Well, yes and no. Here's how I see it:
Between the utility transformer and your panel bus, there is a resistance. Between the panel bus and the AC compressor there is a resistance, and between the panel bus and the lighting circuit there is a resistance.
When the AC starts up, there is a voltage drop ahead of the panel that the lighting circuit sees, and an additional voltage drop in its circuit after the panel that the lighting circuit does not see.
When the AC circuit has a relatively larger resistance, it has a larger voltage drop, which reduces the voltage, which raises the current (but not without limit) which increases its voltage drop even more.
There must be a point where the increase in current is overcome by the reduced voltage, and the current is unable to increase any more, and the voltage drop peaks at some point.
Are you suggesting that, despite the greater voltage drop, the smaller conductors allowed a greater current increase, enough to cause a greater voltage drop on the before-the-panel conductors?
And, that the larger AC conductors reduced the overall voltage drop so much that the current on the entire service was reduced, and not increased, by the lower circuit resistance, because it's a motor load?
If so, okay, I'll buy that.
However, I would have loved to take bith voltage and current readings on both the main and the compressor terminals during start-up.
Added: I understand that upsizing helps with start-up voltage sag. What I question is whether that results in a gretaer or lesser current. Does the current increase faster than its own current-caused voltage drop drops?
