Composite video over twisted pair (Cat5) with variable gain

Discussion in 'General Electronics Chat' started by mamcinty, Jan 27, 2015.

  1. mamcinty

    Thread Starter New Member

    Sep 11, 2013
    5
    0
    First I'd like to introduce myself and thank you for any help that you can offer. All that I know about electronics design is self taught so I very likely have made some stupid mistakes. This is not my first project but it's particularly difficult because I don't know much about analog design. This is not a project for a client.

    I need help modifying an in progress design. This device is a receiver in a composite video differential transmission system. The receiver (this design) is implemented with the AD8130 while the transmitter is implemented with the AD8131.

    It works beautifully but only with 1000ft of cable. This is because I implemented the reference design without thinking or properly reading...it was a very busy time and I felt like an idiot when I realized what happened.

    I need to redesign for variable cable length (gain). While the original circuit was designed for 1 to 100Mhz, I am only really concerned with 1 to 10Mhz being properly compensated.

    Here is the existing schematic
    View attachment Video-Receiver-v10.pdf

    and PCB
    [​IMG]

    This is the reference design
    View attachment AD8130_Differential_Video_Receiver.pdf

    and the entire datasheet for the AD8130
    View attachment AD8129_8130.pdf

    I'm not stuck with the AD8130 so feel free to suggest alternate options.

    Matt
     
  2. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    You mean exactly 1000ft of cable, or anything less than that? How long cable do you actually need?
     
  3. mamcinty

    Thread Starter New Member

    Sep 11, 2013
    5
    0
    I need to be able to provide the user of the device an adjustment (a pot, dip switch bank, etc) that will allow adjustment between 0 and 1000ft.
     
  4. crutschow

    Expert

    Mar 14, 2008
    13,016
    3,235
    You could use an adaptive cable equalizer such as this.
     
  5. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,540
    1,251
    If you are happy with the 1000 ft performance, then it probably is over-peaked at shorter distances. Simplest would be a dipswitch to select different values of C1/R1/both on the ADI app note figure 134. At 10 MHz the board layout shouldn't be a problem.

    ak
     
  6. RichardO

    Well-Known Member

    May 4, 2013
    1,235
    384
    That is a great chip that crutschow referenced but I don't think it will work for your analog signal.

    Did you remember to drive the cable from a video source with 120 ohms output impedance? If your video is from a standard 75 ohm source then you have to add the extra 45 ohms to get proper impedance matching to the cable. Similarly, The output impedance of your receiver must match the video device's input which is probably 75 ohms. You need to add a 75 ohm resistor from your AD8130 output to the connector.

    Even with proper termination, the cable losses may still be a problem. Losses in the cable have both an ac and a DC component. The R/C network in your schematic can compensate for some of the ac losses but they do not compensate for DC losses caused by the resistance of the cable wires. I don't have a good feel for the magnitude of the losses...

    There may be chips available to compensate for losses in your cable but if not you can roll your own version. The simple-minded approach is to do an AGC loop that servos the sync tips to the proper amplitude. Since the sync and video have a fixed relationship, the video gets adjusted to the right amplitude as well as the sync. (This only works if the video chain does not clip the video or sync).
     
  7. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    I am quite sure that CAT5 cable is 100 ohms not 120, but nevertheless your point is still valid.
     
  8. RichardO

    Well-Known Member

    May 4, 2013
    1,235
    384
    If so, then his input termination (120 ohms) is incorrect.
     
  9. crutschow

    Expert

    Mar 14, 2008
    13,016
    3,235
    That's correct. I mixed my digital with my analog and they don't mix. :oops:
     
Loading...