| ??? 02/28/07 17:12 Modified: 02/28/07 17:13 Read: times |
#133947 - Those were different times Responding to: ???'s previous message |
Erik Malund said:
Richard Erlacher said:
There may have been complaints, but I think I'd have heard about them. I started hearing about reset issues and about runaway controllers only after internal FLASH became popular. MAY??? I fail to count the number of RC reset devices I heard of (and I have seen some) "if it does not start, turn off power, wait a minute, turn it on again". That MAY have been acceptable, but I know of (industrial) cases where a RC reset caused major mishaps. Have you ever seen a pattern knitting machine going astray? were it not for the hours it takes to clear the mess, it would be funny (it is quite 'colorful'). I reworked it with a supervisor, and the yarn went where it should every time. Yes, if you suffer from supervisorfobia there may be ways around it e.g. a switch after the load electorolytics that only switch power on after it is stable AND has a supercap to help when glitches come and ...., and .... but is inserting the appropiate chip not a simpler and cheaper solution? Erik PS. The unfortunate thing is that the companies that uses uCs in the multimillions, and thus are the most interesting customers to the manufacturers, typically make 'toys' and thus you see the appnotes and datasheets with the stupid RC reset to indicate "if you use our chip it will be cheap" Hmmm ... you consider kitchen appliances, telephones, etc, to be toys? Supervisors didn't appear on any of Signetics' or AMD's or Intel's boards. I suspect your'e suffering from selective memory. Back in those days, i.e. the 80's, there were multiple sources for nearly everything. There were distributors who actually had to compete. There were manufacturers' sales offices, with REAL application engineers. If Motorola's guys had gotten hold of a major weakness sch as this, requiring a $2 supervisor for a $2 chip, do you really believe it wouldn't have been mentioned? I normally learned more about Motorla's products from the Intel or AMD guys than from the Motorola salesmen, and I was pretty close with all of 'em back then, since some of my clients used Motorola 6801's and 6805's while others used i804x and i805x MCU's. The Motorola and Hitachi parts always seemed quite happy with their 0.01 uF cap to gnd and 10 k-ohm pullup to Vcc. The question is, does that positive-going reset interact at all with the pierce oscillator in some peculiar way during a particularly slow power-up, that it fouls up the internal reset? Does it behave in precisely the same way with an external crystal oscillator? Has anybody ever actually doen a rigorous study of this behavior? Was it done by someone who didn't have a dog in the fight, i.e, someone who wasn't selling a product that was part of the controversy? If so, where are the details? RE |
