The NEC doesnt REQUIRE voltage drop to be at or below a certain percentage.
The NEC (as amended by the applicable AHJ) is, by definition, the legally enforceable bare minimum which MUST be met, regardless of any other considerations. It does not even pretend to represent anything approaching "Best Practice". If all you are concerned about is squeaking by the electrical inspection, then you would find AWG 1 AL perfectly acceptable for this application. But merely "meeting code" is virtually never the same thing as "doing the right thing" -- and this case is no exception to that rule.
Rather, it only recommends 3% for feeders and 5% for branch circuits!
Which is far too loose for my taste, especially in an application where you can expect motors to comprise a large/important part of the load. Probably
THE worst thing you can do to a motor, both in terms of its day-to-day performance and its long-term life expectancy, is to feed it with marginal/inadequate voltage. At a 3% V.D., we're already down 7.2V (relative to 240V); add that "NEC recommended" 5% for the branch circuit, and we're now down more than 18.8V. That's just plain too much, no matter what the NEC might "allow". And note, we have not yet even touched on the fact that, on any given day, that "240V" putatively supplied by the PoCo
might actually be as little as 216V! How happy is your "240V" air compressor going to be, trying to start up (probably against a partially pressurized tank) on only 197 volts?
Since #1 would be at 3% VD, I fail to see how thats barely adequate since no more than 100a can be ran through the wire without the breaker tripping and the only way for voltage drop to increase would be to pull more than 100a through the wire!
It's not that simple -- and I'm pretty sure you know that. For one thing, just how precise (and over what sort of time constant) do you think the trip-point tolerance on that main breaker is, anyway?
Now if the wire size was at or smaller than what was needed for 3%, then I would say barely adequate!
Which, as shown in my previous post, is precisely the case here. The calculated minimum required CM area is within ~1.5% of that nominally provided by AWG 1. If you don't consider that to be right there "at what's needed for 3%", then it's clear that you're just splitting hairs for the sake of supporting your argument. Consider: How precise was the OP's estimate of a 140-foot run length? And how much additional "effective run length" are you factoring in to account for the additional resistance imposed by each mechanical connection (of which there are at least four in the complete round-trip circuit loop, even presuming an unbroken wire with no intermediate junctions)? Care to re-run the calculations for an effective wire length of, say, 143 feet? So yes, as I've said multiple times now: THIS IS AN EDGE CASE. And like ALL edge cases, the prudent thing to do is go to the next-larger size wire.
But even 4% VD would be acceptable to most machines!
Maybe. Maybe not. There are too many variables involved to make any sweeping pronouncements either way. But even presuming that is correct, what size wire would we then need to use for the branch circuit(s), to hold THEIR additional V.D. to no more than 1.0%? More to the point, why put yourself in a position to need to worry about that?
Wylie I'm with you on this. 2ManyProjects, your reasoning makes no sense to me. If the OP's loads are going to be approaching a 100A then he needs a bigger feed because of loads, not because of VD.
On the contrary... When designing the system, if the panel CAN support a 100A draw, we have to assume that it WILL impose (at least) that much draw on the feeder, at least sometimes. Now obviously, in most "real life" cases, people don't routinely max out their incoming service. But that is irrelevant, really. Since it CAN happen, the only prudent assumption is that it WILL happen; and then design for the worst-case scenario.
Look, guys... As I've said from the very beginning, the OP very probably CAN "get away with" AWG 1 AL. That has been the context of all my subsequent comments, and is not under debate. So citing code requirements for the "bare legal minimum" as some sort of counter-argument completely misses the point I've been trying to make. The real question is, is it wise to settle for that "bare legal minimum" -- especially when it is relatively easy (at this stage in the game) to render the issue moot, and pretty much ensure the problem will never crop up regardless of how the 1,001 other variables which have yet to be nailed down happen to shake out?
