In the first part of this investigation, published elsewhere, a convenient method of measurement of the ionization threshold (IT) and losses in layer insulation by electrical means was devised. A supporting set of methods for the validation of the electrical measurements was developed in this work. Samples having two or three layers of a Litz wire insulated by either Ultem or a composite Nomex-Ultem system were used. Adiabatic heating produced by dielectric/ionization mechanism was compared with simulations and heating produced by a known power source. The results were found to be in good agreement. With increased confidence in the electric method, layer insulation life above IT at several tens of kilohertz and pulsed stresses was investigated. Multiple samples, both unpotted and potted, were stressed by pulsed HV above IT and driven to failure. The test protocols ensured insulation overheat below 30 degrees C, and thus, only electrical aging mechanism was assumed. Although the tests were conducted in a wide range of frequencies, voltages, and losses, insulation life appears to be determined by the energy deposited in the insulation, of which the lion's share comes from the ionization losses, rather than by other parameters. For two layers of 5-mil Ultem, the energy density to destruction, on average, was 1.7 and 9.5 kJ/cm(2) for unpotted and potted samples, respectively. Failed samples were examined thoroughly. Wide-area damage was typical for all, which served as an indication of uniform ionization. Wire damage by plasma was also noted as larger wire rigidity and multiple black spots on the enamel surface. Study of the temporal IT behavior spanning three years is briefly reported.
