mtwieg, good info on the free compilers. Free is good. Yes, one of the pins connects to the programmer interface, which also handles the in-circuit debugging. You can set one or breakpoints in the code, hit the icon to download and run your code on the chip, and then it will halt at your spot(s), at which point you can see the value of any of the registers or your own variables, including arrays in hierarchical format, and in your choice of hex or decimal. If I recall correctly you can write to registers, also, though I've never found a need to do that.
jderimig, I agree, if I already had programmed PICS or ATMELs, I would probably stick with that. But since I was starting fresh, (like ghost appears to be) I was happy to go with Zilog because its development kit was such a good value, I wanted to use a higher-level language, and the features, price, and performance of the MCUs themselves were competitive. Math is part of the C package, as are lookup tables, and an I2C interface is in the Zilog website library, for the Zilog MCUs that have an I2C peripheral (not the XP series.
, but they do post code for SPI). I know PICs have a huge following and I'm sure there's lots of good code out there for doing a variety of tasks. I have found the Zilog library to be helpful and to have code examples for most of the MCU features and interfaces. And of course, there's a ton of C code out there for other functions not specific to MCUs. I also like to use the XP series features to reduce the parts count and therefore the number of things I might screw up in the analog signal conditioning.
mtwieg and Guy,
The 13-bit functionality for the A/D is definitely there, and I've used it with success. In my last project I was making a controller to balance 8 li-ion cells, and I had a very simple voltage divider for the whole battery stack that made those extra 2 bits really come in handy. I started with their example code that loses the 2 bits of precision, and later I modified the code to just multiply the 13-bit register output by a scale factor that was large enough to preserve all of the precision when I right-shifted the resulting bits to get the voltage values in mV. I can dig up plots of the before and after voltage traces if you're interested, where you can clearly see the jaggies go away in the 13-bit version.
I don't know why they downplay the capability the way they do the in the spec. I ran across a similar thing where in the version of the spec they had last summer, they describe how the differential amp can be used in a 20x mode, but in the current version, they removed references to that feature in and even labeled the register bits that controlled it "reserved." I never used that feature myself, so perhaps they had gotten complaints that it wasn't working well enough, but that's definitely not my experience with the 13-bit A/D. If someone wants to try out the 20x differential input feature (for a resistor bridge pressure transducer interface without external components, for example), I'll post an excerpt from the older version of the documentation.