Open GoogleCodeExporter opened 9 years ago
Makefile is for Arduino boards - it has settings for most common Arduino
boards. Fuse settings cover the
oscillator circuitry as well as the microprocessor - for custom boards you will
need a custom Makefile, as you
have a custom circuit.
Extended fuse: FD is equivalent to 05 (only bottom 3 bits are significant, and
Avrdude complains if the top
bits are nonzero)
High fuse: You have the incorrect BOOTSZ settings. Remember Optiboot is small.
Low fuse: Your oscillator settings are different. This may or may not be
correct depending on your circuit.
Original comment by cathed...@gmail.com
on 6 May 2010 at 1:26
I DON'T HAVE A CUSTOM CIRCUIT, I HAVE A PRO MINI WITH A 328P - standard,
supported.
The Pro Mini is in no way a custom board, either with the 168 or 328, at 8, or
16
Mhz. the 20Mhz would be custom but that is produced by the script.
Original comment by tz2026
on 6 May 2010 at 1:46
I won't update "WontFix", but you should note it won't work to flash NORMAL,
PLAIN
ARDUINOS correctly.
Original comment by tz2026
on 6 May 2010 at 1:47
Sorry for misunderstanding - original issue did not specify the target board,
only "my atmega328p". Ease up on
the caps please.
Please confirm exact board - which voltage variant, which CPU speed, and I'll
investigate.
Original comment by cathed...@gmail.com
on 6 May 2010 at 2:03
I have to admit I'm thoroughly confused at this point - apparently some
different
types and models (I searched for an answer and got several which I have thrown
together and con-fused). I was trying to use AVR-Studio and it has settings
which
just confuses me more, at least for the 328p. It even says the bootloader is at
0x3x00 instead of 0x7x00 for those fuse bits (4.18 SP3).
I will try to sort it all out and get back with a definitive answer. I think I
merely exposed a longstanding question - what should the fuse bits be.
Original comment by tz2026
on 6 May 2010 at 3:54
To sort out that address confusion:
Atmel refer to addresses as words, but GCC has byte level access to the flash,
so it refers to addresses as
bytes. So GCC addresses are twice the equivalent Atmel address. Its not a bug -
just a different perspective.
All the fuse and lock bits should be correct for the appropriate boards. They
have been based on the latest
Arduino distribution, but with changes to accommodate the smaller bootloader.
If they are incorrect, I would
like to know exactly which board is incorrect, what you propose it should be,
and why.
As we don't have a clear defect yet, I'm relabelling this as a low priority
'Other' issue until we have more data.
Happy to reassess when more data comes in.
Original comment by cathed...@gmail.com
on 17 May 2010 at 7:19
The fuse settings atmega168 is wrong, the bootsize is set to 512 words instead
of the 256 words. Extended fuse should be 0xFC instead of 0x02
Original comment by superkoe...@gmail.com
on 18 Sep 2010 at 10:38
Agree with comment #7. Changing EFUSE to 0xFC allows me to install and use
optiboot with avrdude on my freeduino
Original comment by chris.ku...@gmail.com
on 26 Jan 2011 at 6:13
[deleted comment]
Why do the "maximum_size" and "high_fuses" settings vary amongst ATmega328
boards? Isn't Optiboot always 512 bytes for this chip, corresponding to a
HFUSE of 0xdc?
From the relevant sections of the Makefile and boards.txt:
at 16MHz
========
atmega328_isp: HFUSE = DE
--
atmega328o.bootloader.high_fuses=0xde
atmega328o.upload.maximum_size=32256
at 8MHz
=======
atmega328_pro8_isp: HFUSE = DE
--
pro328o.bootloader.high_fuses=0xdc
pro328o.upload.maximum_size=30720
--
lilypad328o.bootloader.high_fuses=0xdc
lilypad328o.upload.maximum_size=31744
Original comment by tim...@gmail.com
on 7 Mar 2011 at 4:36
Whoops, bootloader memory is expressed in 16-bit words. So 512 bytes = 256
words, and the Makefile fuse settings are all correct. But I still have some
doubts about the Optiboot boards.txt settings for the pro328 and lilypad328
boards.
Original comment by tim...@gmail.com
on 9 Mar 2011 at 1:58
Original comment by wes...@gmail.com
on 29 Oct 2011 at 5:27
Original issue reported on code.google.com by
tz2026
on 3 May 2010 at 7:57