John Anderson wrote:
That might be something to think about as a good varient: Just get rid of the on-chip RAM and Flash because there isn't enough to be useful for larger projects (Or don't worry about adding more on-chip memory).

NXP now do families that include Zero Flash, in M3 and M4 and Arm9 cores.
I think they do all include a BOOT ROM, so they can load and run from SRAM, (eg from USB) and load external FLASH.

Zero Flash is also how the entire RabbitSemi series have gone, and the Parallax Propeller.

The BOOT loader detail is important, and in Zilog's case, it may be easier to keep a single tapeout, but not make the Flash extreme in size.
Otherwise, they have to support ROM development and debug, and two mask sets...

hi,
nxp has also uC that can execute code from SPI/QuadSPI serial external memory , no need to cache/copy/download it into internal ram.

John, you're right, for higher-res ADCs off-chip is the only real option. But sometimes we don't need every bit of resolution - 10 bits would be fine for most of my ADC stuff (temp, battery temp, battery level, discharge rate, motor temp, etc, I'm happy with +/- 25mV resolution).

Jim : I'm not convinced about zero flash. I've played with some NXP M-series samples, and it seems to require at least another step in hardware design AND in programming technique. At least, it took me many days longer to get a working prototype when compared with the on-chip models that worked out-of-the-box. (At least some of this has to do with my bandwidth and skills, obviously!)

Externally flash mirroring SRAM images (a la ASICs, etc) is an ok solution in some situations, but as a small developer, it makes my design time longer and more expensive, both in terms of designing/evaluating and prototyping (let alone production). I generally only use off-chip memory for table lookups (flash ROM sine tables, pattern matching, and logging telemetry, etc. into external SRAM. So I'm more than happy to design with off-chip memory as long as I can do the initial development on-chip. This also means fewer prototypes! (this is a big deal for a little feller like me!)

I guess this also depends on the final external memory support - and the development tool facilities! If I just need to change a single CLI switch or project setting, initial development using offboard memory wouldn't be such a heartache.

One thing I've learned as I've bootstrapped myself into Harvard-style MCU development and design over the last 10 years - there's no such thing as too many hardware timers! Plus, the flexibility of allowing timers to daisychain usually offsets any additional device purchase cost, even for folks like me.

Tom Ormiston wrote in Early Feb :
We have two new products launching this month.

I wonder what those ' two new products launching in Feb' were ?
Does a new package count as a 'new product', or a SW release ?

Does "Captain Zilog" count as a product?

Nice production values on the e-mag, btw. Very sexy. Now, if someone else could spend an equivalent amount of time (and money, I guess) on solving existing customer problems, Lieutenant Customer would probably team up with Captain Zilog and fight electronic evil all over the place!

I'm sorry, I know, I should play nice here... But truly, what was running through my head as I read the magazine - it's now 8 months since I last heard from Zilog on my showstopping eZ80 technical query, and I can't seem to raise the case anymore... Maybe there's been another re-org behind the support site ? Or the support guys are moonlighting as inkers now?

(I did enjoy the emag though, very post-modern and even funny in parts).

Well, it's been seven years since Zilog introduced new microcontroller silicon. Over here we've designed out all zilog mcus.

I guess it's true, zilog is as dead as disco.