Someone is trying to convince me that EMT runs for parallel feeders have to be physically next to each other (they're on two different tiers of a conduit rack). The conductor length for each conduit is the same. Of course they're insistent without any code backup.
As long as each conduit contains all phase and neutral conductors, including the EGC, you can route them along any path you need. Provided that you follow all other parallel set rules.
You can route them 30 feet apart from one another, as long as they end up being the same length. Although that construction would stick out as an example of poor planning.
In the case of isolated phase installations, which are much more restricted, you need to be careful not to magnetize the vicinity of the conduits. And therefore route the isolated phase conduits as closely as possible to one another. In my opinion, I don't see the value in doing an isolated phase installation. To me it makes much more sense to balance the magnetic field in the conduit, and not even think about it outside the conduit.
Parallel feeders need to be constructed in such a way that no one set becomes a preferred path for current. All paths must be set up to share the current equally.
1. At least 1/0 or larger, for the current-carrying conductors.
2. Same length. In a practical sense, the obvious question is how rigid is this rule?
3. Terminated in the same manner.
4. Be of the same conductive material with the same type of insulation.
5. Routed in the same size and type of conduit, or within the same conduit.
6. Routed through the same ambient temperature and environment.
7. Routed with the same quantity of current-carrying conductors in each conduit.
#2 is an interesting rule. Consider a 500A feeder, built with two sets of 250 kcmil Cu in separate conduits. Suppose you constructed it such that one set ended up being 45 ft, and the other set ended up being 48 ft. If you calculate the current in each set in the case of a full 500A load, you get 242A in the 48 ft set, and 258A in the 45 ft set. Assuming wire resistance dominates over other ohms in the system. On paper, this is a problem, because the wire is rated for 255A. I see this as a very realistic example, and understand exactly how it could happen, even for an installer who knows and understands the intent of this rule. Unfortunately, the NEC doesn't quantify just how different the lengths can be in practice.
I'd also like to speak to #7. Suppose you have three sets, and you think you can group two of them in the same conduit. Think again. Because now the remaining third set isn't heated as much due to the number of active wires in the conduit. The remaining third set becomes a preferred path for more current, because it isn't operating as hot.