Measure small time periods with 1ns resolution

Discussion in 'General Electronics Chat' started by masteropie, Oct 3, 2008.

  1. masteropie

    Thread Starter Member

    May 17, 2008
    26
    0
    I am looking to measure around 1us time periods with a pic down to 1ns resolutions for experimenting with lasers and ultrasonics. I have an idea, seams fairly simple.

    Sample at 1ghz and use some sort of serial to parallel shift register to change the data into 16bit so it can be entered into 16bit ram every 16 clock cycles. Then the pic can take its time and measure the string of 1s in ram. For example: 00000011111111000000000 would be 8 nanoseconds.

    It seams like it world work to me. Just i have no idea know how to make 1gb/s serial into 16bit parallel. Everything available seams like it is way to slow, I must be looking at the wrong thing.
     
  2. scubasteve_911

    Senior Member

    Dec 27, 2007
    1,202
    1
    I found some 9-bit serial-parallel converters that can easily run that fast.

    MC100EP142

    OnSemi and Micrel make some pretty fast ECL logic.

    Your ram idea seems a bit complex, you would know this if you ever interfaced to a RAM chip. You need a controller to run fast ram, because it needs refreshing, etc.

    Typically, they use very fast counters with triggers to measure quick events. So, you'd run a counter at over 1GHz and trigger it when you are timing something, then shut it off, then figure out how many pulses it took knowing your frequency, which gives you the time.

    Steve
     
  3. masteropie

    Thread Starter Member

    May 17, 2008
    26
    0
    Yes, a counter does sound much easier. However i can't find any 20bit+ counters that can operate at 1ghz. Would a 9 bit shift register then a 9 bit counter be used?
     
  4. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    A 9 bit shift register can store 9 bits. A 9 bit counter can store 512 counts. Usage is usually either/or.
     
  5. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    Can you provide us with an explanation of what you are trying to measure that needs to have 1E-9 seconds?

    hgmjr
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Our OP apparently wants to get close to 1 meter of resolution using a laser (ballpark).

    Using such frequencies with ultrasonics would be somewhat pointless, as the speed of sound is so slow in comparision that even a large array of counters or shift registers would've shifted out by the time the return came back.
     
  7. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    Yep. The use of ultrasonics to obtain nanosecond resolution would be like using a yardstick to measure the width of a human hair.

    hgmjr
     
  8. masteropie

    Thread Starter Member

    May 17, 2008
    26
    0
    I would like to measure other things also. Possibly a ultrasonic flow meter also which would require 5ns resolution for 1cm/s with a transducer spacing of 5cm. Just would like this as a general tool.

    That is why i was thinking ram since a simple chip could store 16 million counts. 24bits of storage should also be able to hold 16 million counts. It guess it is more complicated then it seams due to the speed.
     
  9. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    One centimeter per second flow measurement with 5 nanoseconds resolution? Sounds a bit like measuring miles with a micrometer.
     
  10. scubasteve_911

    Senior Member

    Dec 27, 2007
    1,202
    1
    You can cascade the 9bit counter that I mentioned. I was actually thinking about doing a laser TOF system with this stuff before, but decided against it. I wanted better resolution and it wasn't possible with what I could get my hands on.

    Steve
     
Loading...