Pushing ultra-high-speed photography to the limit.

Capturing Visible Electrical Discharge

The spark discharge across a gap with a high-voltage potential is only visible for between 10 to 20 nano seconds*. To capture a meaningful sequence of the spark behaviour not only requires a frame rate of 1 Billion frames/second, which itself is pushing the boundaries of ultra-high-speed photography, it also requires pinpointing the critical moment when the developing spark is visible.


High voltage spark discharge is, with the possible exception of lasers, one of the fastest events to capture using a camera, lasting only a few nanoseconds.

Specialised Imaging worked with a client to demonstrate the Specialised Imaging SIMX16 camera’s ability to capture these events with a view to using the camera to characterise the behaviour of electrical sparking across different dielectrics for high voltage transformers used in power stations.

The laboratory experiment was set up with the objective of capturing a high voltage spark jumping from a brass disc anode to a metal plate cathode. The set up presented two key challenges beyond the 1Bfps camera frame rate and light sensitivity for a very short duration self-illuminating event. Firstly, triggering the camera correctly relative to the event and secondly, protecting the equipment from the inherent high voltage electromagnetic pulse.


Pinpointing the event

The nanosecond scale event duration requires the camera to be triggered without delay and from the event itself. To do this excludes the use of signal conditioning equipment, but this then heightens the risk of damaging the camera by using kilovolt event voltages. To achieve a safe and instant trigger an induction loop on the event high voltage trigger line was employed. This is simply a wire coiled around the trigger wire to provide a consistent instantaneous (insignificant nanosecond delay) trigger directly from the experiment and ensured an acceptable trigger voltage (up to 50V) entered the camera.

Although the camera could be triggered in this way, the correct camera image capture timing took a combination of experience and refinement. Based on their experience, the client’s team were able to estimate when the spark would occur. The SIM design is 16 independently programmable cameras (or Channels) so initially Channel 1 had a relatively long exposure time (10μS), to give it the maximum possible chance of seeing the event as an over bright flash. The remaining 15 channels were configured with exposure times of 1/15th of CH1 exposure time and triggered consecutively – CH2 starting with CH1 and CH16 ending with CH1. This pseudo division of CH1 allowed quicker refinement of the correct triggering for the spark. As a rough analogy, using only one long exposure time of 10μS for a 10nS event would be like trying to initially find a 1 second event within 16 minutes of footage. The trigger delay from the event was adjusted until the spark was seen in Channel 1 and another channel. The process was repeated but reducing all channel exposures to 1/15th of the previous and using trigger time of the Channel where the spark was seen each time. When using this procedure, the gain of each channel was carefully adjusted to ensure the spark was seen without damaging the intensifiers.



Protecting the camera

High voltages can damage any complex scientific instrument such as a SIM camera, by either the spark itself or the electromagnetic (EM) pulse generated by it. To protect the camera a Faraday cage around it was used for this experiment, where the lens view was unaffected by the mesh when placed close enough to it. The use of copper power and ethernet cables are two ways to compromise Faraday cage protection. To overcome these, the SIM fibre optic communication option was used to remove any metal connections, and the power cable was within copper mesh trunking from the camera Faraday cage to the mains socket located a significant distance away.


10nS = 0.000,000,010 seconds


Read the Application Note. Click here.