For the starter buckets in a 480V MCC, should the arc flash calculations be done on the MCC bus and then these numbers would apply to all buckets? Or should each bucket be handled individually using the PD in each bucket. Essentially creating a new bus for each bucket on the load side of its' breaker?

In the studies that I have done, I used the one Cat for the whole MCC.

It is unlikely that a significant arc flash will be confined to a single bucket and not propagate to the vertical bussing in each compartment, so almost of my customers 'use' a single arc flash value for an entire MCC. Some customers consider an integral MCC main device to be sufficiently barriered from the rest of the MCC that it will have time to perform its function prior to becoming engulfed by an arc in a bucket. But no customer wants to calculate values for each individual bucket.

Thanks for the replies. What about for the main breaker? It seems that for the main breaker, you would have to go to the upstream device in order to calculate the arc flash for the space the main breaker is in. Since when you open the door for the main breaker compartment, the line side of the breaker could be exposed. Depending on what's upstream, the potential incident energy at this point could exceed 40 cal/cm2. Should arc flash calculations for the main breaker use the upstream PD?

For MCCs like you describe I use the following approach:

1 Label on the MCC main or incoming line cubicle with incident energy based on the upstream OCPD clearing time.

Generic labels on each bucket that refer to the main or incoming line cubicle arc flash data. That way if you have to relabel there is only one change per MCC.

MCC Labels

I have a situation where the customer has an MCC with no main and the bus is a Cat 1. There are a few 10 hp motors but most of the loads are circuit breakers to control panels that feed small motors (1/2 to 2 hp), small starter for 2hp loads, or CB's for 4kw unit heaters. I'm guessing it would be best to model each load individually? I know they can be modelled as a lump load but then I wouldn't have the protective device/bus hazard category for labels. Any thoughts or suggestions?

If you have cable and breaker data then it makes sense to model the loads.

Yes, the main bus numbers should be used for each bucket. This is because the line side of the bucket disconnect has almost the same energy level as the MCC Main bus.