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

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
Michael Karas
02/20/12 19:56
Read: 1087 times
Beaverton Or
USA


 
#186086 - You need some specs first.
Responding to: Mahmood Elnasser's previous message
Before deciding on a design approach you will need to determine some set of specs for what you are trying to do. You started this by specifying the frequency range of the signals but there are some important things missing. Some things that need to be specified are:

a) How fast can the frequency of the input change?

b) Will there be step changes of the input or will it slew slowly?

c) How fast does the output geared up frequency need to track the input?

d) What accuracy does the output to input ratio have to maintain?

e) Does the output frequency need to be edge synchronized with the input?

Yes, it could be a very effective design to measure the input period and then produce an 10X output frequency with a timer on an MCU. Several things to look for in my book are to look for period measurement for the input that gates from edge to edge in hardware instead of having a timer interrupt latency be part of the measurement count. Another thing is to use a timer of the output that supports a full count reload capability similar to a Timer 2.

You may also need to do a little bit of filtering of the input measurements so as to prevent cycle to cycle dither in the output.

If you find an MCU with a 16-bit counter for the output at 10Hz you will need a counter that is clocked by a maximum of 655.3KHz. This will determine the frequency steps that you can attain at the output. A larger counter would afford use of a higher rate clock and permit better resolution.

Michael Karas


List of 52 messages in thread
TopicAuthorDate
frequency multiplier x 10      Mahmood Elnasser      02/20/12 14:29      
   You need some specs first.      Michael Karas      02/20/12 19:56      
      and even more questions for your tutor        Jim Granville      02/21/12 12:42      
   Analog solution...      Kai Klaas      02/21/12 05:49      
      Input wave form      Per Westermark      02/21/12 06:09      
      Too Bad...      Michael Karas      02/21/12 06:16      
   have you considered a "rate multiplier"?      Richard Erlacher      02/21/12 08:18      
   PCA?      Erik Malund      02/21/12 12:51      
      PCA Yes      Michael Karas      02/21/12 15:38      
   Update      Mahmood Elnasser      02/22/12 00:43      
      2 i.c.s plus 6 R, C      David Prentice      02/22/12 03:03      
      4046 / 7490      Michael Karas      02/22/12 03:36      
      I put my bet in the digital solution      Daniel Contarino      02/22/12 13:33      
         digital solution      Mahmood Elnasser      02/22/12 13:46      
            MCU / CPLD      Michael Karas      02/22/12 13:57      
            Simple digital solution...      Kai Klaas      02/23/12 05:59      
               Engine too fast      Mahmood Elnasser      02/23/12 06:48      
                  it could be made simpler        Jim Granville      02/29/12 19:22      
                     cute      Erik Malund      03/01/12 08:07      
               OR      Erik Malund      02/23/12 06:52      
   Frequency Multipiler      Satish S Munot      02/29/12 10:06      
      Motivation?      Per Westermark      02/29/12 10:14      
         Motivation ?????????      Satish S Munot      02/29/12 10:55      
            Motives      Per Westermark      02/29/12 11:24      
               OK      Satish S Munot      02/29/12 12:40      
                  motivation guess      Maarten Brock      03/02/12 01:14      
                     I had the same feeling...      Kai Klaas      03/02/12 08:25      
                        and if you looked at any of his previous posts ...      Richard Erlacher      03/02/12 20:02      
   A number of thins concern me ...      Richard Erlacher      02/29/12 12:31      
      answers      Mahmood Elnasser      02/29/12 14:45      
      Boring...      Kai Klaas      02/29/12 16:05      
         And in this case...      Daniel Contarino      02/29/12 18:11      
            Yes ... but ...      Richard Erlacher      02/29/12 20:12      
               Making what you need, from what you can get..      Jim Granville      02/29/12 22:14      
                  slogans      Satish S Munot      02/29/12 23:59      
                  Is that what the display needs?      Richard Erlacher      03/01/12 06:38      
                     10-700 Hz is expected output frequency - not input frequency      Per Westermark      03/01/12 06:57      
                        I did make one mistake      Richard Erlacher      03/01/12 07:18      
                           old rpm analog (or digital) display is what they have      Jim Granville      03/01/12 12:01      
                              bursts...      Kai Klaas      03/01/12 13:19      
                                 See LM2917 for ancient tacho examples      Jim Granville      03/01/12 13:40      
   PLL      Mahmood Elnasser      03/18/12 06:57      
   More detail?      Richard Erlacher      03/18/12 12:30      
      Yes, more details, please...      Kai Klaas      03/18/12 12:43      
         Details      Mahmood Elnasser      03/31/12 00:24      
            This can have to do ...      Kai Klaas      03/31/12 09:25      
               sharp edges      Mahmood Elnasser      04/01/12 03:59      
                  Which phase-detector?      Richard Erlacher      04/01/12 08:21      
               re: this can have      Andy Peters      04/02/12 10:49      
                  What I would do...        Kai Klaas      04/03/12 07:40      
                     Very Low Frequency      Mahmood Elnasser      04/19/12 15:30      
                        Give it a try...      Kai Klaas      04/19/12 17:14      

Back to Subject List