Re: [avr-chat] avrdude : chip not responding

[Vincent Trouilliez]

> I am getting tired of seeing crap signal, so I will unplug the cable,
> and look at the signal stragiht from the parport connector of the PC,
> using just two tiny bits of wires to help the scop's probe get in the
> pin holes. If the signal aren't perfectly/decently square at this
> point... :-/

Okay, I just did it... scope's probe straight into the parport's pic
holes.... result is appalling !! :-O :-/

First, as I thought, the MISO line remains perfectly inactive, so
whatever signal I managed to see yesterday on this line with the AVR
disconnected, has to be indeed a crap cable and/or board soldering, so
this needs investigating and fixing for sure.

However the PC's parport is indeed far from perfect either !
I looked at SCK and MOSI and both show similar signal shapes :

1) the rising edge starts "rounding" badly from as low as 3.2V and
doesn't settle until about 50% of the "high" half-period of the clock
cycle ! Worse.... the high-level is far from 5.0V... it's only 4.2V !
I am shocked...
Still, according to the data sheet, 0.6Vcc is enough to  warrant a logic
one to ber read. 4.2V is over 0.8+ Vcc so should be okay. However the
rising edge starts detériorating from as low as 3.2V, hence 0.64Vcc, not
far from the threshold then, so the input AVR's MOSI input might
"hesitate"... not good ! 
So the outcome is, so far, that also 4.2V is very far from 5.0V, the AVR
should be perfectly happy with it (in theory at least), however the
rising edge could do with some buffering or even Shcmitt triggering so
to speak.
And the cable also need to be check for sure !

Well, at least I am progressing, it's now pretty clear that I have crap
signals, so I will keep working in this direction...


Regards,


--
Vince