Electric fence - High voltage arcing problem...

Discussion in 'General Electronics Chat' started by byrona, Sep 16, 2007.

  1. byrona

    Thread Starter New Member

    Sep 16, 2007
    2
    0
    Hey, Problem is...

    An arc on the fence can somehow induce a higher voltage that results in a flash over on the circuit board.

    Sitting in the lab with an open circuit output the voltage sits at around 7kV. Simply due to the capacitive effects of the fence a channel running on the fence will normally have a slightly higher output. The flash over occurring on the board must be in the region of 14 kV.

    Energizer explained..

    We have an energizer with 3 channels. All channels fire simultaneously, two of which output a negative pulse and the third outputs a positive pulse. This energizer is used on security fences so the fence is usually clean but there is a high level of capacitance due to the close proximity of one wire to another.

    On the fence Channel 1 and 3 will be interleaved for the upper portion and channel 2 and 3 will be interleaved for the bottom part of the fence - this meaning that the potential between any two adjacent wires will be the highest obtainable.

    Each channel then has a return so that the fence integrity can always be monitored. This return goes through an opto-coupled network to read the voltage on the fence.

    SITUATION explained..

    The fault has occurred on a real site, I am using a prototype test area that I have managed to simulate the occurrence in. The outputs from the energizer go through a long lead out before reaching the fence. This isn't necessary as the energizer is physically close to the fence it is just done on our test site to simulate a worst case fence example. This long lead out is something I have been thinking of removing, although I am not sure yet as to why it would contribute to this phenomenon.

    To demonstrate the flash over I use lightning diverters to create a controlled arc on the fence. These lightning diverters simply have two metal plates that connect via wires to the fence and ground or from one fence to another. The gap between the metal plates can be controlled so the voltage at which it arcs can be specified. I have 4 lightning diverters. Three are placed between each fence and earth, and the last one is placed between fence 2 and fence 3. Obviously because there is a higher potential between these two fences the arc gap can be set almost twice the size as the other 3 while still obtaining an arc.

    I setup an arc between Fence 2 (negative) and Fence 3 (positive) so that the gap in the lightning diverter is as large as possible while still arcing. Then I adjust the arc gap between all three fences and ground randomly, the flashover will eventually happen. I can’t get it to regularly occur.

    In one fixed setting of arc gaps on the 4 lightning diverters I counted 9 flashovers on the PCB in 5 minutes, during this time I didn’t change any part of the setup. Sometimes no flash over will happen for over 20 minutes, other times flash overs occur anytime a setting is changed on one of the lightning diverter.

    The gap must be as large as possible on all lightning diverters. With a small gap it appears too much energy is transferred and the flash doesn’t happen.

    My initial thoughts are that the fence is acting as an inductor and the sudden decrease in induced current at a point in the inductor results in the voltage rising beyond it to try and maintain the total field effect.

    I have tried to make this description as complete as possible, while at the same time not wanting it to look like a book so no one reads it… If you have more questions about the setup please ask!

    I am looking for suggestions as to how this is happening and how I could regularly get it to happen and how it can be prevented. Any thoughts that seem reasonable I will be happy to test out.
    thanks, Byron.
     
  2. byrona

    Thread Starter New Member

    Sep 16, 2007
    2
    0
  3. JoeJester

    AAC Fanatic!

    Apr 26, 2005
    3,373
    1,157
    Glad to see you solved your problem.
     
  4. cumesoftware

    Senior Member

    Apr 27, 2007
    1,330
    10
    Just quoting...
    So don't be mad. It is not our obligation to help. In most cases it takes one day or more for someone to answer
     
  5. ixisuprflyixi

    Active Member

    Sep 16, 2007
    52
    1
    oh, arcing was the problem. lol
     
Loading...