What conditions cause the losses in the spark gap to be V*I rather thn I^2*R? This is new to me. Matt Behrend