| 01:35.15 | *** join/#openjtag VoodooZ (~stephane6@CPE00c0f021712c-CM014110003785.cpe.net.cable.rogers.com) |
| 01:36.42 | VoodooZ | [g2], ping |
| 02:02.05 | [g2] | VoodooZ, pong |
| 02:04.40 | VoodooZ | Hey, |
| 02:04.52 | VoodooZ | Are you guys using USB on your JTAG project? |
| 02:05.33 | [g2] | ep1220 is |
| 02:06.07 | VoodooZ | I"m asking because I apparently managed to burn my I2C bus AGAIN! I have no clue what happened but the RTC is not responding anymore. |
| 02:06.48 | VoodooZ | So I'm considering building a USB hub using an AVR that will extend I/O out and read other sensors so my slug can read it all in one big chunk via USB. |
| 02:07.42 | Tiersten | Have you reset the RTC? |
| 02:08.19 | VoodooZ | Irregardless. I rebooted and all and it still doesn't work with no other i2c devices on the bus. |
| 02:08.40 | [g2] | bummer :( |
| 02:08.42 | VoodooZ | I was just curious to know how hard it would be to code for USB data xfer |
| 02:09.02 | VoodooZ | I don't feel like buying another one when this one still works fine. |
| 02:09.19 | VoodooZ | I don't care much about time. I'm just afraid linux might not like it though. ??? |
| 02:09.24 | Tiersten | It doesn't respond at all? |
| 02:10.06 | VoodooZ | nope. |
| 02:10.44 | VoodooZ | but that's ok as I always wanted to learn to program USB. |
| 02:11.10 | VoodooZ | I noticed quite a big frame rate drop when using the software I2C anyways so... |
| 02:11.56 | Tiersten | Linux will be fine with the lack of RTC |
| 02:12.05 | Tiersten | It will just start over at a fixed time each time you reboot |
| 02:12.11 | VoodooZ | good. |
| 02:12.41 | VoodooZ | I just hope it won't keep trying to access it until it times out as the bootup times is already long enough like that. :) |
| 02:13.08 | Tiersten | Remove the RTC driver if you want to stop that |
| 02:13.18 | VoodooZ | ok. |
| 02:13.37 | Tiersten | It will time out but it'll be faster just without it |
| 02:13.43 | VoodooZ | Should I have my robot code make USB calls or have the kernel do it. is this what libusb does? userspace? |
| 02:14.12 | Tiersten | Yeah. libusb is a userspace library to talk to USB devices |
| 02:14.29 | Tiersten | Is it actually 100% working now? I thought it was good enough for scanning now but there were still some underlying flaws? |
| 02:14.40 | VoodooZ | I heard bad things about it. |
| 02:15.14 | VoodooZ | so I guess I'd be better off writing an actual kernel driver module for my USB device and have my robot code mmap it right? |
| 02:15.23 | Tiersten | You can do what puppy does and directly access usbfs |
| 02:15.23 | dyoung | look at puppy. |
| 02:15.24 | Tiersten | That should work |
| 02:15.27 | dyoung | echo |
| 02:15.29 | dyoung | :-) |
| 02:15.31 | Tiersten | :) |
| 02:15.44 | Tiersten | mmap it? |
| 02:15.45 | VoodooZ | really? much simpler probably too right? |
| 02:15.48 | Tiersten | ew |
| 02:16.14 | VoodooZ | I'll be mostly requesting a block of data every iterations of the sensor loop and then writing the motor commands out |
| 02:16.18 | Tiersten | It's not as simple as just using libusb but it's easier than writing a whole driver |
| 02:16.32 | VoodooZ | ok. |
| 02:16.48 | VoodooZ | I'm starting to read about USB and how the protocol is handled. |
| 02:17.05 | Tiersten | AVRs have built in USB? |
| 02:17.05 | VoodooZ | command packets vs the others and all... |
| 02:17.11 | Tiersten | (Not looked at them recently) |
| 02:17.19 | VoodooZ | the newer ones apparently do |
| 02:17.28 | Tiersten | You basically have multiple channels called endpoints (don't ask. it's a stupid name yes) |
| 02:17.35 | Tiersten | Each one is only one direction |
| 02:17.43 | VoodooZ | but I might actually try that software based USB stack for the AVRs first. |
| 02:17.46 | Tiersten | You can have 16 both ways. 0 is always the control channel |
| 02:17.48 | VoodooZ | i see. |
| 02:18.05 | Tiersten | The AVRs have the required hardware though yes? |
| 02:18.11 | VoodooZ | So I would request one for reading and one for sending my motor commands. |
| 02:18.22 | Tiersten | I wouldn't try interfacing one that doesn't have the USB hardware |
| 02:18.36 | VoodooZ | really? |
| 02:18.53 | Tiersten | Yes. There are strict timing requirements |
| 02:19.06 | VoodooZ | I don't want to have one of those FTDI USB 2 serial chip thingy as they limit me. |
| 02:19.07 | Tiersten | You can't just bitbang it out like serial |
| 02:19.37 | VoodooZ | yeah, but it works. It's limited to 1.5mbit but that's enough. It's still plenty fast comapred to 100Khz I2C. |
| 02:19.48 | Tiersten | what works? |
| 02:20.06 | VoodooZ | the AVRUSB thing. check it out:http://www.obdev.at/products/avrusb/ |
| 02:20.30 | VoodooZ | There is another opensource one around too. |
| 02:20.43 | VoodooZ | and it even uses my favorite AVR! mega8. |
| 02:20.53 | Tiersten | ew. bitbanged USB |
| 02:21.00 | Tiersten | it is the slow one however |
| 02:21.08 | Tiersten | Only mice and keyboards really use that speed |
| 02:21.12 | VoodooZ | slow what? USB mode? |
| 02:21.14 | Tiersten | Yes |
| 02:21.28 | VoodooZ | Wouldn't that be plenty for motor commands at 20Hz? |
| 02:21.42 | Tiersten | How tight are you timing requirements? |
| 02:21.45 | Tiersten | are your |
| 02:21.57 | VoodooZ | I read sensors and write resulting motor commands every 40-60ms. |
| 02:22.03 | VoodooZ | Not that tight. |
| 02:22.07 | Tiersten | It's okay if you miss it by a few ms however? |
| 02:22.14 | VoodooZ | I guess. |
| 02:22.29 | VoodooZ | Until I realise otherwise :) |
| 02:22.33 | Tiersten | Should be okay then... |
| 02:23.03 | Tiersten | It's not isosynchronous so you're not gauranteed when it will complete |
| 02:23.18 | VoodooZ | well, I'll give it a quite try anyways. cheaper and faster that buying a new AVR. |
| 02:23.23 | Tiersten | Yeah |
| 02:23.25 | VoodooZ | that's fine. |
| 02:23.30 | VoodooZ | for now at least. |
| 02:23.52 | VoodooZ | so I'll check that puppy package then. |
| 02:24.00 | Tiersten | Yup |
| 02:24.14 | Tiersten | Your AVR run at 12MHz? |
| 02:24.19 | VoodooZ | and then If I feel more advanturous I'll write my own driver. |
| 02:24.23 | VoodooZ | 16Mhz |
| 02:24.32 | VoodooZ | which is overkill in most cases. |
| 02:24.37 | Tiersten | You'll have to change it for 12MHz |
| 02:25.04 | Tiersten | Timing again |
| 02:25.05 | VoodooZ | yeah, but just the master one which will talk to the slug and fetch updates from all the sensors. |
| 02:25.41 | VoodooZ | I might have it read the sensors via I2C so it will essentially become a USB-to-I2C hub. |
| 02:25.47 | Tiersten | Depending on how tight the timing requirements of the code are, you might need to dedicate a single AVR to do the USB and another one to do everything else. Not hard though and still cheap |
| 02:25.56 | Tiersten | Depends on how much stuff you're actually doing |
| 02:26.10 | VoodooZ | yep. exactly what I wanted to do. |
| 02:26.17 | VoodooZ | not that much. |
| 02:26.24 | Tiersten | What on earth are you doing to the I2C bus anyway?? |
| 02:26.36 | VoodooZ | I already have mega8's co-processors for most of my sensors |
| 02:26.46 | VoodooZ | and one on the motor controller. |
| 02:27.06 | VoodooZ | I have no clue yet. I must have zapped it with 5V by accident. |
| 02:27.15 | VoodooZ | It was too slow anyways. |
| 02:27.16 | VoodooZ | :) |
| 02:27.19 | Tiersten | that's what you did the first time around though! |
| 02:27.29 | Tiersten | you like wearing wool jumpers? :) |
| 02:27.33 | VoodooZ | yeah I know :( |
| 02:27.58 | VoodooZ | I noticed quite a performance dropped in my vision code when actually enabling the I2C read. |
| 02:28.20 | VoodooZ | so that will help motivate me to move to USB. |
| 02:28.27 | VoodooZ | which I was too lazy to do in the first place. |
| 02:28.38 | VoodooZ | Plus It's good experience. |
| 02:28.57 | Tiersten | yeah |
| 02:29.26 | Tiersten | you using a USB camera? |
| 02:29.38 | VoodooZ | yep |
| 02:29.41 | VoodooZ | usb 2.0! |
| 02:30.04 | VoodooZ | there was no driver so I had to hack the experimental one from a guy I found online. |
| 02:30.23 | VoodooZ | so I have good example code if I ever move to a kernel module. |
| 02:30.31 | VoodooZ | (for my I/O module) |
| 02:30.57 | Tiersten | Use a FPGA ;) |
| 02:31.26 | VoodooZ | heheheh |
| 02:31.43 | VoodooZ | I know! that's another thing I want to try eventually. but It's overkill for now/ |
| 02:33.03 | VoodooZ | I just tried the RTC again (running on AC instead of battery this time) and it still fails so... |
| 02:33.13 | [g2] | Tiersten, hey FPGA boy... when are you gonna sends some VHDL/Verilog our way ? |
| 02:33.17 | [g2] | :) |
| 02:33.25 | VoodooZ | I guess the RTC module is just another module built internally right? |
| 02:33.30 | VoodooZ | hehe |
| 02:34.31 | Tiersten | You'll have to talk to ka6sox when he gets back :) |
| 02:42.03 | VoodooZ | Tiersten, here's the original AVRUSB: http://www.cesko.host.sk/ |
| 02:43.35 | Tiersten | Thanks |
| 03:26.39 | VoodooZ | night |
| 12:19.05 | *** join/#openjtag prpplague (~billybob@72.22.135.245) |
| 12:33.14 | *** join/#openjtag VoodooZ (~VoodooZ@CPE00c0f021712c-CM014110003785.cpe.net.cable.rogers.com) |
| 13:04.07 | *** join/#openjtag VoodooZ (~VoodooZ@CPE00c0f021712c-CM014110003785.cpe.net.cable.rogers.com) |
| 13:25.02 | *** join/#openjtag [g2] (~g2@g2.nslu2-linux) |
| 14:41.04 | *** join/#openjtag ep1220 (~NN@gate.epatec.at) |
| 16:05.27 | *** join/#openjtag ByronT (ByronT@ByronT.nslu2-linux) |
| 16:08.06 | *** join/#openjtag prpplague (~billybob@72.22.135.245) |
| 16:52.09 | *** join/#openjtag ChanServ (ChanServ@services.) |
| 16:52.09 | *** join/#openjtag prpplague (~billybob@72.22.135.245) |
| 16:52.09 | *** join/#openjtag ByronT (ByronT@ByronT.nslu2-linux) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag ep1220 (~NN@gate.epatec.at) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag [g2] (~g2@g2.nslu2-linux) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag VoodooZ_Work (~VoodooZ@CPE00c0f021712c-CM014110003785.cpe.net.cable.rogers.com) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag Tiersten (tman@Tiersten.nslu2-linux) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag rwhitby-away (~rwhitby@rwhitby.nslu2-linux) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag dyoung (~dyoung@dyoung.nslu2-linux) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag nslu2-log (dyoung@dyoung.nslu2-linux) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag jbowler-away (~jbowler@jbowler.nslu2-linux) [NETSPLIT VICTIM] |
| 16:52.09 | *** join/#openjtag cbpage (~cpage@ip68-6-39-177.sb.sd.cox.net) [NETSPLIT VICTIM] |
| 16:52.10 | *** join/#openjtag jacques (~jacques@jacques.nslu2-linux) [NETSPLIT VICTIM] |
| 16:52.10 | *** join/#openjtag ka6sox-office (~tking@netblock-66-159-209-52.dslextreme.com) [NETSPLIT VICTIM] |
| 16:52.10 | *** join/#openjtag pjames (~jcouzens@mail.lost-carrier.com) [NETSPLIT VICTIM] |
| 16:52.10 | *** mode/#openjtag [+o ChanServ] by irc.freenode.net |
| 17:51.32 | *** join/#openjtag nslu2-log (dyoung@dyoung.nslu2-linux) |
| 17:51.59 | *** join/#openjtag ka6sox-office (~tking@netblock-66-159-209-52.dslextreme.com) |
| 18:57.04 | *** join/#openjtag beewoolie-afk (~beewoolie@florence.buici.com) |
| 20:54.41 | *** join/#openjtag [g2] (~g2@g2.nslu2-linux) |
| 22:25.50 | *** join/#openjtag jamie (~jamie@jamie.active.supporter.pdpc) |