Erik, please be careful before you go off on a tirade about something you've assumed but not actually read.

I didn't say that many people use mode 0. I don't recall reading that anyone else said that either. What I said was that I want to use mode 0. I also said that the majority of the 805x app's I've recently (in the past 4-5 years) had in front of me use the serial port only for debug-console purposes if at all. None of them communicate with other systems in their target application. This doesn't mean that it doesn't happen, but only that the stuff I see doesn't do that, at least not yet.

Mode 0 is not terribly useful in a standard 805x so long as it's tied to the oscillator frequency of a 12-clocker because every 12 clocks it accumulates another bit, so in 8 instruction cycles, you have to dispose of another byte. That uses up lots of processing bandwidth. A 4-clocker or a 1-clocker has a bit of an advantage, though. It can dispose of a byte in just a few cycles which can be less than a bit of data, so it can do lots of useful work while the serial hardware assembles the bytes for it. An interrupt can tell it when to snag the now assembled byte, and there's plenty of time to execute and dismiss the isr before the next byte is assembled, or, in the case of a one-clocker, even before the next bit is shifted in.

I believe I wrote you what I'm planning for this Mode 0 serial port, which is a sample path to facilitate the accumulation of samples of data taken from a floppy disk drive that may or may not be hard-sectored. Hard sectored media were popular (a) before there were LSI FDC's because they were easy to implement using a USRT and some logic, or in any of a dozen other ways, and (b) because they held more data than the FM-encoded soft-sectored (IBM-3740) formats allowed, typically ~4KB per track rather than ~3. 20+ years ago, these were popular formats for use in medical instruments of various sorts, particularly imaging machines. As the media have aged, it's become difficult for them to be preserved, as the regulatory bodies in many countries demand. About a decade ago, I had to reverse-engineer an FDC for hard sectored media for a company in India who had to deal with deteriorating media in an environment that demanded longer term storage of their already 20-year-old data. This meant that they had to read, error-correct, and renew their data. Now I have to repeat that process with a device that's sufficiently general to handle both hard and soft-sectored floppy media via a USB port. I think the use of the serial port in mode 0 is well suited for this sort of operation. It also provides a place for me to speed-test the DS89C420's I've got on hand, since they were "rated" for 50 MHz operation, which was later termed unattainable because the internal FLASH memory didn't operate at that speed. The internal XRAM, which can be mapped in to either code or data space can, however, and a routine to respond to an interrupt when the CPU is running at 48 MHz, which should result in a 4 MHz sample rate in Mode 0 with a divide-by-12 clock, if the CPU is operated from a 12 MHz oscillator, would work very well, indeed, and, in fact, it should work out just fine at half that rate. That's where mode 0 comes in. Nobody thinks it's in use everywhere. However, I've already thought of yet another application, though that may be beyond the Maxim part's capabilities.

Now, you can go ahead and make your point, but I'm not sure what you're trying to prove.

RE