Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
08/01/07 16:21
Read: times


 		 #142594 - Not so fast, there, Pilgrim ...
Responding to: ???'s previous message
Erik

Erik Malund said:
So the very common 10µF + 8.2K might not work well with crystals other than 12Mhz.
who cares, ANY RC reset "might not work well".

Erik


I'd be very interested, first of all, in what you define as "working well" along with how, exactly, you define it.

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.

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.

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.

From what I've gathered from several years' observation of traffic in this forum, the typical case is that someone builds a single circuit with the manufacturer-recommended 8.2k-ohm pulldown (internal or external) in combination with an external 10 uF cap to Vcc, and subsequently finds it "doesn't work" in some sense. That's ultimately blamed on the RC reset. He/she then inserts a supervisor of some sort into the circuit, subsequently observes different behavior, and declares the problem to be solved.

It looks to me like a "leap of faith" rather than rigorous study.

Has anyone ever performed actual comparative testing? If so, why haven't they published the test conditions and results?

From what I can see, we're all walking on very thin ice believing what we're told, as opposed to what we read in published data. No MCU maker has yet published a specification that he guarantees will characterize his product with respect to RESET behavior under specified conditions of Vcc risetime, brownout, and fall time. No supervisor manufacturer has gone out on a limb and specified the behavior of his device under those conditions with sufficient rigor either. Of course, so long as MCU makers don't specify the precise behavior of their product, the supervisor manufacturer can't guarantee a proper RESET or much of anything else.

Now, I can't say whether these RC resets work properly or not, since they're not well characterized. There do seem to be a large number of failure conditions attributed to them, though. I can't claim, either, for the same reason, that supervisors of one sort or another do any better, though there seem to be lots of arguments that they do.

Several people have stated that they have their circuit with supervisor in the field, in thousands, and have had no problems. Nearly all MCU makers, however, continue to state that their parts work fine with the RC reset. I don't know who's telling the truth and who's not, or even that the claims are conflicting. How can we KNOW what's really the case?

RE






List of 189 messages in thread
TopicAuthorDate
Reset Accuracy            01/01/70 00:00      
   Read This            01/01/70 00:00      
   throw it away!            01/01/70 00:00      
   That is the nature of RC resets!            01/01/70 00:00      
   Inhereted design            01/01/70 00:00      
   1st Lesson            01/01/70 00:00      
      P.S. Welcome            01/01/70 00:00      
      is IS an iteresting question, isn't it?            01/01/70 00:00      
         more on supervisor            01/01/70 00:00      
            same thing            01/01/70 00:00      
            Extendec ASCII Characters            01/01/70 00:00      
               That's cool Joe            01/01/70 00:00      
               Not ASCII            01/01/70 00:00      
               Beware - Your results may vary            01/01/70 00:00      
         Thoughts on RC Resets            01/01/70 00:00      
            Schmitt leaves problems?            01/01/70 00:00      
               even at startup you can get into trouble            01/01/70 00:00      
                  the decline and f.. naah rise of the \'51            01/01/70 00:00      
                     the lpc9xx is the glaring BAD example...            01/01/70 00:00      
                        it's not about NXP, though I woldn't forgive them            01/01/70 00:00      
                           it's far past that time...            01/01/70 00:00      
                     only partially so            01/01/70 00:00      
            not quite            01/01/70 00:00      
            Lynn ... can you see anything wrong with ...            01/01/70 00:00      
               how about shooting a sparrow with a RPG?            01/01/70 00:00      
                  You've got to ask yourself ...            01/01/70 00:00      
                     Which manufacturers?            01/01/70 00:00      
                        everybody is            01/01/70 00:00      
                           errare humanum est...            01/01/70 00:00      
                        It's up to us to see that they fix it!            01/01/70 00:00      
                  What's the point?            01/01/70 00:00      
                     I do noy 'believe' I KNOW.            01/01/70 00:00      
                        You can't prove a negative ...            01/01/70 00:00      
                           when it's been gone for 100,000 "unit hours"            01/01/70 00:00      
                              correction            01/01/70 00:00      
                                 How do you monitor this?            01/01/70 00:00      
                                    how stupid can you be            01/01/70 00:00      
                                       so YOU really haven't ever looked at any of them?            01/01/70 00:00      
                                          sure I did            01/01/70 00:00      
                                             it's just as I said ... YOU guess and ship ...            01/01/70 00:00      
                                                Same question            01/01/70 00:00      
                                                   You force them in your test fixture            01/01/70 00:00      
                                                that I do not want to do my job is an insult            01/01/70 00:00      
                                                   It's always about your ego, isn't it, Erik?            01/01/70 00:00      
                                                      adding insult to injury            01/01/70 00:00      
         make it opposite            01/01/70 00:00      
            Answer is wrong            01/01/70 00:00      
               No, is not!            01/01/70 00:00      
                  RE: No, is not!            01/01/70 00:00      
                     Oops.            01/01/70 00:00      
                        the "small embedded printf"...            01/01/70 00:00      
                           LED blinking            01/01/70 00:00      
                              LCD init delay            01/01/70 00:00      
            possibly (most likely) incorrect post            01/01/70 00:00      
      supervisor IC            01/01/70 00:00      
   Check reset Schematic and capacitor .            01/01/70 00:00      
      not necessarily            01/01/70 00:00      
   RC time            01/01/70 00:00      
      crystal frequency is a negligible problem            01/01/70 00:00      
      who cares            01/01/70 00:00      
         Not so fast, there, Pilgrim ...            01/01/70 00:00      
            reset toujours            01/01/70 00:00      
               correction            01/01/70 00:00      
                  I didn\'t know this is the price of a bit...            01/01/70 00:00      
                     the origin of the \'bit\'            01/01/70 00:00      
                        Almost but not quite ...            01/01/70 00:00      
                           wasn't that a gold coin?            01/01/70 00:00      
                  So you assert that that's the one that works?            01/01/70 00:00      
               What is this "reset problem" solved by supervisor            01/01/70 00:00      
                  the final effect is on peripherals not the micro            01/01/70 00:00      
                  not, "this" but all            01/01/70 00:00      
                     What did you mean to say, Erik?            01/01/70 00:00      
                        proof? many ways, here is an example            01/01/70 00:00      
                           I don't doubt any of that, but ...            01/01/70 00:00      
                              the skinny            01/01/70 00:00      
                                 just another reset-related story            01/01/70 00:00      
                        true reset-related stories            01/01/70 00:00      
                           Relevance?            01/01/70 00:00      
                              yes, relevance.            01/01/70 00:00      
                                 the second example ...            01/01/70 00:00      
                                    I gave you two SOLID examples            01/01/70 00:00      
                                       I agree with Erik and Jan            01/01/70 00:00      
                                          Start at the end ...            01/01/70 00:00      
                                          I just think you're all fooling yourselves ...            01/01/70 00:00      
                                       Vexing, perhaps. Solid? Not in the least!            01/01/70 00:00      
                                       What\'s so \"solid\" about an anecdote?            01/01/70 00:00      
                                          Who gives a hoot            01/01/70 00:00      
                                    let\'s put it in another way            01/01/70 00:00      
                                       Here\'s what I want from you ...            01/01/70 00:00      
                                          A particular experience            01/01/70 00:00      
                                 Just consider THIS thread ...            01/01/70 00:00      
                                    well, what about this            01/01/70 00:00      
                                       there's a difference between a fake and a fix            01/01/70 00:00      
                                          I believe you are the only one with that opinion            01/01/70 00:00      
                                             So, who has actually tested their product?            01/01/70 00:00      
                                                incoirrecrt            01/01/70 00:00      
                                    my experience with RC reset            01/01/70 00:00      
                                       Invalid conclusion            01/01/70 00:00      
                                          here is YOUR misconception            01/01/70 00:00      
                                             You\'ve got it wrong ...            01/01/70 00:00      
                                                I can recreate this            01/01/70 00:00      
                                                   maybe you can ...            01/01/70 00:00      
      IN1232N            01/01/70 00:00      
         something is wrong            01/01/70 00:00      
         MAX1232 is better!            01/01/70 00:00      
   What do you mean by "erratic and random"?            01/01/70 00:00      
   RE: something is wrong            01/01/70 00:00      
      Would you be please so kind...            01/01/70 00:00      
         RE: Would you be please so kind...            01/01/70 00:00      
            Hhm,...            01/01/70 00:00      
      What\'s the power supply doing at power-on?            01/01/70 00:00      
   8051 Family Reset Function            01/01/70 00:00      
      Power ON reset            01/01/70 00:00      
         defined power on            01/01/70 00:00      
            inadequate power supply            01/01/70 00:00      
               what does that have to do with            01/01/70 00:00      
                  it has everything to do with it.            01/01/70 00:00      
                     Reset            01/01/70 00:00      
                        ... and what does it do as Vcc decays?            01/01/70 00:00      
                           Defence            01/01/70 00:00      
                              You needn't defend the party line ...            01/01/70 00:00      
                                 correct if you are NOT using a supervisor            01/01/70 00:00      
                                    bidir reset, if there is an internal supervisor            01/01/70 00:00      
                                       Oh, that's what 'bidirectional' means here            01/01/70 00:00      
                                          how do you call that property?            01/01/70 00:00      
                                          Maybe that's your RESET problem!            01/01/70 00:00      
                                             i do not            01/01/70 00:00      
                                                You don't what?            01/01/70 00:00      
                                                   I am sorry, Richard            01/01/70 00:00      
                                                      You do that often ...            01/01/70 00:00      
                                    lower            01/01/70 00:00      
                                    IF you're using too weak a PSU            01/01/70 00:00      
                                       another red herring            01/01/70 00:00      
                                          Well maybe it's not a red herring after all            01/01/70 00:00      
                              Ricxhard, posted            01/01/70 00:00      
                                 supervisor is irrelevant during Vcc decay            01/01/70 00:00      
                                    BULLSHIT            01/01/70 00:00      
                                       I've observed it, while you haven't even looked            01/01/70 00:00      
                     unpredictible is a strong word            01/01/70 00:00      
                        just how would YOU predict the behavior, Jan?            01/01/70 00:00      
                           CMOS            01/01/70 00:00      
                              Yes, that's fine ... well, maybe ...            01/01/70 00:00      
                     male cow manure            01/01/70 00:00      
                        and what's the value of the internal pulldown?            01/01/70 00:00      
               ... and who\\\'ll pay the bills?            01/01/70 00:00      
                  Not only that, but...            01/01/70 00:00      
                  It's too costly if it doesn't work            01/01/70 00:00      
                     get me a muckraker and a pair of boots            01/01/70 00:00      
                        Well, I've observed it, and you haven't even tried            01/01/70 00:00      
      Flash micros need a different treatment!!!            01/01/70 00:00      
         to Kai KISS            01/01/70 00:00      
            Of course, good alternative!            01/01/70 00:00      
         Why do the manufacturers not say that?            01/01/70 00:00      
            one is one            01/01/70 00:00      
            we have answered this many many times before            01/01/70 00:00      
               we really don't disagree ... at least on this            01/01/70 00:00      
                  sure you can            01/01/70 00:00      
            The same story again and again and again...            01/01/70 00:00      
               Clearly, we see things differently            01/01/70 00:00      
                  We aren't so differently...            01/01/70 00:00      
                     I have pulled every string I have ....            01/01/70 00:00      
                        That would be really awesome!!            01/01/70 00:00      
                     it would be awesome            01/01/70 00:00      
                        then I'll call the whole thing off            01/01/70 00:00      
                           No, that's not it            01/01/70 00:00      
                              Would this solve your problem?            01/01/70 00:00      
                                 What do you mean?            01/01/70 00:00      
                                    more details please            01/01/70 00:00      
                                       Maybe this answers some questions            01/01/70 00:00      
                                          Blocking the BBRAMs            01/01/70 00:00      
                                             Why?            01/01/70 00:00      
                                                certainly            01/01/70 00:00      
                                                You asked for a methode, and I showed you one            01/01/70 00:00      
                                                   I agree, what you suggest might work            01/01/70 00:00      
                                                      slowpoke            01/01/70 00:00      
                                                      Because of the enormous currents?            01/01/70 00:00      
                                                         I considered that ...            01/01/70 00:00      
                                                            Of course!!            01/01/70 00:00      
                                                               I've been looking for simple ways            01/01/70 00:00      
                                                more suspition - and more questions            01/01/70 00:00      
                                                   It wasn't a thorough test            01/01/70 00:00      
                                          more questions...            01/01/70 00:00      
                                             I don't understand the questions ...            01/01/70 00:00      
                                                the questions are intended...            01/01/70 00:00      
                                                   there are a number of forces at work here            01/01/70 00:00      
                        Take care....            01/01/70 00:00      
                           You pointed that out once before ...            01/01/70 00:00      
                        We did, of course!            01/01/70 00:00      
                           That\'s probably an adequate statistical basis            01/01/70 00:00      

Back to Subject List