Shop OBEX P1 Docs P2 Docs Learn Events
Output rise and fall time — Parallax Forums

Output rise and fall time

CRST1CRST1 Posts: 103
edited 2008-09-26 05:07 in Propeller 1
Does anyone know the specs for the rise and fall time of the propeller outputs?· When I measure them I get about 12 usec. If this is right then it looks like the fastest output out of the pins would be about 40khz. Can anyone verify this or tell me I'm wrong. How fast can the propeller output a signal?

Thanks

Comments

  • LeonLeon Posts: 7,620
    edited 2008-09-25 20:52
    You are obviously measuring them wrong. What sort of termination did you use? What sort of scope?

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
  • Mike GreenMike Green Posts: 23,101
    edited 2008-09-25 20:54
    The Propeller datasheet has that information. See www.parallax.com/tabid/442/Default.aspx.

    Oops. The datasheet has lots of information, but not rise / fall time. Considering that the cog counters can produce
    a 160MHz square wave, the rise and fall time would need to be less than 1.5ns.

    Post Edited (Mike Green) : 9/25/2008 9:01:34 PM GMT
  • CRST1CRST1 Posts: 103
    edited 2008-09-25 20:59
    I could not find the time anywhere on th site.
  • CRST1CRST1 Posts: 103
    edited 2008-09-25 21:10
    I measured the output unloaded and tied to a 74LS245 driver chip with a 200mhz fluke scope. does there need to be more of a load to switch faster?· I also tried multiple prop chips. Does anyone know or have the specs on the outputs?
  • LeonLeon Posts: 7,620
    edited 2008-09-25 21:13
    You must have misread the scope.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-09-25 21:26
    You're doing something wrong with your measurement if you get 12µs. Are you measuring it right at the output pin? Do you have a good ground connection to the probe cable's shield? Some scopes have a bandwidth selection. Perhaps yours is not set to the full available bandwidth.

    Here's what I get (unloaded output):

    attachment.php?attachmentid=55999

    -Phil

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    'Still some PropSTICK Kit bare PCBs left!
    prop_rise_time.gif
    640 x 480 - 12K
  • mirrormirror Posts: 322
    edited 2008-09-25 21:41
    Phil,
    I want one of those oscilloscopes!!! 2.015ns - I'm truly impressed with the extra 15ps.
    When doing some high speed propeller 2 pin bit banging using my 80MHz USB scope, I connected the probes one way around, then the opposite way around, and then took the average position between those two very ugly looking traces. The traces looked like sinewaves that were approximated with 5 points, giving them a not-very-sober trapezoidal look!

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  • LeonLeon Posts: 7,620
    edited 2008-09-25 21:51
    I'd have expected 2-3 ns, as obtained by Phil. 12 us is just plain silly! Perhaps his scope really is 200 mhz, rather than 200 MHz.

    smile.gif

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
  • hippyhippy Posts: 1,981
    edited 2008-09-26 03:32
    I spent hours checking code for bugs when a customer complained to my employer that our software didn't work with their product and they had traced it down to a pulse which was meant to be 4uS as being 'well out of spec'. That eventually got costed as 'our mistake' for not asking if their scope was calibrated smile.gif
  • mirrormirror Posts: 322
    edited 2008-09-26 05:07
    I made a mistake that really was 'my mistake' as a very new engineer 15 years ago. I designed my first programmable logic device into a product which then had an occasional 'glitch' during development. Most of us were convinced that it was the firmware at fault. The firmware had 10's of thousands of lines of code and the 'glitch' was a random memory corruption. It sounded very much like a stray pointer!

    I checked and re-checked the signals to the memory chip, and everything looked fine within the measuring limits of or HP Logic analyser, which had a resolution of 5ns.

    Eventually I redid all the timing calculations by hand, taking into account the delay through every stage of the PLD. Under worst case conditions there was a 4ns hold-time violation to one of the ram chip signals, and fixing that removed the memory corruption! That was the only error found in the logic!

    We could have probably saved weeks of engineering time with a nano-second resolution oscilloscope, but we all thought it was a software bug.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sign In or Register to comment.