I'm using ADuC836 with 85SMX watch crystal and all is working fine. With the
same crystal and ADuC848/847 samples, the oscillation don't start. I try 3
samples but nothing. why? could oscillator section be different from 836 to
847/848 chip? could be that all the 3 samples are damaged?
There is no difference between the parts. If it is working on the 836 then
there is no reason why if shouldn't work on the 847.
One thing I will say though is that Roberto should ensure that there are no
capacitors (decoupling) used on the XTAL1 and XTAL2 pins. There is sufficient
internal capacitance to ensure oscillation.
Confirm that the crystal is indeed across XTAL1 and XTAL2 and that when the
part is powered up that oscillation can be seen on XTAL2.
This should be 32.768kHz. Do not load XTAL1 as this could damp the crystal
sufficiently to kill oscillation.
If XTAL2 is not oscillating (with the correct crystal connection and the part
powered up) then I would suggest that either
1. the part is not actually powered up correctly (the 836 and 847 are not pin
2. the crystal is either faulty or is sufficiently far away such that the
impedance, parasitic capacitance and inherent inductance of the traces are
loading the crystal sufficiently to kill the oscillation.
The PLLCON SFR is only of use to select a particular core operating speed.