Re: [avrdude-dev] Decreasing programming speed for mega88

[Joerg Wunsch]

As Tomas Bergh wrote:

> I am not sure of how this kind of mailinglist works and since the
> forum at http://savannah.nongnu.org/support/?group=avrdude allmost
> seems to be dead I try this type of sendout instead. If this is the
> wrong way of doing it, please dont be angry on me!

Ouch!

I didn't even know that support tracker is activated...

Brian, I think we should have a quick look at the recent questions
posted there, and then close it down.

In general, I'd say either this mailing list, avr-chat, or
avr-gcc-list are the best places to ask.

> I have built a galvanically separated programmer for avr devices and
> got some problems when I am programming the mega88 device. Usually I
> use ponyprog but since it is not supporting the mega88 device I try
> with avrdude. I am able to program my device when running at 8MHz,
> external crystal, but not with 20MHz, external crystal. The high
> fusebit-byte is left unchanged and the low fusebit is set to
> 0xf7. One thing that is interesting is that I can program the device
> with 8MHz and then change the crystal and the device then seems to
> work fine at 20MHz! This makes me suspect that my programmer is
> alittle slow so I would like to set down the programming speed of
> avrdude.

Well, there has been a discussion about this (was it here or on
avr-chat?  I forgot) a few days ago.  It seems right now, this can
only be done in the sourcecode itself, as the respective feature was
merely meant as a debugging aid only.

However, I guess slowing down avrdude wouldn't help you.  Based on
your description, I'm rather afraid that your edges behind the
optocoupler are too slow, and while the device tolerates that at a
lower clock frequency, it doesn't anymore at full speed.  Obviously,
slowing down avrdude wouldn't help you as the edges would still be the
same.  (It would only help for the opposite case, if avrdude were too
fast for the device's clock frequency to follow, i.e. programming
works at higher clock but fails at low clocks.)

Alternatively, try adding a couple of gates behind the optocoupler, so
the edges will become steep again (due to the amplification imposed by
the gates).

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)