Malte,
This is going to be a very tough problem to solve, I fear. However, I have
been doing some experimentation here and have noticed that every time I
probe T0 on the baord and leave the probes attached the MCU remains stable.
What I have done is simulated the capacitance of the probe by soldering a
22pF ceramic capacitor between T0 and ground. At present my board is holding
stable. I know this does explain much but it will give me time to think
about what on earth is going on here! What happens if you revert to the
original resonator and capacitor combo and try the trick with the T0 line.
Do you get the same effect or is the board just as bad? It might also be
helpful to note that this board is susceptible to random bank switching just
after the top panel controls are adjusted. Which goes back to this whole S/H
data/IC6/probes on T0 conundrum...
Cheers,
Andy
On 01/09/2010 20:05, "Malte Rogacki" <
gacki@gacki.sax.de> wrote
>
> Andrew;
>
> I'm still in the "theory" stage of everything.
>
> It appears that my measurements as far as amplitudes went were probably
> completely off. They were done with a "x10" probe setting since this seemed
> to affect the operation of the KLM-367 least. However the waveforms and
> values looked very different from what I am getting now with a "x1" probe
> setting.
>
> Using two 33pF caps (as suggested by the 8048AH datasheet) did not work out
> well (instant lockup). Neither did using a quartz oscillator with the stock
> value caps (very unstable).
>
> I also experienced more stable behaviour when doing measurements yesterday.
>
> Right now I am using a 10pF ceramic cap for C23 (instead the 22pF cap) and
> the synth has yet to crash. I'll let it run for a few hours.
>
> One thing that was odd (or may reflect on my measuring abilities...) is
> that the signal for XTAL1 and XTAL2 is reading not 6 MHz but around 10 to
> 12 MHz. One of the new boards reads even 18 MHz!
>
[Non-text portions of this message have been removed]