Which is better: AC or DC Testing? Will these tests hurt my generator? Here are the facts:
- DC high potential test equipment is relatively small and easy to transport. AC equipment is comparatively difficult to transport (large and heavy).
- DC testing stresses the dielectric in a different manner than AC. DC tends to stress the end windings more. DC tends to be more sensitive to low resistance contaminates on the surface of the windings.
- AC high potential testing stresses the stator insulation system much as it would during normal operation. AC tends to stress the straight section of the stator windings more. AC is more apt to flush out high resistance insulation weaknesses that DC might not.
- DC can be used in a controlled over-voltage testing methodology where the test can be terminated prior to winding failure. AC is basically a go/no-go proof test.
- AC and/or DC high potential tests will not fail an otherwise sound insulation system. Sufficient research has been performed within the industry to conclude that windings which fail hipot already had either systemic and/or specific issues.
- AC and/or DC high potential testing does age the stator winding insulation. However, a one minute AC or DC high potential test equates to only a 0.0004% reduction in service life per test.*
In conclusion, AC high potential testing is better suited for acceptance proof testing of new equipment (i.e. coils, bars, and stator windings). DC testing is better suited for in-process proof testing, maintenance proof testing, and controlled over-voltage testing of new and/or used equipment. AC or DC high potential testing should not be used as the sole source of diagnostic or acceptance data. An experienced and qualified generator testing specialist will recommend a testing protocol specifically suited to your machine’s life cycle and operating history.
* A one minute AC or DC high potential test equates to approximately 11 hours of useful life expended. Assuming that a stator winding has a nominal 30-year life expectancy (263K hours), this would equate to only a 0.0004% reduction in service life per test.