**** BEGIN LOGGING AT Tue Oct 10 02:59:57 2006
Oct 10 12:29:49 <zumbi> hello!
Oct 10 12:30:38 <vmaster> zumbi: hi
Oct 10 12:30:46 <zumbi> :-)
Oct 10 12:31:49 <zumbi> i'm planning to add daisy chain jtag to a circuit with many FPGAs -i'm new to daisy chain- so...
Oct 10 12:32:55 <zumbi> ... jtag is for programming purposes. The question i have is if you know if i have to do daisy chaining with eeproms or FPGAs, and if you know some example schematic
Oct 10 12:34:43 <vmaster> daisy chaining is rather easy - just connect TDI->TDO from one device to the next, and connect TCK and TMS (optionally TRST) to each device in parallel
Oct 10 12:35:16 <vmaster> you might run into problems with too many devices in the chain though
Oct 10 12:35:31 <zumbi> vmaster: but memories (EEPROM)? (there are 3 devices)
Oct 10 12:36:11 <vmaster> is there a jtag port on these eeproms?
Oct 10 12:36:45 <zumbi> vmaster: i guess, it's an eeprom for that FPGA
Oct 10 12:37:31 <vmaster> Xilinx?
Oct 10 12:38:12 <zumbi> yes
Oct 10 12:38:37 <zumbi> http://direct.xilinx.com/bvdocs/publications/ds123.pdf
Oct 10 12:38:50 <zumbi> those are programming memories
Oct 10 12:39:32 <zumbi> anyway, vmaster, thank you very much
Oct 10 12:41:57 <zumbi> vmaster: about openjtag, how is the project status ? do you need some volunteer help?
Oct 10 12:43:30 <vmaster> i don't think there's a lot happening here lately. someone started working on a ARM9+FPGA board (iirc) to allow high JTAG speeds, but haven't heard about this project for a while
Oct 10 12:45:18 <vmaster> regarding your fpgas: if you have JTAG available at each FPGA, you can configure them on the fly, without writing the EEPROM each time
Oct 10 12:46:11 <vmaster> the datasheet mentions problems with some devices that follow the standard close enough - some change TDO on the rising edge of TCK, and some on the falling edge
Oct 10 12:46:56 <vmaster> s/that follow/that don't follow/
Oct 10 12:46:57 <purl> vmaster meant: the datasheet mentions problems with some devices that don't follow the standard close enough - some change TDO on the rising edge of TCK, and some on the falling edge
Oct 10 12:47:27 <zumbi> i see
Oct 10 12:48:32 <zumbi> well, thanks
Oct 10 12:48:40 * zumbi reads openjtag webpage
Oct 10 12:49:04 <[g2]> vmaster I think things are warming up a little
Oct 10 12:49:48 <vmaster> [g2]: yeah, i saw AchiestDragon online, but I believe this is only temporary?
Oct 10 12:51:15 <[g2]> vmaster I'm talking about something else
Oct 10 12:51:27 <vmaster> [g2]: ah, ok :)
Oct 10 12:55:42 <[g2]> vmaster what kinda of jtag speeds do you see with the Olimex usb jtag device ?
Oct 10 12:56:26 <vmaster> the FT2232 achieves about 1.5-2 MHz effective JTAG clock while debugging an ARM7/9
Oct 10 12:56:48 <vmaster> that's good enough for downloading at ~120kb/s
Oct 10 12:57:02 <[g2]> that's pretty fast
Oct 10 12:57:20 <vmaster> fast enough for most purpose, imho
Oct 10 12:57:28 <vmaster> s/purpose/purposes/
Oct 10 12:57:29 <purl> vmaster meant: fast enough for most purposes, imho
Oct 10 12:57:52 <[g2]> yeah fast enough that it's not painful
Oct 10 13:06:28 <zumbi> have you used opencores' usb core for these purposes?
Oct 10 13:16:45 <wookey_> some of the balloon people have developed a fast jtag widget too:
Oct 10 13:16:46 <wookey_> http://balloonboard.org/~lwithers/urppd/
Oct 10 13:17:07 <wookey_> using cypress EZ-USB device
Oct 10 13:17:18 <wookey_> released under GPL2
Oct 10 17:06:39 <prpplague> vmaster: http://pastebin.ca/196499
Oct 10 17:17:39 <vmaster> hey prpplague 
Oct 10 17:18:32 <vmaster> could you run again with debugging enabled (-d)
Oct 10 17:18:34 <vmaster> ?
Oct 10 17:36:26 <prpplague> vmaster: you turn on the highdrive for your ft2232 based devices?
Oct 10 17:37:28 <vmaster> only for the one that doesn
Oct 10 17:37:32 <vmaster> 't have buffers
Oct 10 17:37:35 <prpplague> hmm
Oct 10 17:37:49 <prpplague> i'm having alot of problems with one of prototype boards
Oct 10 17:38:24 <prpplague> vmaster: its communicates but get a ton of errors
Oct 10 17:38:34 <vmaster> did you try lowering the speed?
Oct 10 17:38:37 <prpplague> vmaster: yea
Oct 10 17:38:42 <prpplague> vmaster: all the way down to 10
Oct 10 17:40:02 <vmaster> you removed R9/C9 on that board, too?
Oct 10 17:40:20 <prpplague> yea
Oct 10 17:41:13 <prpplague> vmaster: whats interesting is that my board worked fine with the leds, but these act like the leds are drawing too much current
Oct 10 17:41:37 <prpplague> vmaster: i've removed the leds for now, but it still seems like something isn't right
Oct 10 17:42:03 <prpplague> i'm wondering if dlpdesign changed their default firmware settings
Oct 10 17:42:30 <prpplague> vmaster: did you use the linux example code to turn on the high driver outputs?
Oct 10 17:42:45 <vmaster> yeah, the windoze stuff never worked for me
Oct 10 17:43:25 <prpplague> vmaster: i have a very hard time with ftdi docs, do i simply set the value in the eeprom to 0x01 to enable high drive?
Oct 10 17:44:00 <vmaster> the eeprom stuff is barely documented, iirc
Oct 10 17:44:40 <vmaster>     Data.AIsHighCurrent = 1;        // non-zero if interface is high current
Oct 10 17:44:58 <vmaster> that's what I changed in sample/EEPROM/write/main.c
Oct 10 17:47:14 <prpplague> vmaster: ok gotcha
Oct 10 17:48:44 <prpplague> vmaster: hmm, you have to write all the values back out?
Oct 10 17:49:15 <prpplague> vmaster: hmm, what is the max current you have on your devices?
Oct 10 17:56:34 <prpplague> vmaster: don't suppose you could give me a eeprom dump of one of your working devices?
Oct 10 18:02:44 <vmaster> prpplague: sorry, just had a doner kebap for lunch
Oct 10 18:03:19 <vmaster> prpplague: i'll send you a dump and the .c file i've used to program it
Oct 10 18:03:43 <prpplague> vmaster: thanks
Oct 10 18:08:31 <vmaster> prpplague: mmd.ath.cx/ft2232_highdrive.c
Oct 10 18:08:57 <vmaster> dumping the eeprom is only possible with libftdi, right?
Oct 10 18:09:21 <prpplague> vmaster: seems to be possible with the d2xx
Oct 10 18:09:37 <prpplague> vmaster: under samples there is a read function
Oct 10 18:24:01 <vmaster> http://mmd.ath.cx/ft2232_highdrive.c
Oct 10 18:25:54 <vmaster> http://mmd.ath.cx/EEPROM.dump
Oct 10 18:26:14 <vmaster> the output from read/main.c is ASCII, if you'd like, I could modify it so it dumps to a file
Oct 10 18:55:47 <prpplague> vmaster: thanks
Oct 10 18:56:00 <prpplague> vmaster: think we found the problem with one of the jtag lines on the main board
Oct 10 18:56:07 <prpplague> vmaster: its very noisy
Oct 10 19:15:33 <vmaster> prpplague: could you figure out why?
Oct 10 19:16:21 <prpplague> vmaster: looks to be a board layout issue
**** ENDING LOGGING AT Wed Oct 11 02:59:56 2006