Maybe not. My happy news was premature. Rich's suggested changes have gotten rid of the glitches I was seeing in the data that I was writing to, and reading from, the CF, but I'm still seeing the "sector offset" problem. This is with both my home-grown utility AND MYIDE (of course, I changed both to separate the ready and busy routines). It is also (sadly) with the Kingston CF.
For example, if I start at LBA 1 and fill it with hex 31 ('1'), and then go on to LBA 2 and fill it with hex 32 ('2'), etc. etc. When I go back and look at these sectors (read them back), they are offset by 1. In other words, if I again start at LBA 1 and read it, I get 512 hex 32 ('2'), and not what I expected. This offset continues as I read out in increasing LBAs. Curiously enough, the hex display on the card shows the correct LBA! The only way I can get rid of this is (and this is in my home-grown utility) to write a random sector (I've been using LBA 3Eh) in the initialization part of the code. Just once! If I do that, then everything is OK, and I see sector contents that I expect to see. If I don't, that 1 sector offset persists throughout my tests.
BUT, at least I'm not seeing stray garbage in what I read and write. I haven't seen any of that since I followed Rich's hardware and software suggestions.
That's some progress .... I guess.
(Oh, this is with a CompuPro CPU-Z (at 4 MHz), a CompuPro RAM 17 (64k static), and the N8VEM serial card.)
Roger