Scuzzbox arrives!

[scuzz]

So, a brief update. Since today marked my last day of lecture, and I had a spare couple of hours, my friend and I assembled the first chunk of the scuzzbox.

Partially assembled scuzzbox

scuzzbox_part1_assembled_scaled.jpg (79.12 KiB) Viewed 842 times

It only has the power section and the actual LPC2388 (plus supporting circuitry), and the RAM. I'm going to try to get some output when I get braindead trying to study.

WheeeeeeeeeeeeeeeeeeeeeeEE!!!!

[lightfoot256]

Wow, very cool. Keep us posted with the progress!!

[scuzz]

Wow, very cool. Keep us posted with the progress!!

I'll do my best to keep you guys up to date! I'm really excited to see so much interest in this .

Right now I'm wrestling with getting the ARM toolchain built, and I'm also trying to develop a base file to compile from. I currently have what appears to be a working gcc, ar, as, ld, and possibly gdb, but I'm not willing to upload the code until I'm sure I don't accidentally set the code protect bits.

I've also just been trying to find things like the "lpc23XX.h" file and similar chunks, but most of it seems to come through this company called KEIL software who have special licensing rules which I don't really want attached to the kernel code.

Anyway, I'll post some actual information after wednesday when I'll have finished my last exam, and will have some time to have truly pulled out some hair before I ask for too much help

[lightfoot256]

Lol, after seeing your post it spurred me on to look back into all those little ARM7 microcontrollers and start planning my own super uzebox! Unfortunately - although I did electronics back in school - I'm really a software engineer who just complains that the hardware is never "quite" right with a passion to do it all myself and have it all working juuuuussst the way I want. hehe.

Anyway, now that the 1284 is starting to filter into mainstream I should return back to my sonic port

Chris

[paul]

Looking great, scuzz. Can't wait to see it up and running.

Anyway, now that the 1284 is starting to filter into mainstream I should return back to my sonic port

That'd be neat, too

[scuzz]

The learning curve enters the game! So now that I have some of it built, I'm trying to do all the fun stuff involved in getting the build environment set up and whatnot.

I now have a working gcc-4.4.0, gdb-6.8, ar, as, ld, and the std libraries from newlib. I can compile code and run it in an ARM7 simulator! Woo!

Now the obnoxious part is trying to get my arm-usb-ocd Jtag chain to connect to the ARM. I was trying to use urjtag but it keeps complaining that "TDO appears to be stuck at 0" or "TDO appears to be stuck at 1". I have a sinking feeling that I failed to swap TDI and TDO, but the schematic for the olimex LPC2378 and LPC2478 dev boards appear to match what I did.

So I'm thinking it's possibly this could just be an issue with urjtag. So as a possible solution I'm attempting to compile openocd, though so far I have had poor luck... Does anyone have any idea if "./configure --build=i686-pc-linux-gnu --host=arm-elf --enable-ft2232_libftdi" isn't correct? The main issue there is that when I attempt to compile with those options I end up with

Code: Select all

../../src/helper/replacements.h:133:2: error: #error no usleep defined for your platform

I'm going to try some fairly silly stuff to see if urjtag is really at fault, but if anyone has any inspiration, anything would help.

Thanks!

[scuzz]

Update: It looks like it's probably working!

I'm still playing around with getting everything up and running properly but I got Openocd 0.4.0 up and running and it appears to both see and be able to flash the board. I'm still trying to get the code to execute, but something in the built-in USB bootloader appears to be making it switch into Thumb mode and not letting go. This may be one of those "I overwrote the bootloader and now it's b0rked" issues but if so I'll just have to wipe it out as I wasn't planning on using it anyway.

I'm also documenting all of this and writing bash scripts to set it all up for the next unfortunate soul to attempt this. I'll work on cross-compiling all of this and setting up eclipse in windows after it's all working, don't worry

Anyway I'm excited! I can't wait to get the LED blinking and onto the more exciting kernel code!

[scuzz]

Debug debug debug!

I've been trying to get code running and have all sorts of stuff which seems to be working, except, of course, that I still can't get my code running properly. I decided to take a look at the errata for the chip, and realized that I didn't put anything connected to the VBAT pin (the battery voltage for the real-time clock module). This seemed completely reasonable during design, but apparently this causes the module to not fully start up on the revision of the chip I'm using! Blast!

So, a green-wire solution and we'll see if I can get some blinking LEDs to show soon.

[scuzz]

THE LED! IT BLIIIIIINNNNNNNNKS!!!!!!!!!!!!!

Here's a link to a video!

So many hours of work... soooooooo satisfying to see the LED finally blink .

This confirms that the super-basic baseline systems are working, and that the compiler is, in fact, compiling for the system properly. I still haven't gone through the rigamarole of actually making the clock input go to good use (this is being run on the 4MHz internal clock), nor does it use the RAM, but it *DOES* use the ARM and the JTAG interface, and that's a damn fine start .

Unfortunately you'll probably be sad to hear that it doesn't yet boot on its own properly yet, this is something which I need the debugger whipped out for and I have to kind of trick it into actually running the code. This looks like it's because I'm making some mistakes in compiling it, but that likely has to do with me failing to compile the standard library stuff which calls my main routine.

Never fear though! I'll get this working! With what I have I should no be able to output video shortly, though getting the stuff booting properly without needing the debugger is probably my next step .

IT LIIIIIIIIIIVES!!!!

[lightfoot256]

yay, Good work!

Strange how your board looks so big with all the detail but when you place it next to something you can see that it's actually quite small!

Keep it up, can't wait to see it output video! hehe.

Chris