| ??? 08/01/07 22:10 Read: times Msg Score: +1 +1 Good Answer/Helpful |
#142603 - reset toujours Responding to: ???'s previous message |
Richard Erlacher said:
I'd be very interested, first of all, in what you define as "working well" along with how, exactly, you define it. That's very easy. RESET shall be equal to VCC while the latter is less than a defined value (bigger than VCCmin) plus a time sufficient for the oscillator to start (plus the 24 clocks, but that's negligible). Richard Erlacher said:
I've seen no evidence that any 805x has ever "worked well" within the constraints of any rigorous definition, at least insofar as its RESET behavior is concerned. Is the zillions of perfectly working installations (you mention below) under the widest and wildest power condition while having standard reset IC, insufficient evidence for you? Richard Erlacher said:
As far as I've seen, no MCU manufacturer has rigorously defined the controlling parameters, such as Vcc rise time, oscillator startup time, or any other oscillator behavior, response and duration of internal operations associated with RESET, etc. Are you willing to pay the price (in all sorts of meaning of this word) of such definition? Manufacturers traditionally fail to provide much more vital information, either from negligence or deliberately. They for example consistently insist on the fact that you shall not operate the chip outside a certain voltage range; which IMHO directly implies that the chip shall come from the factory with a battery and you shall never disconnect that battery nor allow it discharge. You certainly would find this implication ridiculous, but think about it. Some of the details can be simply deduced from "common knowledge" about CMOS structures, but that needs further thinking, and that hurts... :-) Richard Erlacher said:
AFAIK, no supervisor manufacturer has done any better with respect to these parameters, so you can't really say that insertion of a supervisor is, in any sense, a fix for the numerous problems that have traditionally been blamed on flaky RESET behavior of various 805x chips. Well, if the reset IC fixed these problems using any other mechanism than providing a valid RESET signal, real or esotheric, I don't really care - it did solve my problem that's all. Look, Richard, I can agree with you that it's "unscientific" to drop in the reset IC and forget the case - but it works. Maybe there is real work behind it and it just simply did not get published, or simply has been forgotten. There is quite a lot of such things in engineering - you either start from Maxwell's equations or similar first principle each time you want to solve something - and go crazy until you solve the simplest circuit; or you give in as all of us, stupid engineers, and work with assumptions and simplifications and paradigms, taken for granted without much ado (withing certain boundaries); all based on work of our predecessors, much of which is already forgotten, too. You know I am really curious and tried to get involved in this matter slightly more perhaps, and if I would have the resources (time, money, equipment) I would gladly go and dig as deep as needed to give you (and others) the only and definitive answer based on exact observations, simulations and calculations; but I don't think this will ever happen. And I also don't think any of us here will ever be in that position. So, please, give up, and quietly drop in that $1+ reset IC next to your sub-$1 '51. And, if anything goes wrong, please feel free to blame me, personally... But, first of all, please, don't take this matter SO seriously, relax, have fun... Jan |
