Had you taken time to read the posts to which you claim to be responding, you'd realize how far off you've drifted.

I've never seen a JTAG channel without the pullups to which you refer, and I've been using 'em since the early '90s when Cypress, Altera, Lattice, and XILINX started pushing 'em as programming tools.

If you'd taken the time to read the post against which you're constantly railing, you'd much better understand what's going on. I've been considering a test. You're going on about some nebulous concept, a concept which your responses to Lynn's remarks clearly indicate you can't possibly understand, else you wouldn't say what you say.

The reason I demand the packages I've said I want to use IN MY TEST FIXTURE, is that many, Many, MANY of the currently offerred 805x's (not just the old timers) are available in those packages and share a common pinout, which makes them suitable for testing multiple vendors' products in the same test fixture. Obviously, you've never participated in any sort of testing, else that would have occurred to you by now.

As it happens, I'm also interested in testing the components that I intend to continue to use in upgrading existing applications, some of which app's are decades old, and perhaps expand the product offerings from among which I select them.

No one but you says, "There is no problem ..." though several have said, in one way or another, that they're satisfied with the results they've observed with the use of one or another supervisor.

I'm just keeping an open mind. I've seen no trouble with FLASH corruption, since I don't use the internal FLASH, at least not yet, in the components I do use.

I'm willing to set up a real test that would establish how frequently over a test of perhaps 100 million power cycles, collectively, varying, first one at a time, and later in combination, several seemingly critical parameters and comparing the result attained with RC reset vs. the MAX1232, and, later, perhaps other supervisors and making statistical comparisons of the outcomes. (BTW, That would render me eminently qualified to testify in court, about this particular matter, as I'll have directly observed the event in question and be both professionally and academically qualified to interpret the result.)

Needless to say, occasional observation of an upset that occurs "from time to time" resulting in FLASH corruption or some other malfucntion, or observation of a circuit or two while switching power on and off a hundred, or even a thousand times, would probably not meet the mark.

You mentioned the figure 99.995% a number of times in your remarks responding to Lynn's remarks. That's 50 ppm. How, exactly, did you arrive at that very precise figure? What's the margin of error?

I've caught you spewing numbers you couldn't support before. In the past, this has resulted in you quietly tiptoeing backward. Savvy readers probably take that into consideration and simply ignore the bombastic remarks and random numbers you often spout. Everyone, by now, realizes you have poor grasp of the difference between fact and conjecture. In fact, there's plenty of basis for one to question your awareness of a difference between fact and fiction.

All along, however, you've failed to grasp that the "issue" is question about whether the malfuntion initally of concern is a consequence of faulty reset control or of faulty power control and filtering. It could even be a question of improper oscillator management. I'm just thinking that this should at least once, be quantified, so people see what they gain with a supervisor, and under what conditions.

Now, you quote me saying, "Your favorites are, of course, inelegible because they're not offered in the most popular packages and supply voltage. " to which you respond,


I'd say that, having taken a cursory look at the offerings from ATMEL, NXP, TI, Analog Devices, etc, I've concluded that only a few offer a common and readily socketable pinout, and those are the DIP-40 and PLCC-44. If you know of a package such as TSOP, TSSOP, PQFP, BGA, or the like that can be socketed in a common (meaing the same, since I'm not going to build more than one of these test fixtures with six devices on board), ubiquitous, inexpensive (and having more than one vendor), and reliable test/burn-in socket, and having precisely the same pinout, then please share that information with us, particularly if it's also shared by the current generation Maxim/Dallas parts that currently interest me. Otherwise please stop trumpeting your lack of comprehension.

RE