Yeah, it probably depends on how one defines the diffusion time, it's not such a precise concept. I actually think rtomes approach is a good one, that the average diffusion time may be usefully defined as the total radiant energy divided by the luminosity. With your numbers, that comes out about 7 x 10¹³ seconds, or about 2 million years. The Kelvin-Helmholtz time would then be about 60 million years.