I was not quite happy with the fact that while the distribution in #39 fits the observed data quite well below 50 amps it doesn't fit that well above 50 and we are trying to use it to see what might happen at 200 amps. So I looked at the data set again up to about 3 this afternoon but this time fit the log of the probability. The figure below shows a much more convincing fit than in #39 and so I am much more confident in its projections.

While at it I also looked at a data set from up north going back to early October. The results of a log fit to that are in this picture

The most startling thing to me (the nice fits over the regions of measurement are of course pleasing too) is that while the electrical systems, installed appliances, HVAC gear etc. in the two houses are quite different the shapes of the distributions are very similar and the shape parameters (lambda) confirm this is being numerically quite close. Note that the vertical axis is now in seconds rather than minutes.

The McLean curves predict that 200 amperes would be exceeded for about 10^2.6 = 398.107 seconds per year. It is reasonable to assume that distributed among these 6 minutes there would be a period or periods of time long enough to cause the tripping of a 200 amp fuse or breaker. This reinforces the conclusion that is prudent to have the next size service i.e. 400 amperes. The projected time for demand greater than 400 amps is seen to be 10^-4 sec i.e 100 microseconds. It's clear that a service greater than 400 amps is not necessary and 600 would be absurd. But that is what the pros recommended based on the breakers in my panels.

In Ogden the house is smaller and hasn't nearly so many gadgets and doodads so the loads are generally smaller. The Ogden curve suggests that 200 amps will be exceeded 10^0.8 = 6.3 seconds a year. That is much closer to the edge and I would have been frustrated, had I this perspective, that I could almost get away with a 200 amp service there but I wan't given a choice. The electrician did what he thought he had to against my specific instructions (I was down in Virginia when he installed this stuff). 200 amps probably is the right size for Ogden even so it is so close to borderline. The curve suggests that 400 amp would be exceeded for about 4 nS. I don't think any breaker of fuse is going to pop in that time.

All this is fascinating, at least to me, but I don't see how it is going to be of help to OP as his is new construction and he has no history. So I still think the best he can do is count up the pole-amperes planned for his panels and divide by 4 or go through the details of the code load calculation.

I'd like to thank you guys for getting me thinking about this. This method of looking at the load data has got me confident that it really means something. Maybe I can use it as a basis for suing the electrician up north!

While at it I also looked at a data set from up north going back to early October. The results of a log fit to that are in this picture

The most startling thing to me (the nice fits over the regions of measurement are of course pleasing too) is that while the electrical systems, installed appliances, HVAC gear etc. in the two houses are quite different the shapes of the distributions are very similar and the shape parameters (lambda) confirm this is being numerically quite close. Note that the vertical axis is now in seconds rather than minutes.

The McLean curves predict that 200 amperes would be exceeded for about 10^2.6 = 398.107 seconds per year. It is reasonable to assume that distributed among these 6 minutes there would be a period or periods of time long enough to cause the tripping of a 200 amp fuse or breaker. This reinforces the conclusion that is prudent to have the next size service i.e. 400 amperes. The projected time for demand greater than 400 amps is seen to be 10^-4 sec i.e 100 microseconds. It's clear that a service greater than 400 amps is not necessary and 600 would be absurd. But that is what the pros recommended based on the breakers in my panels.

In Ogden the house is smaller and hasn't nearly so many gadgets and doodads so the loads are generally smaller. The Ogden curve suggests that 200 amps will be exceeded 10^0.8 = 6.3 seconds a year. That is much closer to the edge and I would have been frustrated, had I this perspective, that I could almost get away with a 200 amp service there but I wan't given a choice. The electrician did what he thought he had to against my specific instructions (I was down in Virginia when he installed this stuff). 200 amps probably is the right size for Ogden even so it is so close to borderline. The curve suggests that 400 amp would be exceeded for about 4 nS. I don't think any breaker of fuse is going to pop in that time.

All this is fascinating, at least to me, but I don't see how it is going to be of help to OP as his is new construction and he has no history. So I still think the best he can do is count up the pole-amperes planned for his panels and divide by 4 or go through the details of the code load calculation.

I'd like to thank you guys for getting me thinking about this. This method of looking at the load data has got me confident that it really means something. Maybe I can use it as a basis for suing the electrician up north!

Last edited: