Before you can create an arc fault detector system, one must first be able to create reproducible and stable arcs. Here is
a photograph of a 400kW DC arc. The arc in this photograph is a small conductor racing around the gap between two parallel
copper doughnut shaped electrodes. In the first photograph you see what a normal camera or a human eye sees. The arc is moving
so fast that it looks like a light circle. The second photograph was taken with a high-speed shutter to help stop the action.
Here you can see the two circular electrodes.
Off to the right of the two electrodes is a photocell looking across the face of the electrodes. The photocell is looking
down a long small diameter tube so that it can only see the arc for the brief period of time that it crosses immediately in front
of the sensor. A high-speed data system measures the time between revolutions of the arc around the electrodes by looking at the
signal from the photocell eye sensor. Since we know the dimension of the circumference of the electrodes and have time we can
now compute the speed of the arc. We have measured arc speeds in excess of 700 MPH.
The second photograph allows you to see the structure of the electrodes. What we have done here is to magnetically
constrain the arc to run around the perimeter of the electrodes. This is very useful for several reasons. By allowing the arc
to move we reduce the localized concentration of heat and thus the need to repair or replace the electrodes after every test. By
constraining the arcs location we can now position sensors such that we can guarantee that the sensor will have the arc in its
field of view. Finally this geometry allows easy measurement of speed of arc motion. Now look more closely at the second
photograph and you can see that the arc did not extend completely around the electrodes. The arc is bright enough to expose the
film in a very brief time. The motion of the arc creates a smear of light in the gap between the electrodes. High-speed movies
of the arc in many different geometries have been used to create a knowledge base that allows us to be able to accurately predict
the moton of an arc inside of any switchboard.
Figure 3 shows the electrodes without the arc present. Once can see the photocell tube on the right hand side of the
electrodes. Heavy metal turnbuckles hold the electrodes in proper parallel alignment. After several test the electrodes become
covered with a heavy layer of copper oxide and the adjacent switchboard surfaces are covered with a layer condensed copper oxide
smoke.
Figure 4 shows what happens if the arc is not constrained. The arc struck at the bottom of the two parallel copper
electrodes. The electrodes were wrapped with a heavy fiberglass electrical insulating tape. While this tape can insulate red
hot metal objects, it can’t prevent the motion of the arc. The arc went up the electrodes while eating through the tape and
through the side of the switchboard behind the electrodes. Notice that the arc paused at the tip of the electrodes before it
went out. The tips of the electrodes used to have square corners, but now they are well rounded. The electrodes were
electrically insulated from the side of the switchboard. The long jagged hole in the side of the switchboard is collateral
damage from the arc.