Time Domain Reflectometer

Discussion in 'The Completed Projects Collection' started by Sparky49, Mar 11, 2012.

  1. Sparky49

    Thread Starter Active Member

    Jul 16, 2011
    834
    417
    INTRODUCTION:
    A TDR (Time Domain Reflectometer) is an essential piece of equipment when working with cables and communications. However, commercial TDRs are very expensive, remaining in use with only large companies. This project aims to provide a cheap, functioning TDR for those interested in its basic operation and theory.

    THEORY & MECHANICS:
    This TDR works on a very simple principle, that waves are reflected from a fault along the cable. TDRs not only are able to discover the type of fault, but also the distance the fault is – useful in real life when digging up a mile of telephone cable would be impractical.
    Put simply, this TDR works by emitting a pulse which is split at a BNC junction. Part of the signal goes straight to the oscilloscope, whilst the other travels down the cable. If there is a fault, then the pulse will be reflected back to the oscilloscope where it can be analyzed.

    SCHEMATIC & COMPONENTS:
    Schematic1.jpg [1]
    2 10nF Capacitors
    1 22nF Capacitor
    1 4n7 Capacitor
    1 1nF Capacitor
    1 220pF Capacitor
    1 47pF Capacitor
    1 220μF Elec. Capacitor
    1 78L06 6V Regulator
    1 1N4148 Diode
    1 74AC14 Schmitt Inverter
    1 15K Resistor
    1 150R Resistor
    5 220R Resistor
    1 22R Resistor
    1 47R Resistor
    1 BNC Connector
    1 5 pole, 2 way switch
    3 Jumpers

    PCB:
    PCB2.jpg [2]

    CONSTRUCTION:
    Construction was very simple. A big thank you to Nerdegutta for making the PCBs.

    IMG_0596.jpg


    OPERATION:

    This is also simple. Connect the TDR to the BNC Junction. Then connect one end to the oscilloscope and one end to the cable under test. If there is no break, then you should have a clean, square pulse.

    If not, then a pulse will be reflected, and the shape will reveal the type of fault.

    Please note that the cable must be less than half of the wavelength that corresponds to the main frequency of the pulsetrain. Otherwise the reflected pulse will become mixed with the output pulse. This can be calculated with the simple formula:

    λ=c/v

    REFLECTIONS:

    Below are some reflections, which can be used to determine the type of fault.

    NO FAULT:

    [​IMG]

    OPEN:

    [​IMG]



    SHORTED:

    [​IMG]


    RESISTANCE:

    [​IMG]



    INDUCTANCE:

    [​IMG]


    As can be seen, each has its own very distinct reflection.

    I hope this has been useful.

    Sparky


    CREDITS:
    [1], [2] Nerdegutta
    EAGLE Files Nerdegutta

    FURTHER READING:
    http://www.epanorama.net/circuits/tdr.html
    http://www.elektor.com/magazines/2011/october/time-domain-reflectometry.1948603.lynkx
     
    Last edited: Mar 18, 2012
  2. Georacer

    Moderator

    Nov 25, 2009
    5,142
    1,266
    Stash point

    edit: After a LOT of frustration, I gave up to the authoritative forum client who wouldn't let me stuff more than 6 images in a post.
    Please settle for the thumbnails in the first 3 pictures.

    I 'd like you to check if the images are in the correct order, I 'm afraid I may have messed up.
     
    Last edited: Mar 12, 2012
    Sparky49 likes this.
  3. Sparky49

    Thread Starter Active Member

    Jul 16, 2011
    834
    417
    Thankyou very much Georacer, I will remember the 6 picture limit for future posts.:)

    Any questions or thoughts?
     
  4. Georacer

    Moderator

    Nov 25, 2009
    5,142
    1,266
    Actually, yes. You mention the wavelength. The wavelength of what? The square pulse contains all of the frequencies.
     
  5. Sparky49

    Thread Starter Active Member

    Jul 16, 2011
    834
    417
    Sorry, I meant the duration between each pulse.

    If you use λ=Fc, this tells you how far the pulse can travel before the next pulse.

    You want the cable length to be less than half of this distance, so the pulse can be emitted, reflected and arrive at the oscilloscope before the next pulse, otherwise the pulses and reflections get mixed up.

    It is for this reason that a selection of capacitors give a choice of time constant, allowing the duration to change depending on the length of cable.

    The longer the cable, the greater time you want between pulses.
     
  6. Georacer

    Moderator

    Nov 25, 2009
    5,142
    1,266
    Got it. Please correct it in the first post too.

    Can you also do something for the low-res schematic?
     
  7. RRITESH KAKKAR

    Senior Member

    Jun 29, 2010
    2,831
    89
    Hi,

    Your project is very interesting can you please explain more how we know that cable are short and open like this from oscilloscope.?!!
     
  8. Sparky49

    Thread Starter Active Member

    Jul 16, 2011
    834
    417
    Thanks for your interest.

    Take a look at the picture which is labelled 'OPEN'. The first pulse is from the TDR. The second pulse is reflected from the fault. If this pulse goes above the 0V line, then it is open. If however, this second pulse goes below the 0V line, then we know that it is shorted.

    The other pulses are other reflections, but we don't have to worry about this.

    I hope you understand. :)

    Sparky
     
  9. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    Good to see someone building that Elektor project or something similar.

    Well done.
     
    Sparky49 likes this.
  10. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    You will note when I write a long article the first thing I do is create a bunch of stash points. You can always get rid of them if needed, but you can not create them later. Nice thing about being a moderator is I can bring back deleted posts (usually).

    When people start posting to your article whatever space you have set aside is it. The exception is stash points for attachments, but 6 is the limit in either case.

    I have used TDR's as test equipment. I liked the ability to measure the impedance of the transmission line over distance (which is just an extension of measuring a break in the transmission line over distance, which is a function of time).

    I have no problem with this project being posted however. If someone ever comes up with something better it can be posted too (along side). Projects need not be practical, as they also teach. This was not meant to describe yours, just a general observation.

    TDRs were originally invented to find breaks in underwater cables, usually trans-oceanic. Given when these cables were laid you could not always get a diver to the site it made the proposition and repair much more practical. A good commercial TDR can locate a break within a couple of feet over a thousand miles of cable.

    I remember when a gaggle of engineers borrowed ours to find a break in the data cable that feed our production test consoles. It was an early version of ethernet, 10Base2. Turned out some construction types cut the cable when they were working on the elevator. That was one expensive cut, I always wondered if it was discussed with them.

    Thank you for hosting the information locally, what is so hard for other people to understand about the concept?

    Thanks for the project. It is interesting.
     
    Last edited: Mar 12, 2012
    Sparky49 likes this.
  11. KJ6EAD

    Senior Member

    Apr 30, 2011
    1,425
    363
    If I understand correctly, your cable length limitation is based on the fact that you're launching a pulse train into the cable under test instead of a single pulse. Why not eliminate this limitation?
     
    Sparky49 likes this.
  12. Sparky49

    Thread Starter Active Member

    Jul 16, 2011
    834
    417
    Studiot;

    Yes, it was the Elektor article which inspired me. I believe they had an article a month before the project which was also very interesting.

    Bill;

    Thanks for your kind words. I agree, projects can be great for teaching. I have learned quite a bit from this little project.

    KJ6EAD;

    Put simply, my scope doesn't have storage. If I didn't repeat the signal, then I'd have to have lightning fast eyesight for the 8μs it appeared on the screen. :)

    Thanks for looking.
     
  13. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    I did once experiment with TDR to look for breaks in post tensioning cables in bridges, but my experiments failed. That is they failed to locate known breaks but found spurious ones.
     
    Sparky49 likes this.
  14. Sparky49

    Thread Starter Active Member

    Jul 16, 2011
    834
    417
    Then it was a success no?
     
  15. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    It was successful in showing that TDR (at least at the frequencies I was using) is unreliable for testing post tensioning cables.
     
    Sparky49 likes this.
  16. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    You might think about using a computer sound card oscope, if it is fast enough. That is pretty equivalent to a storage scope.
     
    Sparky49 likes this.
  17. Sparky49

    Thread Starter Active Member

    Jul 16, 2011
    834
    417
    Well...

    I'm torn between making a solar flare detector, or... a tube based oscilloscope:D.

    I reckon it would be cool to have the tdr and its own oscilloscope in the same case.

    Adding a further element to the TDR.

    Either way, I'd like to do something with valves - their glow is mesmorising...:eek:
     
  18. JMW

    Member

    Nov 21, 2011
    88
    8
    The waveforms are sort of similar to the old HP TDR's I have used. Very nice post. I'll look into building it. Do you factor in the velocity factor the the cable under test?
     
    Sparky49 likes this.
  19. Sparky49

    Thread Starter Active Member

    Jul 16, 2011
    834
    417
    You can do, but I just used the speed of light, much quicker and I know it off the top of my head. :)
     
  20. RRITESH KAKKAR

    Senior Member

    Jun 29, 2010
    2,831
    89
    Can you explain bit more about the pulses you are using ...
    and in the open why there are two pulse at both right and left side of oscilloscope...??
     
Loading...