Re: [avrdude-dev] avrdude problem with Attiny13 - works fine in AVRstudio

[Haberler Michael]

Am 16.02.2009 um 22:45 schrieb Joerg Wunsch:

> As Michael Haberler wrote:
>
> > I eventually figured out that the Kee Electronics JTAGICE mkII clone
> > fails to implement CMD_SPI_MULTI properly.
>
> I think for the JTAG ICE mkII code, we could even change AVRDUDE to
> use "read signature" rather than "spi multi", because there are only
> two clones I'm aware of (the Kee Electronics one, and an experimental
> version of usbprog, but I'm not sure whether they even do implement
> ISP at all), so there's little risk of breaking something else.

it's not just the signature read which is broken - all memory reads  
using CMD_SPI_MULTI (eeprom, fuses, flash etc) are broken.

so that would mean switching to  
CMD_READ_{FLASH,EEPROM,FUSE,LOCK,SIGNATURE}_ISP to work around it.

chip erase with avrdude currently doesnt work either, but I havent  
nailed down why yet.

> Anyway, as the STK500v2 implementation of the "read signature" command
> is almost identical to using "read multi", there's no real advantage
> anyway.  For sure, it would be good if Kee Electronics changed their
> firmware for full compliance with the described protocol.

I massaged them a bit and they seem to be willing. Let's see what they  
come up with.

To be fair, they advertise AVRStudio & IAR compatibility only. Maybe  
they see the light in a market beyond manufacturer products.


> What's interesting with your observation is that obviously, Kee
> Electronics really wrote a firmware version of their own.

yes, it's lacking bug-for-bug compatibility at this point

--

a newbie question: why is it when the atmega16 data sheet says in  
table 116 that the serial programming instruction for chip erase is:

Chip Erase (Program Memory/EEPROM) $AC $80 $00 $00

and in avrdude.conf it says:


chip_erase       = "1 0 1 0  1 1 0 0    1 0 0 x  x x x x",
                   "x x x x  x x x x    x x x x  x x x x";

shouldnt this be more like

chip_erase       = "1 0 1 0  1 1 0 0    1 0 0 0  0 0 0 0",
                   "0 0 0 0  0 0 0 0    0 0 0 0  0 0 0 0";


doesnt make a difference for the bugs in question, just trying to  
understand whats going on

-Michael