In general, the issue of port size goes like this:
On the one hand, the closer you are to the chamber, the higher your pressure.
On the other hand, the closer you are to the chamber, the longer your pressure duration (assuming a 16" minimum barrel length). Said inversely, the closer you are to the muzzle, the shorter your pressure duration. At some point too close to the muzzle, your gas impulse won't last long enough to cycle the bolt reliably, so you end up having to increase the port size to really goose that bolt to the rear and then let inertia do the rest.
On the third hand, there's a difference in pressures between .223 Remington and 5.56 NATO ammo.
And on the fourth hand, it is possible to over-gas an AR action and cycle the bolt too fast and too hard. One symptom of an over-gassed action is the buffer hitting the rear of the tube (ie, completely compressing the spring and making noise). Another symptom is that the bolt doesn't pick up the next round from the mag.
So what you have to do when balancing these two issues is:
a) start small. Like 0.049 is a lower bound I'd start with when I know nothing.
b) assemble the barrel, gas block, gas tube, etc.
c) load a magazine with one round. Chamber from the magazine. Test fire. Observe where the brass goes - if it ejects at 3 to 4 o'clock out the side of the ejection port, things are looking promising.
d) remove magazine. Load two rounds. Chamber first round. Fire. Stop. Check to see if the rifle loaded the second round from the magazine. If it didn't, you might have short-cycled. Check the brass of the round in the mag, and try to determine whether the bolt got over the rear of the case rim.
You might need to open up the gas port a little more. If your ejected brass is flinging out of the port at 4 to 5 o'clock and it's ejecting fast and hard, your action might be over-gassed, and you'll have to put in a heavier buffer to slow down the bolt to pick up the next round in the magazine.
e) If your bolt did pick up the next round, fire it and see if the brass ejects where the first round did. If this is all OK, things are looking more promising. Load a magazine with five rounds and cycle through them. If you can rip through all five rounds, you're probably done. If you really want to test, then run five rounds of NATO ammo and five rounds of .223 SAAMI ammo through it, then inter-mix them to insure that the action cycles with both types of ammo.
If the rifle ejected at 2 o'clock or further forward, you're probably under-gassed, and you need to increase the size of your port.
If your port is closer than about 3.5 to 4" from the muzzle, NB that you might end up with an issue about the duration of the gas impulse to the BCG. Don't try to locate your gas port too far forward.
When I'm increasing port size(s), I go up by about 0.005" at a time. That moves the ejection direction (for me, on rifle and mid-length gas systems) by about "half-o-clock" or 30 minutes. I've never done one of those carbine-length barrels with a custom barrel where I had to drill a port. Custom barrel guys seem to want longer barrels.
Maximum port sizes I've ever seen have been about 0.090".
One way to slam-dunk this once is to drill a max-sized (0.085" or so) port and use a gas block with an adjustable gas valve. Then you test fire the rifle and fool around with the valve to get it right.
