Linux on 7" mini netbook ARM-VT8500 ?

[litch84]

Hej hej!

I recently bought the netbook equipped with the ARM-VT8500 and I tried everething you said about how to fix the keyboard/mouse problem but i didn't manage to make it work. Does't anybody know how I can figure out if the problem comes from the keyboard or the motherboard?

If the software restore failed,

The keyboard has 3 clips across the top and two screws underneath that hold it in. You could try taking it out and making sure the ribbon connector is properly in it's socket.

The mouse ribbon cable is accessibe via taking all the bottom screws out and unclipping the front of the case. The ribbon cable connect to the mobo just left of center.

Make sure both are in properly, and for the next statement's sake DONT SCRATCH THE CASE!

If still not working, perhaps send back for warranty?

By the way I'm also looking forward for a Linux version to be available for our mini netbook.
Thanks;

Um, that's the purpose of this thread, to find some way of putting Linux on the ARM 926 based SoC Netbooks....

So far PrFaas has managed to get a serial console working, I've compiled several Kernel's that 'support' the demo boards that the CPU's are based on. We're basically waiting on one of us to put the two together and get a Kernel running properly on it.

I'm waiting on parts from the UK for the serial connection, and I beleive PrFaas is getting around to building the Kernel based on the config I've supplied whenever he has the time... Others are trying their attempt at reverse engineering the WinCE kernel and other various paths towards getting a greater understanding of the platform.

We're not far off, once we get a 2.6.x Kernel going, the LCD won't take much with serial debugging and we're well on our way to slapping a filesystem on it and booting X.

[litch84]

The LCD controller is based on the ARM PL011.

I'll get you the LCD model # soon.

[PrFaas]

I just made an 'odd' observation on my 7" machine: About those gnd/rx/tx/5V pads that work as a serial port/u-boot console: the pad marked with 5V does *not* 'carry' 5 Volts.... I never noticed because i re-used an existing RS-232 interface which has an independent 5.0 Volts supply. The actual supply voltage on the pad marked 5V is *not* 5.0 Volts, but 'only' 3.3 Volts. This does mean that if anyone attempts to copy 'my' RS-232 buffer/driver he/she *is* going to have a little problem: the circuit is designed for 5.0V supply voltage, not for 3.3 Volts. On my board, there *is* a 5.0 Volts (well, i actually measure 5.2 Volts), but the pad marked '5V' is not 'it'.

I found an actual 5 Volts at the following two locations:
- when looked upon the board in the position where you can see the 2 Gig Flash-ship, the two 'bottom' pins of the USB connectors carry 5 Volts
- There is a small pad marked 5V near the front of the board that also carries 5 Volts....

I'll see if i can 'gimp' something to make some arrows at those locations....

Perhaps it would be easier to attach a 3.3 Volts 'native' RS-232 buffer to only the 'original' pads.... BTW: note to self: *never* blindly believe what is printed on a PCB

[litch84]

I just made an 'odd' observation on my 7" machine: About those gnd/rx/tx/5V pads that work as a serial port/u-boot console: the pad marked with 5V does *not* 'carry' 5 Volts.... I never noticed because i re-used an existing RS-232 interface which has an independent 5.0 Volts supply. The actual supply coltage on the pad marked 5V is *not* 5.0 Volts, but 'only' 3.3 Volts. This does mean that if anyone attempts to copy 'my' RS-232 buffer/driver he/she *is* going to have a little problem: the circuit is designed for 5.0V supply voltage, not for 3.3 Volts. On my board, thet *is* a 5.0 Volts (well, i actually measure 5.2 Volts), but the pad is not 'it'.

I found an actual 5 Volts at the following two locations:
- when looked upon the board in the position where you can see the 2 Gig Flash-ship, the two 'bottom' pins of the USB connectors carry 5 Volts
- There is a small pad marked 5V near the front of the board that also carries 5 Volts....

I'll see if i can 'gimp' something to make some arrows at those locations....



I have the machine running with the keyboard & touchpad removed at the moment.

I just used the 5v from the USB +ve rail.

[Matriark TerVel]

Thank you for taking the trouble to convert my ASCII-art to a more decent diagram

Only one remark w.r.t your 'schema': i did mention to connect the inputs of the unused inverters to either 5V or gnd; I should perhaps have put that in the diagrams as well . The reason for connecting them is that the unused inverters could start to draw quite some current if their inputs are left floating.

No problem.

Digital Electronics was never my best subject, unfortunately. At times, I wish I would have gotten more into it than I did.

Oh, and on very close inspection: on my board, the 'order' of the pads for the RS-232 is different from what you indicate on the schema; Could well be that 'what you draw is what you see' on your board, but on 'mine' the pins are gnd, tx,rx,5V in that order, from left to right.

I drew them it what seemed to be the logical order. I don't recall what the order is exactly.

I just made an 'odd' observation on my 7" machine: About those gnd/rx/tx/5V pads that work as a serial port/u-boot console: the pad marked with 5V does *not* 'carry' 5 Volts.... I never noticed because i re-used an existing RS-232 interface which has an independent 5.0 Volts supply. The actual supply coltage on the pad marked 5V is *not* 5.0 Volts, but 'only' 3.3 Volts.

It's a mystery then why they labeled the pad "5V." If it's really 3.3, shouldn't it be possible to get around the need for the independent 5V supply by basing the circuit on an 74LVT14?

[PrFaas]

No problem.
It's a mystery then why they labeled the pad "5V." If it's really 3.3, shouldn't it be possible to get around the need for the independent 5V supply by basing the circuit on an 74LVT14?

For the receiver, that would make things easier No more trouble to limit the output of the inverter to 3.3 Volts. I'm not quite sure if the transmitter would be 'seen' by a normal RS-232 input if the 'high' voltage is only 3.3 Volts, that could be a bit of a problem. Would be worth a try though...

FYI: 'My' circuit is also not *really* RS-232-conform: the output voltage for 'no signal' should -according to 'spec'- be below -2.7 Volts ...

Do you have an 74LV14 'around'? You'd have to change the zener diode to a 3.3 Volts one and 'drop' the entire circuit at the output of the receiver inverter (connect directly to the CPU input.. ) A simpler circuit that's for sure LV is also a low-power chip, so that should be no problem.

FYI: status: compiling an 'arm' kernel at the moment....

[Matriark TerVel]

Do you have an 74LV14 'around'? You'd have to change the zener diode to a 3.3 Volts one and 'drop' the entire circuit at the output of the receiver inverter (connect directly to the CPU input.. ) A simpler circuit that's for sure LV is also a low-power chip, so that should be no problem.

I don't have an 74LV14 lying around unfortunately If I did, I'd try to build the circuit.

[PrFaas]

I don't have an 74LV14 lying around unfortunately If I did, I'd try to build the circuit.

I'll see if i can get a 'ride' on a Farnell order 'at work'. I think i can still suffer the expense of 2.9 Euro's; For 10 of them that is

I've got my bulky 'external' buffer print working ok for now, but -in the end- i'll want the serial buffer 'inside'. Not only for the experiments: i have some 'use' for a serial port on this machine.

I do hope we'll be able to at least find the connector pins of the CPU module eventually: the CPU itself is likely to have more than one serial port, and i'd love to have one that u-boot does not 'babble' to.

NB: i had to edit the #148 post a bit: after the objcopy, you still have to 'gzip linux.bin', or else mkimage has no input

And: yes, i've got a kernel compiled, converted and in the tftp directory by now Attempting to boot it will be for tomorrow, i guess: for that, i also need to 'hook up' the serial console, isolate the network, 'down' the router..... It is a bit late for all that now...

[lauriesherrod]

I made the terrible mistake of buying one of these from some company in China. It has WinCE. After about 3 hours the Internet Explorer started with a message "IESAMPLE encountered a serious error and must shut down". I "chatted" with tech support and they wanted me to send it back to China. I did - it cost $30 in postage! It came back and worked another 3 hours and got the same thing again. This time I convinced them to send me the files to reload the computer.

Now I cannot get that to work. I extracted the files - its autorun.exe, 3-4 other files, and a folder with subfolders like Windows, Programs, etc. I put the files on a SD card and tried the autorun.exe and it blew away most of the files on the computer's drive - so many that it could no longer see the SD card.

So I hand copied the files back using a flash drive into folders like \Windows and \System, etc. I got it back to where it was before - IESAMPLE still giving errors. I tried booting with the SD card in but that did nothing. I tried running autorun again and it repeated the same thing.... grrrr....

I tried running the file from a flash drive, same thing.
I tried copying the files and folders to the computer's drive and running the autorun and it did the same thing.

The lovely error I get is "system disk no existed".

The tech support cannot speak English and they keep disconnecting me when I ask questions.

I'm pretty technically competent if anyone can tell me how to load an OS on one of these. This one has a 2 GB flash memory of some kind - no way I can find to get into anything like BIOS. It does start up into a bit of Windows.

Thanks!!!

[litch84]

NB: i had to edit the #148 post a bit: after the objcopy, you still have to 'gzip linux.bin', or else mkimage has no input

And: yes, i've got a kernel compiled, converted and in the tftp directory by now Attempting to boot it will be for tomorrow, i guess: for that, i also need to 'hook up' the serial console, isolate the network, 'down' the router..... It is a bit late for all that now...

FYI: The objcopy had no effect, I noticed the file size was the same, done a binary comparison of the 'before and after' - no difference.

If you do a 'make ARCH=arm CROSSTOOL=arm-x-whatever-'

It'll make the desired binary under arch/arm/boot/zImage

I'm still a bit hazy on loading address, I've seen some talk that the kernel should be loaded at 0x8000 because it uses the first 32K for paging/other vector based table....

Then again, I've seen other scriptcmd files that load directly to 0x00000000...

Can anyone shed some light?

ALSO:

The MAX232 supports 3.3v TTL input operation - see datasheet:
http://pdf1.alldatasheet.com/datashe.../MAX232IN.html

Page 3 under "recommended operating conditions"