+ Post New Thread
Results 1 to 12 of 12
  1. #1
    Newbie level 4
    Points: 536, Level: 4

    Join Date
    Sep 2016
    Posts
    5
    Helped
    0 / 0
    Points
    536
    Level
    4

    [moved] Digital Transition Capture

    Hi everyone,

    I need a circuit is designed to detect whether there has been at least one transition during the past clock period. This image show detail the operating principle of the circuit. I am not specialized in digital, so i don't know this circuit is possible? Please give me some information about the circuit.

    Click image for larger version. 

Name:	Capture.PNG 
Views:	14 
Size:	6.0 KB 
ID:	153571

    Thanks you!

    •   AltAdvertisement

        
       

  2. #2
    Advanced Member level 1
    Points: 2,489, Level: 11

    Join Date
    Aug 2016
    Posts
    441
    Helped
    69 / 69
    Points
    2,489
    Level
    11

    Re: Digital Transition Capture

    Hi,

    A simple FSM will do it.

    You need two pipelined registers, reg1 and reg2. Keep the current data in reg1. On the next clock edge, move the content of reg1 to reg2. Always compare each bit in reg1 with corresponding bit in reg2 using XOR gate. You'll get a '1' when the contents are different. You'll get a '0' otherwise.

    - - - Updated - - -

    OR the outputs from all the XOR gates.
    Last edited by Akanimo; 8th June 2019 at 20:02.
    -------------
    --Akanimo.



  3. #3
    Super Moderator
    Points: 51,966, Level: 55

    Join Date
    Apr 2011
    Location
    Minneapolis, Minnesota, USA
    Posts
    12,746
    Helped
    2539 / 2539
    Points
    51,966
    Level
    55

    Re: Digital Transition Capture

    I wonder whether some kind of flip-flop detector would work. However I'm not certain whether you can get it to it work on a single up or down transition, or whether it requires an entire cycle.

    From the opposite standpoint what does your output do? It stays high most of the time, and it drops when it sees no transition during a clock cycle. Therefore it might also be called a missing pulse detector. A 555 timer IC can be made to do that.



  4. #4
    Newbie level 4
    Points: 536, Level: 4

    Join Date
    Sep 2016
    Posts
    5
    Helped
    0 / 0
    Points
    536
    Level
    4

    Re: Digital Transition Capture

    Hi Akanimo,
    Thanks for your help.
    This circuit work with clock frequency is smaller than the input signal frequency.
    Therefore, I think it is impossible to use the your method.

    Hi BradtheRad,
    Thanks for your help. The method use 555 timer IC is very suitable for my case. But I can't use the IC 555, I need a integrated circuit in my IC. Do you have any other method to design a missing pulse detector?
    Thanks,



    •   AltAdvertisement

        
       

  5. #5
    Advanced Member level 1
    Points: 2,489, Level: 11

    Join Date
    Aug 2016
    Posts
    441
    Helped
    69 / 69
    Points
    2,489
    Level
    11

    Re: Digital Transition Capture

    What are your frequencies? List them so we know. Even the 555 has limitation in frequency.

    Is that the original clock or a derived clock?
    -------------
    --Akanimo.



    •   AltAdvertisement

        
       

  6. #6
    Super Moderator
    Points: 257,269, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    44,887
    Helped
    13656 / 13656
    Points
    257,269
    Level
    100

    Re: Digital Transition Capture

    The "at least one transition" point needs clarification. There can be potentially many transitions with high repetition frequency, under circumstances they may be detected as zero transitions.



  7. #7
    Newbie level 4
    Points: 536, Level: 4

    Join Date
    Sep 2016
    Posts
    5
    Helped
    0 / 0
    Points
    536
    Level
    4

    Re: Digital Transition Capture

    Hi,
    We need to operate at the frequency greater than 10MHz. Is the IC 555 can work in this case? Is this depend on the technology?
    I wonder if I use 65nm CMOS technology with the same topology of IC555, can I obtain the expected operation frequency?

    Thanks,
    Chau



  8. #8
    Advanced Member level 1
    Points: 2,489, Level: 11

    Join Date
    Aug 2016
    Posts
    441
    Helped
    69 / 69
    Points
    2,489
    Level
    11

    Re: Digital Transition Capture

    10MHz is too high for a 555. 3-MHz is specified for the cmos version.

    BTW, when you say 10MHz, is it the clock frequency or the data. You didn't say which and you didn't mention the other.
    -------------
    --Akanimo.



  9. #9
    Newbie level 4
    Points: 536, Level: 4

    Join Date
    Sep 2016
    Posts
    5
    Helped
    0 / 0
    Points
    536
    Level
    4

    Re: Digital Transition Capture

    Quote Originally Posted by Akanimo View Post
    10MHz is too high for a 555. 3-MHz is specified for the cmos version.

    BTW, when you say 10MHz, is it the clock frequency or the data. You didn't say which and you didn't mention the other.
    Let see the figure below. In which, the clock frequency is 30 MHz, the input data has 0's pulse width is 5ns.
    Click image for larger version. 

Name:	viber_image_2019-06-10_19-46-19.jpg 
Views:	5 
Size:	183.5 KB 
ID:	153604



    •   AltAdvertisement

        
       

  10. #10
    Super Moderator
    Points: 257,269, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    44,887
    Helped
    13656 / 13656
    Points
    257,269
    Level
    100

    Re: Digital Transition Capture

    Your latest waveform makes less sense than the post #1 waveform which changed the output on the rising clock edge.

    Please specify the detection algorithm you want to implement unambiguously, e.g. as a state diagram.



  11. #11
    Advanced Member level 1
    Points: 2,489, Level: 11

    Join Date
    Aug 2016
    Posts
    441
    Helped
    69 / 69
    Points
    2,489
    Level
    11

    Re: Digital Transition Capture

    You have mentioned too many frequencies: 10MHz, 30MHz, and the "input" waveform of your last post shows a frequency that is about double that of the 30-MHz clock.

    Anyways, the idea is that the clock is in tens of MHz.

    I asked you earlier on how you got the clock signal, whether it is a primary clock or a generated clock but you didn't give any response in that regard.

    What I think you can do right now is to use a clock multiplier (a PLL would be sufficient) to generate a fast enough clock signal from your 30-MHz clock so that you can use my approach of Post #2.

    I noticed that your output logic level of Post #10 is reversed compared to that of Post #1. This means you would replace the XOR gate with an XNOR gate.

    - - - Updated - - -

    But this means that the generated clock frequency has to be faster than 1/5ns ( =200MHz).
    -------------
    --Akanimo.



  12. #12
    Member level 3
    Points: 2,053, Level: 10
    Achievements:
    7 years registered

    Join Date
    Dec 2009
    Location
    USA
    Posts
    54
    Helped
    15 / 15
    Points
    2,053
    Level
    10

    Re: [moved] Digital Transition Capture

    Attached is a very "rough" idea for Latch & Hold

    1) The Data Signal is converted from "H-to-L" & "L-to-H" transitions to very narrow Pulses - ( width needs verified )
    2) One, or more, Data Pulses will SET the R-S Latch, asynchronously.
    3) The rising edge of the Clock triggers the D-Latch to transfer the state of the R-S Latch to the OUTPUT Pin.
    4) The rising edge of the Clock also RESETS the R-S Latch - ( I need to SYNC this with a new Data Transition Pulse )
    5) The RESET of the R-S latch can be blocked by an active Transition of the Data Signal - ( currently very poorly )

    The logic to handle a "simultaneous" Data Transition and a Rising Clock Edge RESET needs to fixed by making them always sequential.


    Click image for larger version. 

Name:	Latch-Hold.jpg 
Views:	8 
Size:	174.1 KB 
ID:	153612
    Last edited by summitville; 10th June 2019 at 19:47.



--[[ ]]--