Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Sending signals through the computer?

  1. Dec 24, 2008 #1
    Hello there,

    My project is a Smith Chart calculator/simulator that takes the Zo and ZL of a transmission line and calculates various data like SWR, SWR(dB), return loss, coeff I, E, P, etc. It also constructs and plots on a graphical Smith Chart.

    I want to take this one step further and retrieve my Zo and ZL from a REAL transmission line.

    I want to connect a coax cable via an usb to coax adapter and send a signal through my computer to determine these properties.

    I need some ideas/guidance on this. Should I do this via the USB port? I think essentially I need to make a TDR to retrieve data such as E incident and E reflected.

    It would be ok if I am not able to determine Zo for now and just use a data sheet. But I would like to retrieve data to calculate ZL and then send that data to my program.

    Thanks so much in advance for your help!
  2. jcsd
  3. Dec 30, 2008 #2
    What is your operating system? What kind of adapter is it? I have never used USB for such purpose and with other projects almost always used an USB / RS - 232 adapter. Then using typically a microcontroller for sending and receiving pulses. Immediate thought that comes to mind (I could be wrong) is, that you need some kind of specific software to show the received pulse unless you know something really handy in the OS API. All that I've seen just show data as bytes (8 bits = 8 pulses) and single pulses go to the abyss. This is an interesting project so please feel free to post on your progress.

    Additionally computer interprets pulses as logical 1 or 0 so tdr pulse has to exceed the threshold to be interpreted as 1. Thinking more on this the resolution is quite poor.
    Last edited by a moderator: Dec 30, 2008
  4. Dec 30, 2008 #3
    Hi misgfool,

    Thanks for your input.
    OS: Win XP

    I havent purchased these, just kept them in my bookmarks to investigate what is a good way to go about this project.

    I'm trying to collect information on how exactly I would analyze the pulses. As you mentioned, some piece of software that does that would be handy. I'll update this page as I make progress.

    Thanks again.
  5. Dec 30, 2008 #4


    User Avatar
    Science Advisor
    Homework Helper

    There is a little more to it than this.
    You need some electronics to detect the arrival time of a pulse.
    What voltage? What time resolution?

    You could start with something like this http://www.picotech.com/ reading the values from the device with the supplied library is fairly easy.

    The devices you linked to are for sending digital comunications streams over coax.
  6. Dec 30, 2008 #5


    User Avatar

    Staff: Mentor

    I'm not understanding how those adapters are going to help you, but I could just be missing what you intend.

    That's an ambitious extension to your project, but a very real-world extension, so that's a good thing. Yes, you would basically be making the TDR function of a cable tester, where you drive a fast-rise-time pulse from a known Zo driver down the coax, and monitor for the initial amplitude (for the cable Zo value), and for any reflections that come back (from cable mismatches, or from the wavefront hitting the termination if it is not matched).

    To do this, you would need a driver with a calibrated Zo, and a fairly fast oscilloscope-type digitizer. You might be able to implement these with USB digital-output and USB analog-input modules, if you find ones with speed specs that look good. I'd look at National Instruments website to see generally what is available:


    They are not the cheapest source around, but their quality is generally good, and we buy a lot of stuff from them for our own test equipment needs.

    Now, a less expensive alternative would be to generate some data to simulate what you would get from your DAQ card, if you really had built the whole thing up. Generate the data based on the equations for TLs, and then feed that data back into your Smith chart mapping software. That would be a good intermediate step.

    Alternately, you can make your TDR up out of a signal generator and digital oscilloscope, capture the 'scope data, and feed that into your program. That would be cool, and doesn't cost you the $ of the USB data acq hardware.
  7. Dec 30, 2008 #6
    Many thanks Mgb_phys and berkeman for your responses!

    I was thinking with those adapters I linked to, that I would send a pulse and retrieve reflected data and then analyze that data via some software but that doesnt seem to be a good approach.

    I looked at the lowcost DAQ products at NI you linked me to and yes 220 dollars CDN is a lot of money for a school project.

    I'm not clear on the second suggestion you gave. Would I have a database of TLs lines(for Zo) then use the length and equations to figure out ZL?

    The third suggestion is a route I'm most likely going to take. In my college we have labs with Labjacks(DAQ units but I think they are serial) and Agilent scopes and generators with the Agilent VEE software on the computers at the stations. We also have HCS12s, I will ask my prof what else he has.

    Many thanks for direction on my project!
  8. Dec 31, 2008 #7
    I might be cheaper, but the cool-o-meter would definitely be plummeting.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook