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

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
09/11/07 03:13
Read: times


 		 #144352 - This does not make him learn to swim...
Responding to: ???'s previous message
Richard said:
Is it possible that this problem stems from the fact that the MCU is from ATMEL?

Haha, it was only a question of time that you would formulate this...
No, Richard, in this forum there's a long history about failings with breadboards, not at all to do only with ATMEL's micros. Also, I don't think that it necessarily has to do with the oscillators traces. I think the whole set-up is improper, especially the power supply decoupling.

Richard said:
I do recall, however, that we were able to get crystal oscillators to work just fine back in the days when we allowed 5-10 cm traces between the crystal and the active components.

No, that's too long. I know of several cases, where more than 1...2cm resulted in problems. Not only caused by the inductance of traces, but also by the current loop formed and the associated radiation. If you are lucky, your design with the 10cm long traces might work, but all application notes, dataheets and other papers demand shortest traces! I would heavily recommend the use of shortest traces for the whole oscillator section. By the way, with your 10cm design you would heavily fail the CE radiation test!!!

Richard said:
I'm still thoroughly convinced that the problem lies in (1) the use of the solderless breadboard, and (2) the integrity of the connections between oscillator components, i.e, cap's, crystal, and MCU.

Rubem was just reporting, that his breadboard design works, but the PCB version fails. So, what about that?

Richard said:
The fact that the "breadboard" circuit changes its behavior from day to day shouldn't surprise anyone. No circuit built on one of these is entirely static.

Yes, breadboards are entirely unsuited for high speed designs or fast digital circuits, because they don't allow the use of a solid ground plane but introduce lots of complex impedances.

Rubem must show us his layout to be able to find the cause of his problems. More, we could recommend him a suited layout, what I tried to do in an earlier post.

Kai

List of 125 messages in thread
TopicAuthorDate
89C2051 oscillator does not oscillate (?)            01/01/70 00:00      
   Capacitor value?            01/01/70 00:00      
      20-40pF indeed            01/01/70 00:00      
         it might be the crystal            01/01/70 00:00      
            Probably not            01/01/70 00:00      
               because it\'s crap            01/01/70 00:00      
                  Not same position            01/01/70 00:00      
                     one cap or two makes most probably no difference..            01/01/70 00:00      
                        Not on the pins            01/01/70 00:00      
                           I'd recomend you re-read the spec's ...            01/01/70 00:00      
                           It's the same Vcc            01/01/70 00:00      
                  Hhm...I wouldn't recommend this paralleling            01/01/70 00:00      
                     hummm... really? In what frequency range?            01/01/70 00:00      
                        The inductance of the left resonating with...            01/01/70 00:00      
                           interesting            01/01/70 00:00      
                        Spice models            01/01/70 00:00      
                           Thanks, Kai.            01/01/70 00:00      
                           Numerical calculation, some examples...            01/01/70 00:00      
   Wrong reset scheme!            01/01/70 00:00      
      10µF reset capacitor to Vcc            01/01/70 00:00      
      Isn't that the mfg's recommended circuit?            01/01/70 00:00      
         Which do you mean?            01/01/70 00:00      
            10 uF to Vcc and 8.2K (built-in?) to GND            01/01/70 00:00      
               no, it is not            01/01/70 00:00      
                  If it's in nearly every datasheet ...            01/01/70 00:00      
                     this is your perception of the things, Richard            01/01/70 00:00      
                        Here we go again ...            01/01/70 00:00      
                           ... and again...            01/01/70 00:00      
                              I don't know what to say ...            01/01/70 00:00      
                                 there is NO out of spec power supply            01/01/70 00:00      
                                    Not only...            01/01/70 00:00      
                                       Not allways            01/01/70 00:00      
                                       Have you seen evidence of improvement?            01/01/70 00:00      
                                          Wider spectrum            01/01/70 00:00      
                                             maybe ratio of good/bad crystals is the same...            01/01/70 00:00      
                                                In all the various designs I produced ...            01/01/70 00:00      
                                                   Never had a failure until now            01/01/70 00:00      
                                                   I know of lots of crystal problems...            01/01/70 00:00      
                                                      Could be, but I've never encountered it ...            01/01/70 00:00      
                                                      %$#@! Double-post again ... ...            01/01/70 00:00      
                                       ^%$#@! double-post ...            01/01/70 00:00      
                                    If the datasheet says 1 ms ...            01/01/70 00:00      
                           more about locomotives            01/01/70 00:00      
                              symptoms suggest ...            01/01/70 00:00      
                     steam locomotives            01/01/70 00:00      
                        I like the steam locomotives            01/01/70 00:00      
                           So do I            01/01/70 00:00      
                              me too, said the dog            01/01/70 00:00      
                                 The kids should be smarter than that!            01/01/70 00:00      
                                    kids are kids            01/01/70 00:00      
   It might not be a clock problem            01/01/70 00:00      
      Does not work, sorry            01/01/70 00:00      
   Bad Connections or Crystal ?            01/01/70 00:00      
      A picture may be useful            01/01/70 00:00      
         Breadboards are highly unsuited for micros!!            01/01/70 00:00      
            Makes sense            01/01/70 00:00      
            breadboards are highly unsuited for ANYTHING!            01/01/70 00:00      
               Not at all!            01/01/70 00:00      
                  If you look at the pictures you can see why.            01/01/70 00:00      
                     Yes, it doesn\'t look properly, but...            01/01/70 00:00      
                        It's too risky!            01/01/70 00:00      
                           What is the purpose of your posts??            01/01/70 00:00      
                              My purpose is to point out that there are flaws            01/01/70 00:00      
         *** THE PICTURES ***            01/01/70 00:00      
            Problem may be breadboard            01/01/70 00:00      
               Breadboard correction            01/01/70 00:00      
                  How I programmed the 2051            01/01/70 00:00      
               Must be continuous            01/01/70 00:00      
               there is a good english word \'IS\"            01/01/70 00:00      
            I have the same thing running on my desk...            01/01/70 00:00      
               really ?            01/01/70 00:00      
                  Sadly, it's a trade name            01/01/70 00:00      
                  yea, but...            01/01/70 00:00      
                     again, replacing one uncertainty with another            01/01/70 00:00      
            I see two points            01/01/70 00:00      
   When you\'re ready to give up ... try this!            01/01/70 00:00      
      wow            01/01/70 00:00      
      use the unbuffered parts            01/01/70 00:00      
         It works fine with S, LS, HC, AC, HCT, ACT, etc            01/01/70 00:00      
            It's not so foolproof            01/01/70 00:00      
               I've had pretty good results ...            01/01/70 00:00      
   Many thanks, I'll try some of your tips...            01/01/70 00:00      
      Maybe your AT89C2051 is just damaged?            01/01/70 00:00      
      Overtone            01/01/70 00:00      
      89S51/52 are much more stable            01/01/70 00:00      
   Can you still (re-) program it            01/01/70 00:00      
      Let's hope that Rubem will come back ...            01/01/70 00:00      
         Hi again, sorry            01/01/70 00:00      
            Even with no breadboard            01/01/70 00:00      
               What has the programmer to do with the oscillator?            01/01/70 00:00      
      The chip is fine            01/01/70 00:00      
         What will that tell you about the oscillator?            01/01/70 00:00      
            It is time to tell the manufacturer, that...            01/01/70 00:00      
               oscillator of 2051?            01/01/70 00:00      
                  Sorry, maybe I was too harsh            01/01/70 00:00      
                     The MCU and crystals are all working for sure            01/01/70 00:00      
                        It was not the 7805            01/01/70 00:00      
                           because you shook/dropped/kicked it            01/01/70 00:00      
                        7805 decoupling            01/01/70 00:00      
                           LM79XX need a different treatment!            01/01/70 00:00      
                        Good news            01/01/70 00:00      
                           Is this >good news< ??            01/01/70 00:00      
                           Have you considered the "breadboard" effect?            01/01/70 00:00      
                              Inductance            01/01/70 00:00      
                                 Hey!            01/01/70 00:00      
                                    So what?            01/01/70 00:00      
                                       specifically            01/01/70 00:00      
                                       Let\'s not go overboard ...            01/01/70 00:00      
                                          This does not make him learn to swim...            01/01/70 00:00      
                                             I have only my own experience ...            01/01/70 00:00      
                                          Not Atmel problem            01/01/70 00:00      
                                          things change            01/01/70 00:00      
                                             looking at the wrong characteristic            01/01/70 00:00      
                                                OT: to Richard            01/01/70 00:00      
                                                   OT: I hadn\'t but I have ... Yes, I\'m interested!            01/01/70 00:00      
                                       You are probably right            01/01/70 00:00      
                        do you have a 10-100nF directly across the            01/01/70 00:00      
                           *** PCB and schematic ***            01/01/70 00:00      
                              I cannot find C1            01/01/70 00:00      
                                 Ahh, it is the supercap! Needs some changes!            01/01/70 00:00      
                                    Surprisingly It works like it is            01/01/70 00:00      
                                       No, no...            01/01/70 00:00      
                              I can't really 'read' your PCB but ...            01/01/70 00:00      
                                 It is connected            01/01/70 00:00      
   Is any old recipe.            01/01/70 00:00      

Back to Subject List