha | Atmel has an app note of doing something very similar to this (though they made a shoutcast client) on one of their ARM7TDMIs, and they had to bump own the sample rate |
| 05:23.39 | MonMotha | on a BIG ARM7TDMI, with lots of memory, you might be able to pull off CD spec decoding (44.1kHz, 16 bit output) |
| 05:23.54 | MonMotha | but I'm not aware of any ARM7TDMIs with enough memory |
| 05:24.23 | MonMotha | there are non TDMI variants of ARM7, but they're getting rare, and usually they rely on external memory which puts you into the same scenario as the AVR32, and I think the procs cost more |
| 05:26.03 | edlikestoboogie | why is so much memory needed to decodes mp3s? |
| 05:31.10 | edlikestoboogie | this mp3 decoder uses 37k https://datatype.helixcommunity.org/Mp3dec |
| 05:32.27 | *** join/#edev redblue (i=star@ppp014.108-253-207.mtl.mt.videotron.ca) |
| 05:58.44 | edlikestoboogie | I wish there was a site to compare ARM9 and arm9e microcontrollers like this site for arm7s: http://www.gnuarm.com/ArmDevices_frame.html |
| 05:58.47 | edlikestoboogie | anyone know any? |
| 06:25.52 | *** join/#edev mbuf (n=user@61.16.248.242) |
| 07:15.47 | *** join/#edev [AD]Turbo (n=turbo@host198-64-dynamic.61-82-r.retail.telecomitalia.it) |
| 07:16.55 | [AD]Turbo | hola |
| 07:24.34 | *** join/#edev boris_OmegA (n=boris@194.51.141.172) |
| 07:31.24 | *** join/#edev ccjoe (n=joeg@75-169-18-107.slkc.qwest.net) |
| 07:37.41 | *** join/#edev KazaLite (n=kazaLite@116.58.8.130) |
| 07:50.49 | *** join/#edev lyakh (n=lyakh@p57BD252F.dip0.t-ipconnect.de) |
| 07:59.04 | *** join/#edev JoeLlama (n=snork@unaffiliated/joellama) |
| 08:14.28 | *** join/#edev _12ways (i=TWWAYS@121.134.254.81) |
| 08:16.57 | *** join/#edev TurboAWAY (n=turbo@host63-243-dynamic.6-87-r.retail.telecomitalia.it) |
| 08:45.48 | *** join/#edev harpal (n=Harpal@122.169.108.195) |
| 10:29.07 | *** join/#edev g1powermac_PB (n=g1powerm@unaffiliated/g1powermac) |
| 10:37.51 | *** join/#edev JoeLlama (n=snork@unaffiliated/joellama) |
| 10:46.03 | *** join/#edev chouimat (n=dieu@r2351064.cidc.net) |
| 11:20.17 | *** join/#edev g1powermac (n=g1powerm@76.177.147.36) |
| 13:19.02 | *** join/#edev GPSFan (n=kenm@64.92.145.112) |
| 13:19.02 | *** mode/#edev [+o GPSFan] by ChanServ |
| 13:27.10 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 13:59.26 | *** join/#edev flavior (n=flavior@201-65-98-68.poolip.JPA.embratel.net.br) |
| 14:06.44 | *** join/#edev chouimat|work (n=dieu@209.217.106.98) |
| 14:27.42 | *** join/#edev kgilmer (n=kgilmer@ip-66-80-197-243.atl.megapath.net) |
| 14:37.18 | prpplague | kgilmer: greetings |
| 14:37.32 | kgilmer | good morning prpplague |
| 14:37.57 | prpplague | kgilmer: cookin bugs for breakfast? hehe |
| 14:38.16 | kgilmer | haha. actually instant shark fin soup :) |
| 14:38.46 | chouimat|work | morning |
| 14:38.50 | prpplague | kgilmer: interesting |
| 14:38.53 | prpplague | chouimat|work: greetings |
| 14:39.42 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 14:39.42 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 14:46.12 | wmat | why do people come here, ask questions, then leave. They obviously don't realize that IRC answers have a turnaround speed of 24 hours or so. |
| 14:46.48 | wmat | or they realize that their question was easily answered by googling |
| 14:49.05 | chouimat|work | 24 hours is way too long |
| 14:49.07 | CosmicPenguin | Thats the modern IM/twitter world |
| 14:49.15 | CosmicPenguin | these people are obviously way too busy to sit around and wait for knowledge |
| 14:49.46 | solar | how do I solve problem A? /part |
| 14:50.09 | chouimat|work | CosmicPenguin: s/busy/impatient |
| 14:50.16 | solar | perhaps ppl should just ask the question in the /quit msg. |
| 14:50.42 | CosmicPenguin | Whatever discourages them from asking questions is fine by me |
| 14:54.25 | wmat | heh |
| 14:58.29 | CosmicPenguin | You pay our dues, you get the clue |
| 14:58.29 | CosmicPenguin | s |
| 15:04.31 | *** join/#edev gandhijee (i=akp@host-66-202-34-165.spr.choiceone.net) |
| 15:15.20 | prpplague | hehe |
| 15:15.41 | prpplague | yea, they act like we sit around just waiting to answer questions on a momments notice |
| 15:16.54 | prpplague | CosmicPenguin: man those qualcomm folks are persistant |
| 15:17.04 | CosmicPenguin | they need clue |
| 15:20.21 | CosmicPenguin | http://lkml.org/lkml/2008/9/24/443 |
| 15:20.49 | CosmicPenguin | I need a mutt plugin that will take a email from the LKML and automatically give me the link to lkml.org |
| 15:25.18 | prpplague | looks |
| 15:26.54 | prpplague | hehe |
| 15:27.39 | wmat | read 'mutt plugin' very, very incorrectly :/ |
| 15:28.33 | prpplague | CosmicPenguin: i like that idea, i'd definetly submit a bunch of our drivers |
| 15:29.02 | prpplague | CosmicPenguin: alot of them aren't really that useful to anyone else, but they'd serve as a good base for others |
| 15:32.27 | prpplague | kgilmer: when you get your hammer kit, i'm interested in hearing any feedback you might have |
| 15:32.49 | prpplague | kgilmer: the bug stuff have jtag available? |
| 15:33.39 | wmat | haha, they were optionally going to call TAINT_CRAP, TAINT_INCOMPETENT_VENDOR |
| 15:33.43 | prpplague | CosmicPenguin: although i think the usage of "CRAP" isn't very professional |
| 15:33.54 | CosmicPenguin | call a spade a spade |
| 15:34.04 | wmat | the other option was TAINT_GREG |
| 15:34.12 | prpplague | CosmicPenguin: there are better ways of descibing that |
| 15:34.42 | prpplague | TAINT_SUBSTANDARD |
| 15:35.06 | ohsix | whats all this about? |
| 15:35.18 | prpplague | ohsix: http://lkml.org/lkml/2008/9/24/443 |
| 15:35.19 | CosmicPenguin | Why coddle people? |
| 15:35.26 | CosmicPenguin | if its crap, call it crap |
| 15:35.47 | prpplague | like i said, not very professional using that word |
| 15:35.53 | CosmicPenguin | why not? |
| 15:35.56 | CosmicPenguin | is it going to offend somebody? |
| 15:36.06 | CosmicPenguin | and if it does, perhaps those people need offending |
| 15:37.40 | prpplague | we will just have to argee to disagree on this one, hehe |
| 15:44.17 | wmat | prpplague: in a way, it's labelling unprefessional code appropriately |
| 15:45.57 | prpplague | personally i consider the use of "crap" unprofessional, so you are using unprofessional wording to label unprofessional code |
| 15:46.36 | prpplague | so in essence you are saying "we right professional code, but we don;t have to act professional" |
| 15:46.57 | wmat | heh, fair enough |
| 15:47.23 | CosmicPenguin | The disconnect is that I don't think any of the kernel developers consider the code professional |
| 15:47.27 | CosmicPenguin | it is what it is |
| 15:49.32 | prpplague | CosmicPenguin: i think you are going to find that alot of the vendors who do linux support are going to provide patches to alter that |
| 15:50.30 | CosmicPenguin | You mean distributions? |
| 15:50.53 | prpplague | CosmicPenguin: more like companies like MV, Timesys, and WindRiver |
| 15:50.57 | CosmicPenguin | I doubt either Greg or Dave Jones would change it |
| 15:51.15 | CosmicPenguin | And is there any past presidence for a vendor to maintain a patch against upstream to change wording? |
| 15:51.16 | prpplague | CosmicPenguin: i doubt that as well |
| 15:51.29 | kgilmer | will let you know prpplague |
| 15:51.57 | prpplague | CosmicPenguin: no with the kernel, but with other applications |
| 15:52.05 | prpplague | s/no/not |
| 15:53.20 | prpplague | CosmicPenguin: sysv init in the past has had a number of objectionable error messages and comments in the code |
| 15:53.48 | prpplague | CosmicPenguin: of course i've not looked at the code for sysv init in years since i mostly use busybox init |
| 15:59.44 | prpplague | CosmicPenguin: kind of a moot point anyway, since it is highly unlikely that they will change the name |
| 16:02.27 | prpplague | HA - http://www.reason.com/images/07cf533ddb1d06350cf1ddb5942ef5ad.jpg |
| 16:07.58 | prpplague | gonna have to print some of those up |
| 16:09.38 | prpplague | although they left out the section about acquiring the LPR status when married to a US citizen |
| 16:12.07 | prpplague | i guess we were luckly only took us 3years and 8 months |
| 16:33.30 | *** join/#edev magnetron (n=magnetro@unaffiliated/magnetron) |
| 16:38.14 | prpplague | solar: looks like i have a match on the processor for the didj |
| 16:52.25 | MothaWork | prpplague: that chart does mention that "spouse[s] ... generally receive green cards" |
| 16:52.33 | MothaWork | though it doesn't imply how difficult that may be |
| 16:59.31 | ohsix | that staging stuff is cool, should speed stuff up and give people who want to contribute something easy to start with :] |
| 16:59.43 | *** join/#edev bbradley (n=bbradley@78-105-167-16.zone3.bethere.co.uk) |
| 16:59.58 | ohsix | be it cleaning up an existing start, or starting their own |
| 17:06.06 | *** join/#edev Redhatter (n=vk4msl@2001:388:f000:0:0:0:0:279) |
| 17:09.14 | prpplague | MothaWork: yea |
| 17:09.26 | prpplague | MothaWork: leaves out all the requirements and such |
| 17:12.45 | solar | thought you already figured it out to being some underclocked arm. |
| 17:12.53 | solar | has that changed in the past day or two? |
| 17:15.38 | *** join/#edev lyakh (n=lyakh@p57BD252F.dip0.t-ipconnect.de) |
| 17:16.59 | prpplague | solar: nope, just annoucing it publically |
| 17:17.35 | prpplague | http://www.mesdigital.com/english/Products/product_pollux.asp |
| 17:17.38 | solar | this one right http://www.linuxdevices.com/news/NS4648734230.html |
| 17:17.41 | solar | erp yah |
| 17:17.46 | prpplague | yea |
| 17:19.04 | prpplague | solar: i registered on their website and downloaded a datasheet that has the pinouts |
| 17:31.01 | *** join/#edev EI5GTB__ (n=Paul@213-202-165-31.bas503.dsl.esat.net) |
| 17:33.37 | solar | Anything interesting from the SFLC yet? |
| 17:34.59 | prpplague | solar: landley said they have it in the que, but not expecting anything soon |
| 17:35.24 | prpplague | solar: my unnamed source says that leapfrog is intentionally dragging their feet on post source |
| 17:36.46 | prpplague | man o man, the flyswatters are selling |
| 17:37.05 | prpplague | 20 to 1 for flyswatters to hammers |
| 17:37.23 | solar | cuz of the beagle boards? |
| 17:37.57 | solar | how are the http://tincantools.com/product.php?productid=16144&cat=0&page=1&featured selling? |
| 17:39.41 | prpplague | solar: only about 1/2 of the flyswatter sales are going out with beagle board adapters |
| 17:39.58 | prpplague | solar: so i have no clue what people are using them for |
| 17:40.27 | prpplague | solar: we started on a shield board for the beagle last week |
| 17:41.10 | solar | neat |
| 17:41.35 | prpplague | solar: pretty disappointed about hammer sells though |
| 17:41.38 | chouimat|work | setenv bootcmd 'setenv bootargs $(bootargs) root=/dev/sda7; bootm 0xf8100000' |
| 17:41.44 | chouimat|work | oops wrong window |
| 17:42.32 | MothaWork | at least you didn't attempt to log in as root |
| 17:43.27 | prpplague | solar: i suspect when the current inventory is gone, we probably won't do another run, or if we do, it will be much smaller |
| 17:43.39 | chouimat|work | MothaWork: true |
| 17:50.04 | prpplague | solar: i suspect that it didn't hit the market in the right place, the beagle is what more software people want, the arduino stuff is more for the hobbists, the gumstix is more for the proof-of-concept folks, and the buglabs stuff is on the horizon |
| 17:51.53 | prpplague | solar: btw, looks we've stabalized the hardware for the 8050, should have some units ready to ship soon to you |
| 17:54.54 | Crofton | people are funny, I like to DIP nature of the hammer |
| 17:55.18 | Crofton | but I guess people are just used to weird crap now |
| 17:56.31 | prpplague | Crofton: yea, what gets me, people are always asking for step-by-step instructions on how to do stuff |
| 17:57.18 | Crofton | people like that stuff |
| 17:58.48 | prpplague | yea seems so |
| 17:59.39 | prpplague | if i had it to do over again, i wouldn't have had the beta testers use the wiki, but write specific howtos |
| 18:03.19 | wmat | but howto's can be on a wiki |
| 18:03.56 | wmat | you're right though, people always want step-by-step instructions |
| 18:04.04 | prpplague | wmat: yea, but it seems alot of people want a pdf with very detailed step-by-step info |
| 18:04.17 | wmat | with pictures too, i bet |
| 18:04.21 | prpplague | wmat: yea |
| 18:04.45 | prpplague | wmat: i'm too jaded with the type of docs i normally get |
| 18:04.48 | wmat | prpplague: i can likely add a PDF exporter to mediawiki if there isn't already one |
| 18:05.17 | prpplague | wmat: yea, the howtos on the wiki just aren't detailed enough for alot of people |
| 18:05.41 | prpplague | wmat: funny thing, we get alot of people who don't even know what a wiki is or how to use it |
| 18:06.23 | wmat | they don't have to to use the content though |
| 18:06.33 | wmat | i use wikipedia all the time, but rarely contribute |
| 18:07.45 | prpplague | there are alot of people like landley that simple don't like the informational structure of wiki's |
| 18:08.30 | wmat | also, on the left of the wiki, there's a Printable Version link |
| 18:09.09 | wmat | the "informational structure", wtf does that mean? |
| 18:09.57 | wmat | in my experience that's usually a persons way of saying they'd rather not give up control of the content |
| 18:10.39 | wmat | layout could improve on the elinux wiki though, I admit that |
| 18:11.08 | wmat | but then again, everyone is free to change their template and/or use a mediawiki template to their liking |
| 18:15.03 | prpplague | http://www.robertdyas.co.uk/webapp/wcs/stores/servlet/ProductDisplay?storeId=91&mpe_id=11206&jspStoreDir=rc01&intv_id=10601&partNumber=P_145550&evtype=CpgnClick&langId=-1&catalogId=311&ddkey=ClickInfo |
| 18:15.06 | prpplague | i want one |
| 18:15.31 | prpplague | wmat: mainly it means that he has trouble locating information |
| 18:16.05 | prpplague | wmat: most of the people who complain about the wiki are those who prefer the structure of the linux howtos |
| 18:16.42 | wmat | can't please evryone I guess |
| 18:18.19 | prpplague | wmat: yep |
| 18:18.54 | prpplague | i still haven't figured out what the appeal of the beagle board is |
| 18:19.40 | wmat | prpplague: http://www.engadget.com/2008/09/25/windows-ce-powered-cuol-book-gives-netbooks-a-bad-name/ |
| 18:21.01 | prpplague | wmat: yea, but they are comparing it to a x86 UMPC |
| 18:21.11 | prpplague | wmat: i'd gladly pay $289 for it |
| 18:21.51 | MothaWork | prpplague: somebody here wants to use the beagle board "as is" in a product |
| 18:21.56 | MothaWork | i.e. mount it in a case |
| 18:22.23 | prpplague | MonMotha: wow |
| 18:22.29 | mnemoc | wants to get a gdium |
| 18:22.38 | prpplague | mnemoc: gdium? |
| 18:23.01 | prpplague | MonMotha: the beagle has i/o on all 4 sides |
| 18:23.07 | MothaWork | prpplague: I'm aware |
| 18:23.11 | wmat | prpplague: I guess the appeal is lowcost + expandability |
| 18:23.24 | mnemoc | prpplague: loongson 2F, 10" |
| 18:23.29 | prpplague | wmat: thats just it, it isn't that expandable |
| 18:23.47 | prpplague | mnemoc: sorry, i still don't know what you are refering to |
| 18:23.57 | mnemoc | prpplague: http://www.gdium.com/ |
| 18:24.02 | prpplague | looks |
| 18:24.05 | mnemoc | prpplague: a MIP64 netbook |
| 18:24.10 | wmat | prpplague: "the expandability of today's desktop machines..." |
| 18:24.48 | prpplague | mnemoc: ahh |
| 18:25.14 | mnemoc | prpplague: there is something about having a MIPS desktop that makes me desire it |
| 18:25.14 | prpplague | wmat: that a quote from the website? |
| 18:25.29 | wmat | yeah |
| 18:25.35 | wmat | elinux.org to be exact |
| 18:25.40 | prpplague | wmat: about the only thing currently you can expand with is usb devices |
| 18:27.04 | MothaWork | prpplague: and that's different from many desktop machines how? |
| 18:27.32 | prpplague | MothaWork: its not, but how is that any different than the hammer? |
| 18:27.40 | prpplague | MonMotha: the hammer has 2 usb hosts available |
| 18:27.50 | prpplague | MonMotha: along with a ton of other hardware items |
| 18:29.42 | prpplague | MonMotha: the only thing i can latch onto is that the beagle appeals to hardware-phobic developers |
| 18:30.16 | mnemoc | prpplague: is the hammer (not the samsung inside it) only been used for prototyping and developing or also on mass production devices? |
| 18:31.14 | prpplague | mnemoc: several ipaqs, several phones including the openmoko phones, a number of umpc devices |
| 18:31.42 | mnemoc | oh, cool :) |
| 18:31.44 | ds2 | the hammer is way slower |
| 18:31.46 | prpplague | mnemoc: and a wide range of gps navigation devices |
| 18:32.17 | MothaWork | prpplague: beagle appears to have some nasty errata |
| 18:32.18 | prpplague | ds2: yea, there is a big difference in performance |
| 18:32.34 | prpplague | MonMotha: indeed, they are still having problems with the host controller |
| 18:32.49 | wmat | mnemoc: where can the gnium be bought? |
| 18:33.49 | mnemoc | wmat: nowhere yet... it's supposed to be launched at the end of december in france |
| 18:33.52 | ds2 | also a Cortex-A8 has a lot more mystique then a lowly ARMv4 |
| 18:34.03 | mnemoc | wmat: http://www.gdium.com/2008/09/15/summary-information-from-your-comments-part-1/ |
| 18:34.05 | ds2 | even though most people don't need the difference |
| 18:34.14 | wmat | mnemoc: thx, wtf is with France? |
| 18:34.20 | ds2 | then there is the video out options not directly available on the hammer |
| 18:34.31 | mnemoc | wmat: EMTEC is french :-\ |
| 18:34.36 | wmat | ah |
| 18:35.19 | mnemoc | wmat: then they are supposed to add distributors in the rest of Europe (i think most probably germany and england) |
| 18:35.49 | prpplague | ds2: so the video stuff is a big factor? |
| 18:36.00 | MothaWork | prpplague: also seems to be less documented than I'd like in a dev board |
| 18:36.23 | prpplague | MothaWork: which item? the hammer or the beagle? |
| 18:36.29 | MothaWork | beagle |
| 18:36.34 | prpplague | ahh |
| 18:36.41 | ds2 | prpplague: from all the stuff I hear, yes |
| 18:36.52 | ds2 | esp. the 3D stuff... people want to do media (movie) players |
| 18:37.00 | mnemoc | wmat: the oddest part is that it was designed by the australian branch of EMTEC, but they don't even have plans to sell it there at short term |
| 18:37.14 | MothaWork | ds2: that would be the major market for the OMAP w/ DSP |
| 18:37.50 | ds2 | MothaWork: for the OMAP yes, but for the beagleboard, the DSP isn't really ready for the hobbiest portion |
| 18:38.02 | MothaWork | ds2: I gathered, but it is there |
| 18:38.02 | ds2 | bbl, lunch |
| 18:38.08 | mnemoc | prpplague: somehow I had the idea that the hammer (DIP 40pin) was to big for those devices |
| 18:38.14 | MothaWork | something about having a bunch of hardware that's unusable due to software always irks me |
| 18:38.20 | ds2 | MothaWork: yes, it is at its infancy |
| 18:38.33 | prpplague | mnemoc: i'm sorry i did not understand your statement |
| 18:39.06 | prpplague | mnemoc: oh i misunderstood your earlier statement |
| 18:39.16 | prpplague | mnemoc: no, the hammer is mainly used for proof-of-concept stuff |
| 18:39.33 | prpplague | mnemoc: the samsung stuff is used in alot of devices, but not the hammer itself |
| 18:40.07 | mnemoc | prpplague: that was exactly my first thought :) |
| 18:40.20 | prpplague | mnemoc: most people purchase the hammer to do a test design in the lab, since it is all .1" |
| 18:40.38 | prpplague | mnemoc: once they have a working design, they go ahead with their own boards |
| 18:42.21 | prpplague | http://hackaday.com/2008/09/25/arduino-solenoid-concert/ |
| 18:42.47 | mnemoc | prpplague: I was wondering when you asked what is it about the beagle |
| 18:42.52 | prpplague | you see people doing all kinds of stuff with arduino, for two reasons, it is cheap, and they can program it easily via the wiring/process language |
| 18:43.25 | prpplague | mnemoc: yea, the beagle is selling very quickly |
| 18:43.26 | mnemoc | prpplague: and as I see it, the hammer has a professional target but the beagle is fully targeted for hobist |
| 18:43.55 | prpplague | mnemoc: yea, the hammer was targeted at hobbists |
| 18:44.38 | mnemoc | i was trying to say the oposite :p (my english stinks) |
| 18:45.26 | prpplague | mnemoc: yes, which is my point |
| 18:45.39 | prpplague | mnemoc: more professionals are using the board, when it was intended for hobbists |
| 18:45.48 | mnemoc | ah, ic |
| 18:51.58 | mnemoc | prpplague: btw, did you have a minute for my parcel? :) |
| 18:52.48 | prpplague | mnemoc: going to the post office today after work |
| 18:53.31 | mnemoc | looks at his clock and remembers prpplague is from USA |
| 18:54.17 | mnemoc | prpplague: thanks! :) btw, where in the US are you located? |
| 18:54.30 | prpplague | mnemoc: dallas,texas area |
| 18:55.24 | mnemoc | uh, 7h behind me then |
| 18:55.55 | prpplague | yea, just 2pm here |
| 18:58.23 | *** join/#edev purl (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 18:58.23 | *** topic/#edev is a channel for the discussion of hacking small consumer electronic devices, mostly ARM based; Hammer Development boards as well as Nail Boards are now available at http://www.tincantools.com/ |
| 19:05.52 | g1powermac | holy mother of pearl |
| 19:06.03 | g1powermac | Jack Thompson has been disbarred |
| 19:06.12 | g1powermac | http://yro.slashdot.org/yro/08/09/25/1822207.shtml |
| 19:07.05 | mnemoc | looks for disbar at his dictionary |
| 19:09.52 | g1powermac | oh dear no |
| 19:09.57 | g1powermac | this ain't good |
| 19:09.57 | g1powermac | http://yro.slashdot.org/yro/08/09/25/178253.shtml |
| 19:10.12 | g1powermac | shakes his head at the shear stupidity of the state I live in |
| 19:21.21 | prpplague | ho ho hum |
| 19:24.51 | CosmicPenguin | Wait, wasn't Jack Thompson disbarred a while ago? |
| 19:24.58 | prpplague | CosmicPenguin: on appeal |
| 19:25.02 | CosmicPenguin | ahh |
| 19:25.29 | prpplague | CosmicPenguin: the original judge said for 10 years, the appellet judge said for life |
| 19:25.35 | prpplague | CosmicPenguin: guess he shouldn't have appealed |
| 19:25.45 | CosmicPenguin | don't piss of the judge is the moral of that story |
| 19:26.04 | CosmicPenguin | and the lawyer who represents himself has a fool for a client |
| 19:26.07 | CosmicPenguin | and so on and soforth |
| 19:26.55 | *** join/#edev dijenerate (n=dijenera@65.48.164.209) |
| 19:39.58 | *** join/#edev Ian_Daniher (n=it@204.10.219.119) |
| 20:14.51 | *** join/#edev landley (n=landley@static-71-162-243-5.phlapa.fios.verizon.net) |
| 20:14.51 | *** mode/#edev [+o landley] by ChanServ |
| 20:21.04 | prpplague | landley: greetings |
| 20:22.16 | MothaWork | prpplague: do you know if ARM7TDMI has anything like a cycle counter? |
| 20:22.20 | landley | Morning. |
| 20:22.28 | MothaWork | basically, a really fast running counter I can access quickly |
| 20:22.37 | MothaWork | oh, and from within an interrupt context |
| 20:23.05 | prpplague | MothaWork: that is going to vary from cpu to cpu |
| 20:23.33 | prpplague | MothaWork: you just needing to count time? or something else? |
| 20:23.38 | MothaWork | so the core doesn't feature anything like that |
| 20:23.44 | MothaWork | prpplague: well, I'm looking to timestamp an interrupt, basically |
| 20:24.02 | MothaWork | the timestamp can be meaningless, it just needs to be sufficiently not correlated to other interrupts |
| 20:24.15 | prpplague | MothaWork: the jiffies value isn't enough? |
| 20:24.17 | MothaWork | so if I have a comm interrupt, I want effectively a random timestamp on that |
| 20:24.26 | MothaWork | prpplague: this isn't on Linux (ARM7TDMI) |
| 20:24.34 | prpplague | MothaWork: oh, silly me |
| 20:24.45 | prpplague | MothaWork: baremetal eh? |
| 20:24.46 | MothaWork | I'm running FreeRTOS, but it only gives you a millisecond counter, and that's not high enough res |
| 20:24.55 | MothaWork | baremetal ~ FreeRTOS, yes :) |
| 20:24.56 | prpplague | MonMotha: ahh |
| 20:25.02 | MothaWork | basically, I need to build up a random number generator |
| 20:25.08 | MothaWork | and I need something to feed the entropy pool |
| 20:25.26 | MothaWork | taking an example from Linux, the "exact" arrival time of most interrupts is effectively random as far as the CPU is concerned |
| 20:25.56 | MothaWork | software people have requested a hardware means of generating random entropy. That seems excessive |
| 20:26.16 | prpplague | hmm, interesting problem |
| 20:27.07 | prpplague | MothaWork: the device have a pwm onboard? |
| 20:27.22 | MothaWork | prpplague: it does |
| 20:27.30 | MothaWork | one of my options was to measure jitter between two oscillators |
| 20:27.45 | prpplague | MonMotha: you could set up a pwm with autorestart |
| 20:27.55 | MothaWork | I have a clock for an RF module that I can hook up to a timer input and measure it relative to the CPU clock |
| 20:28.06 | MothaWork | it'll jitter and give you a bit or two every few seconds |
| 20:28.14 | prpplague | MonMotha: combie that with the millisecond counter |
| 20:28.15 | MothaWork | prpplague: yeah, I'm not sure if I can get that fast enough |
| 20:28.38 | MothaWork | basically, I need the least significant few bits to be essentially random, and for interrupts that are mostly periodic, that requires a very fast counter |
| 20:28.40 | prpplague | MonMotha: don't need it to be fast, just clocking and use the other counter together |
| 20:28.59 | *** join/#edev JoeLlama (n=snork@unaffiliated/joellama) |
| 20:29.53 | prpplague | MonMotha: hmm, i'd look at the capabilities of the pwm |
| 20:30.16 | prpplague | MonMotha: user the millisecond counter plus a cycle value of the pwm |
| 20:30.36 | MothaWork | prpplague: well, say the interrupt is not appreciably slower than the rate at which my PWM counter increases |
| 20:30.48 | MothaWork | then the bits aren't random, they're more like somewhat random positive increments |
| 20:31.14 | MothaWork | or, even worse, if the interrupt is actually periodic, then I have to have things basically just measure the jitter in the rate at which the interrupt is serviced |
| 20:31.55 | prpplague | MothaWork: sorry i was going on the assumption you wanted to start with a timestamp |
| 20:32.41 | MothaWork | prpplague: well, that was just one idea |
| 20:32.58 | prpplague | MothaWork: if your current counter can do milliseconds, i'd see how small you could do the pwm to create a much smaller cycle |
| 20:33.14 | MothaWork | if the timestamp increments at a rate that is very fast compared to the interrupt rate, then the least significant few bits are effectively random |
| 20:33.42 | prpplague | MothaWork: i.e. if you could set the pwm to .1 milliseconds |
| 20:34.12 | MothaWork | the interesting thing is that even if the interrupt is perfectly periodic and does not jitter relative to MCU clock, it could still generate random bits since the interrupt latency will vary due to e.g. critical sections |
| 20:34.29 | prpplague | yea |
| 20:35.11 | prpplague | doesn't parallax sell a small i2c device for random number generation? |
| 20:35.16 | MothaWork | they might |
| 20:35.22 | MothaWork | I'm trying to avoid adding hardware |
| 20:35.27 | MothaWork | seems like an unnecessary expense |
| 20:35.54 | prpplague | yea |
| 20:35.58 | prpplague | i can understand that |
| 20:36.57 | prpplague | MothaWork: what is the device clocked at? |
| 20:43.03 | MothaWork | some strange clock in the 48-50MHz range |
| 20:43.46 | MothaWork | looks like I can clock the PWM system at MCK, which means this may be doable |
| 20:43.54 | MothaWork | I'd be curious to see how much entropy I can get |
| 20:44.51 | prpplague | MonMotha: you have any adc's that aren't being used? |
| 20:45.00 | MothaWork | not that aren't hooked up to anything |
| 20:45.47 | prpplague | MonMotha: might try adding in the values of the adc's |
| 20:45.54 | prpplague | MonMotha: how many bits? |
| 20:45.57 | MothaWork | yes |
| 20:46.10 | MothaWork | they are 10 bits, but we are successfully supersampling them to 18 (!!) |
| 20:46.16 | prpplague | lovely |
| 20:46.29 | MothaWork | we can reliably detect thermal effects that don't actually result in a full 1LSB change in analog value at the pin |
| 20:46.31 | prpplague | so should have plenty to seed your values then |
| 20:47.03 | prpplague | MonMotha: interesting |
| 20:47.41 | MothaWork | our LCD backlight is run off +12V through a couple power resistors on the other side of the board (non-ideal, yes). Turning on the backlight causes a small change in A/D readings |
| 20:48.05 | MothaWork | it's less than 1LSB at the 10-bit precision of the raw A/D, but it shows up |
| 20:48.33 | prpplague | interesting |
| 20:49.14 | prpplague | MonMotha: so use the milisecond counter, plus the pwm cycle with autorestart, plus one or more of the adc values, and you should be plenty good |
| 20:49.40 | prpplague | MonMotha: worst case, the counters could be replicated, but the adc values isn't going to be predictable |
| 20:50.33 | MothaWork | well, this is all going to get hashed to sunday and back anyway |
| 20:50.41 | MothaWork | this is just the entropy input into the RNG |
| 20:51.10 | MothaWork | so I need 2-4x the number of bits in as come out for high quality stuff, and about 1/128th the number of bits in as go out for things like nonces and other non key material |
| 20:51.59 | prpplague | MothaWork: ahh |
| 21:03.19 | *** join/#edev roxfan (n=dunno@209.211-200-80.adsl-dyn.isp.belgacom.be) |
| 21:15.10 | prpplague | mnemoc: all boxed up and ready |
| 21:16.33 | mnemoc | prpplague: great! I can't wait to see that nail =) |
| 21:17.18 | *** join/#edev RobotGuy (i=d047c850@gateway/web/ajax/mibbit.com/x-806fe8f1c0e7b16f) |
| 21:24.18 | *** join/#edev Ian_Daniher (n=it@69.61.230.246) |
| 21:37.37 | prpplague | anyone used the verizon cdma usb dongle under linux? |
| 21:37.51 | prpplague | landley: didn't you get a wireless dongle recently? |
| 21:40.17 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 21:40.20 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 21:43.32 | landley | prpplague: I got a bluetooth dongle. |
| 21:43.37 | landley | Two actually, lost the first one. |
| 21:43.50 | landley | So I can use my cell phone to get internet access on my laptop. |
| 21:43.59 | landley | The built in bluetooth doesn't work. It's a Dell. |
| 21:44.00 | prpplague | landley: ahh |
| 21:46.30 | ds2 | do dells ever work? ;) |
| 21:50.16 | *** join/#edev kergoth (n=kergoth@ip72-201-11-30.ph.ph.cox.net) |
| 21:51.09 | kergoth | morning |
| 21:52.18 | *** part/#edev kergoth (n=kergoth@ip72-201-11-30.ph.ph.cox.net) |
| 21:52.41 | *** join/#edev kergoth (n=kergoth@ip72-201-11-30.ph.ph.cox.net) |
| 21:54.45 | *** join/#edev redblue (i=star@ppp199.108-253-207.mtl.mt.videotron.ca) |
| 22:18.50 | solar | only the laptops. |
| 22:32.02 | RobotGuy | solar: I'm not going to get new internet at home - cancelling the Monday installation. |
| 22:52.30 | solar | umm ok |
| 23:14.37 | *** join/#edev lifeeth (n=Praneeth@unaffiliated/lifeeth) |
| 23:26.19 | *** join/#edev kgilmer (n=kgilmer@dsl254-120-154.nyc1.dsl.speakeasy.net) |
| 23:45.35 | *** join/#edev dijenerate (n=dijenera@65.48.164.209) |
