These 220 calcs are good for computing the average load on the service since we assume diversity but when these all end up on a single branch you are looking at the instant load at any particular time.
What you *can* do may not translate to what you *should* do.
