**** BEGIN LOGGING AT Thu Aug 31 02:59:57 2006
Aug 31 14:23:12 <[g2]> vmaster ping 
Aug 31 14:23:52 * [g2] is wondering if you have any thoughts on LPC21xx or SAM micros
Aug 31 14:25:07 <vmaster> hey g2
Aug 31 14:25:18 <[g2]> hey vmaster 
Aug 31 14:25:36 <vmaster> anything in special, or just general?
Aug 31 14:25:41 <[g2]> I'm probably going to be getting some little ARM chips
Aug 31 14:26:01 <[g2]> I need serial and an SD card
Aug 31 14:26:13 <[g2]> actually two serial ports would be nice
Aug 31 14:26:17 <vmaster> heh, that's something they all have
Aug 31 14:26:32 <vmaster> maybe one of the smallest of both families lack the second uart
Aug 31 14:27:13 <[g2]> I don't need the smallest I don't think the cost is much of an issue as these things are <$10
Aug 31 14:27:50 <[g2]> I a little interested in tools (compilers/JTAG)  I'll be using openocd :)
Aug 31 14:27:52 <vmaster> they all have SPI interfaces, so interfacing to SD shouldn't be a problem
Aug 31 14:28:22 <[g2]> so any chips you really like ?
Aug 31 14:28:43 <[g2]> or any to say away from ?
Aug 31 14:28:56 <vmaster> not really, they're mostly the same
Aug 31 14:29:06 <vmaster> the LPCs can run a bit faster, especially the newer ones
Aug 31 14:29:11 <vmaster> up to 70MHz operation
Aug 31 14:29:21 <vmaster> the SAM7X has an integrated ethernet controller
Aug 31 14:29:33 <vmaster> the SAM7 have USB device ports
Aug 31 14:29:41 <vmaster> the LPC214x has USB, too
Aug 31 14:30:09 <vmaster> running out of flash, the SAM7 are slower
Aug 31 14:30:29 <vmaster> you can get cheap devboards for both families from sparkfun
Aug 31 14:30:43 <vmaster> the OpenOCD supports both families, including programming the on-chip flash
Aug 31 14:31:02 <vmaster> the LPC22xx have an external address bus
Aug 31 14:32:14 <vmaster> the memory accelerator module (MAM) of the LPCs allows them to work almost at full speed from flash, only very jumpy code is penalized
Aug 31 14:32:44 <vmaster> the LPC2000 group at yahoogroups is very active, you'll get immediate help for almost every problem, and Philips monitors the group
Aug 31 14:33:42 <vmaster> both families are in-system programmable via UART, the SAM7 also via USB, iirc
Aug 31 14:33:48 <p2-mate> MAM is a fancy new name for cache ?
Aug 31 14:34:00 <vmaster> a very simple form of cache
Aug 31 14:34:08 <vmaster> the flash is internally 128-bit wide
Aug 31 14:34:19 <vmaster> and the MAM is made of two 128-bit latches
Aug 31 14:34:45 <vmaster> so even if a flash access takes two or three cycles, the MAM allows the core to continue full-speed
Aug 31 14:35:27 <p2-mate> as long as it correctly predicts the next 128bit access
Aug 31 14:35:32 <vmaster> yep
Aug 31 14:36:07 <vmaster> the general opinion in the lpc2000 seems to be that the MAM works reasonably well for real world code
Aug 31 14:36:18 <vmaster> s/lpc2000/lpc2000-group/
Aug 31 14:36:18 <purl> vmaster meant: the general opinion in the lpc2000-group seems to be that the MAM works reasonably well for real world code
Aug 31 14:36:25 <p2-mate> which isn't too easy as it's not part of the CPU core (think of function returns using blr)
Aug 31 14:37:00 <p2-mate> s/blr/bl
Aug 31 14:37:21 <[g2]> sounds like the LPC2K family is a winner
Aug 31 14:37:37 <vmaster> heh, that's probably because i used the more than the sam7
Aug 31 14:37:43 <vmaster> str7 seems to be competitive, too
Aug 31 14:38:12 <[g2]> have you used the sparkfun boards ?
Aug 31 14:39:15 <vmaster> i have a LPC2294 header board, a SAM7H64 header board and the SAM7EX dev board
Aug 31 14:53:11 <[g2]> vmaster do you use gcc as a compiler ?
Aug 31 14:53:24 <[g2]> gcc/binutils/gdb etc..
Aug 31 15:01:23 <vmaster> yeah, sure
Aug 31 15:02:12 <vmaster> i use the toolchain from codesourcery, since gnuarm stopped providing ia-32 linux binaries
Aug 31 15:02:20 <vmaster> or the ELDK from Denx
Aug 31 15:02:46 <vmaster> people are using the OpenOCD with Eclipse, but I prefer commandline gdb
Aug 31 15:05:31 <prpplague> vmaster: you like the lpc and sam series?
Aug 31 15:06:17 <vmaster> mhh, yeah, they're nice and cheap
Aug 31 15:07:43 <prpplague> yea, just wish they had a little bigger flash/ram combo
Aug 31 15:07:58 <prpplague> thats one reason we are doing the arm9 device in a 40-pin dip package
Aug 31 15:08:09 <vmaster> mhh, they're about to get bigger variants out
Aug 31 15:08:18 <vmaster> LPC2880 for example
Aug 31 15:08:51 <prpplague> still not big enough to run uclinux effectively
Aug 31 15:09:36 <vmaster> mhh, no, but they're powerfull replacements for all the 8/16 bitters out there
Aug 31 15:09:46 <prpplague> oh yea
Aug 31 15:10:14 <vmaster> uclinux is going to need a device with an external bus, isn't it?
Aug 31 15:10:22 <vmaster> even a meg of flash is too little
Aug 31 15:10:35 <vmaster> and you'll need several megs of RAM
Aug 31 15:10:52 <prpplague> vmaster: you can do it with 1mb ram and 1mb flash
Aug 31 15:11:03 <prpplague> its tight but it's doable
Aug 31 15:11:15 <vmaster> heh, yeah, but that's academic, imho
Aug 31 15:11:23 <prpplague> vmaster: yea
Aug 31 15:11:35 <prpplague> vmaster: effective use would be 1mb flash and 2mb ram
Aug 31 15:57:17 <key2> is there only one kind of AMR JTAG connector ?
Aug 31 15:57:30 <key2> or each constructor would use its own pinout
Aug 31 16:05:53 <prpplague> key2: there is an official document from IEEE that describes "standard" pinouts
Aug 31 16:06:15 <prpplague> key2: however, most manufacturers do not implement the "standard" pinouts due to space restrictions
Aug 31 16:06:48 <prpplague> key2: most production device simple have test pads which are used with what they call a "needle bed" 
Aug 31 16:07:07 <vmaster> most ARM dev boards i've seen use the 20-pin ARM JTAG
Aug 31 16:07:38 <prpplague> vmaster: yea?
Aug 31 16:07:57 <prpplague> vmaster: i don't own a single arm dev board that has a 2x10 header
Aug 31 16:08:03 <prpplague> vmaster: everyone i have is 2x7
Aug 31 16:08:13 <prpplague> vmaster: (that has a header)
Aug 31 16:08:24 <key2> prpplague: I have a 2x5 connector
Aug 31 16:08:41 <prpplague> key2: arm?
Aug 31 16:08:42 <key2> prpplague: and since the chip its connected to is BGA, i won't really be able to figure out what is what
Aug 31 16:08:45 <key2> prpplague: yeah
Aug 31 16:08:53 <key2> prpplague: the conexant cx82100
Aug 31 16:08:59 <prpplague> key2: thats a non-standard connector but is common
Aug 31 16:09:52 <key2> prpplague: well when I look at the connector, it has the Gnd on two pins one in front of the otherone
Aug 31 16:10:02 <key2> and in none of the 2x5 jtag pinout i've seen, i could see that
Aug 31 16:10:03 <key2> ..
Aug 31 16:10:19 <vmaster> the 2x7 is the old ARM standard
Aug 31 16:10:29 <prpplague> vmaster: yea
Aug 31 16:10:31 <vmaster> they've moved to the 2x10 to have a GND for every signal
Aug 31 16:10:42 <prpplague> vmaster: the 2x10 with the offset gnd lines is much better
Aug 31 16:10:53 <prpplague> vmaster: just the extra pins makes it a pain
Aug 31 16:11:18 <prpplague> key2: yea thats probably just for cross-talk protection
Aug 31 16:11:31 <prpplague> key2: this a dev board?
Aug 31 16:11:37 <key2> prpplague: no
Aug 31 16:11:45 <prpplague> ahh
Aug 31 16:11:53 <prpplague> key2: fun fun
Aug 31 16:12:19 <prpplague> key2: usually what i have to do for hacking is sacrifice a board and depopulate it with a head gun
Aug 31 16:12:21 <prpplague> heat
Aug 31 16:12:43 <key2> prpplague: and then u look with what ?
Aug 31 16:13:02 <prpplague> key2: find the bga package listing in the datasheets
Aug 31 16:13:12 <prpplague> key2: locate the jtag pins on the bga package
Aug 31 16:13:23 <prpplague> key2: they tone it out with a multimeter
Aug 31 16:13:26 <prpplague> then
Aug 31 16:13:27 <prpplague> argh
Aug 31 16:13:30 <prpplague> i can't type today
Aug 31 16:13:44 <key2> prpplague: yeah I got it
Aug 31 16:13:52 <key2> but I'll have to sacrify one of the board :)
Aug 31 16:14:37 <prpplague> yea
Aug 31 16:14:44 <prpplague> price of hacking
Aug 31 16:14:55 <prpplague> key2: i usually get 3 of whatever i am hacking
Aug 31 16:15:01 <prpplague> key2: one to de populate
Aug 31 16:15:10 <prpplague> key2: one for dev platform
Aug 31 16:15:18 <prpplague> key2: one as a reference device
Aug 31 16:16:27 <key2> prpplague: I have an other pb here
Aug 31 16:16:32 <key2> it uses an ARM9TDMI
Aug 31 16:16:39 <key2> and it has its own MAC interface
Aug 31 16:16:48 <key2> and the current kernel is vxwork
Aug 31 16:17:02 <key2> so I dunno how am gonna manage to get the eth working since I don't have those drivers
Aug 31 16:17:24 <prpplague> key2: this a hobby hack?
Aug 31 16:17:32 <key2> yeah
Aug 31 16:17:48 <prpplague> ahh
Aug 31 16:17:52 <prpplague> key2: what device?
Aug 31 16:18:57 <key2> bewane
Aug 31 16:20:38 * prpplague googles
Aug 31 16:21:34 <prpplague> key2: hmm nothing poping up on google
Aug 31 16:33:17 <key2> prpplague: its a bewan ATA
Aug 31 16:34:41 <key2> http://www.bewan.com/bewan/products/voip/bphonebox.php
Aug 31 16:35:37 <prpplague> ahh
Aug 31 16:35:39 <prpplague> lots of fun there
Aug 31 16:36:25 <key2> prpplague: it's a cx82100
Aug 31 16:36:32 <key2> i have the datasheet
Aug 31 16:38:31 <key2> 'http://www.tranzistoare.ro/datasheets/2300/500021_DS.pdf
Aug 31 18:12:57 <key2> someone knows what's the .DLF file format ?
Aug 31 18:38:33 <prpplague> key2: http://filext.com/detaillist.php?extdetail=DLF
Aug 31 19:08:46 <key2> prpplague: I get a vxWorks.z but the file is not a correct archive
Aug 31 19:08:52 <key2> it doesnt have the header
Aug 31 19:08:58 <key2> have an idea of how I could uncompress it ?
Aug 31 19:09:24 <prpplague> sorry now
Aug 31 19:09:28 <prpplague> no
**** ENDING LOGGING AT Fri Sep 01 02:59:56 2006