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

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
12/08/09 06:02
Read: times


 
#171515 - ... glad you asked that ...
Responding to: ???'s previous message
Erik Malund said:
Have you ever observed what your switcher does when the load current goes from 20% of rated current to 80% of rated current? If so, what did you observe?
yes, I have and nothing much

By that do you mean that you didn't look?

Have you ever observed what your linear does when the load current goes from 20% of rated current to 80% of rated current? If so, what did you observe?

I just did, since I had a broken gadget in house (diode failure), and it drops about 45 mV for about 2 microseconds, rings as much as 50 mv for 6 microseconds, and then settles back to 5.0087 volts +/- about 5 mV. It does pretty much the same thing when the load is switched off though the waveform is a mite different. Normal noise level under average load is about 15 mV when the MCU is running. That's on the 5-volt supply rated at 10 amperes, regulated with a 78P05 in TO-3. I'd expect somewhat lower performance from an LDO, as its output transistor has less gain. The 'P05 has a darlington output, as does the 7805 and the LM340. With a pass transistor, the thing is a little less simple, hence a bit slower to respond to load changes.

When I last subjected a switcher rated at 10 amperes to the same load (stepping motor) driven from Vcc, it burst into flame. I did it again at the time, and got the same result. I've never bought a switcher from that supplier again. I also don't use switchers to feed steppers bigger than a small apple any longer.

I suspect that sort of thing happens pretty often in your bus-display environment, so I imagine you've looked at it quite a bit. Of course, it's probably not particularly important, since your display can probably recover within a few milliseconds if the supply drops out. Nevertheless, you've probably seen the worst-case.
since a color display have one 'dose' of 1/2ms the maximum allowable 'dip below min' is in the 10us range for the LED supply. Of course, for the logic supply it is zero.

Erik

dose: (could not find a better name) one way to create color is to display one display for 4ms, another display for 2ms, another display for 1ms, another display for 1/2ms. This gives 15 possible intensities each for red, green and blue



I wasn't so much interested in a "dip-below-min", as, generally, that's never tolerable in Vcc. When you say, "zero," what do you mean? Are you saying that your switcher has zero ripple? Are you saying that your Vcc supply always has a precisely constant load? Are you saying that there's no noise riding on Vcc? I'd imagine that there's something.

List of 77 messages in thread
TopicAuthorDate
a simple 89s52 based led circuit on the breadboard            01/01/70 00:00      
   22uF?            01/01/70 00:00      
      He's probably wondering what GNG is            01/01/70 00:00      
         Yep, I also wonder.... Already edited!            01/01/70 00:00      
      yes my bad, its 22pf not uf            01/01/70 00:00      
         Lots of nice info in datasheets and application notes            01/01/70 00:00      
         your problem identified            01/01/70 00:00      
         Yes...            01/01/70 00:00      
   I see no decoupling            01/01/70 00:00      
   Some hints...            01/01/70 00:00      
   similar problems when i was an infant enthusiast            01/01/70 00:00      
      NO!            01/01/70 00:00      
         a simpler solution            01/01/70 00:00      
            Of course...            01/01/70 00:00      
   Will post results in a week or soo            01/01/70 00:00      
      Maybe it is your program            01/01/70 00:00      
         How would that happen?            01/01/70 00:00      
            wandering into Enchanted Forest            01/01/70 00:00      
               about those test spec's            01/01/70 00:00      
                  Thorough testing? Often an illussion!            01/01/70 00:00      
                     testing is an illusion            01/01/70 00:00      
                        Agreed, with caveats            01/01/70 00:00      
                           Still same problem - no problem seen doesn't prove anything            01/01/70 00:00      
                              I've been saying that for years ...            01/01/70 00:00      
                                 for the umpteenth time            01/01/70 00:00      
                                    You should learn how to spell those words you like            01/01/70 00:00      
                                       I quoted you            01/01/70 00:00      
                                          Yes, but you "tracked down" nothing!            01/01/70 00:00      
                                             I recall the previous occasions ...            01/01/70 00:00      
                                                Focus on the circumstances            01/01/70 00:00      
                                                   'investigation' vs 'reasoning'            01/01/70 00:00      
                                                      Reasoning vs. investigation was popular for millenia            01/01/70 00:00      
                                                         Reasoning            01/01/70 00:00      
                                                            Ptolemy and Eratosthenes...            01/01/70 00:00      
                                                               seek to find game: facts, evidence, philisophy            01/01/70 00:00      
                                                                  Reasoning works fine if you don't omit/ignore the evidence            01/01/70 00:00      
                                 I only do what the manufacturer urged me...            01/01/70 00:00      
                                    Yes, Kai, you did that            01/01/70 00:00      
                                       can't you read            01/01/70 00:00      
                                          That's not what you were complaining about!            01/01/70 00:00      
                                       Please read this document...            01/01/70 00:00      
                                           I read that the first time you mentioned it.            01/01/70 00:00      
                  not perfect and never perfectly debugged            01/01/70 00:00      
                     It's the difference between "good" and "good enough"            01/01/70 00:00      
            Not necessary...            01/01/70 00:00      
               Neither way fixes anything.            01/01/70 00:00      
                  it might not be 'harmless'            01/01/70 00:00      
                     Watchdog should be used for anything semicritical            01/01/70 00:00      
                  Of ourse, illjumping must not happen!!!            01/01/70 00:00      
                     What "depending on"?            01/01/70 00:00      
                        Relax, not you...            01/01/70 00:00      
                           What do you use to "pet" the dog?            01/01/70 00:00      
                              hw timer isr för watchdog kick is dangerous            01/01/70 00:00      
                              Different instructions in different parts of the code!            01/01/70 00:00      
                              why use ALE?            01/01/70 00:00      
                                 Why indeed?!            01/01/70 00:00      
                                    The REAL problem lies in what it tells you.            01/01/70 00:00      
                                       please find just one that ...            01/01/70 00:00      
                                          the typical switcher ripple is too much            01/01/70 00:00      
                                             Have you ever observed            01/01/70 00:00      
                                                ... glad you asked that ...            01/01/70 00:00      
                                                   Only you, Richard            01/01/70 00:00      
                                                      that's not what you were asked, Erik!            01/01/70 00:00      
                                                         Nothing much            01/01/70 00:00      
                                                            We can let others conclude what they will            01/01/70 00:00      
                                 because ...            01/01/70 00:00      
                                 It's what the mfg tells us to use ...            01/01/70 00:00      
   Problem solved!            01/01/70 00:00      
      correction            01/01/70 00:00      
      Driving car with closed eyes...            01/01/70 00:00      
         sleep driving            01/01/70 00:00      
            Of course!!!            01/01/70 00:00      
      thanks for WHAT            01/01/70 00:00      
      No ... You only THINK it has been solved.            01/01/70 00:00      
         question for wire wrapping and some clarification            01/01/70 00:00      
            Rectangular pins that cuts into the wire            01/01/70 00:00      
            If you were to adopt this technique ...            01/01/70 00:00      

Back to Subject List