The hole tolerance is probably going to be driven by manufacturing requirements. But if fatigue/vibration are critical, then you need to maintain joint clamp-up and not allow joint movement. Whatever hole size you select, you are going to need to test the assembly to verify performance. There is no practical method to predict friction and clamp-up loss.

Yes, this is more of a best practise design question.

Where it came from, is the company I'm working for has a historical convention of using the next number drill larger than the bolt, which results in a clearance hole that's actually got less clearance than the positional tolerance on the thread centre for the anchor nut - so in a worst case tolerance stack up you could have sheet metal biting into the threads, which means match drilling every hole (or just ramming the bolt through anyway).

I asked around, but no one was able to give me a definite reason why we didn't go a little larger - at least up to the counterbore size of the anchor nut. Of course, change without understanding is very risky! Closest explanation I got was that you ignore friction for ultimate shear loads, in which case its better to have a close tolerance fit, but that goes against all theory of bolted joints! May as well just have pins or rivets for everything if friction doesn't exist.

Yes, for ultimate loads, friction is ignored as it can't be counted on. But friction is critical for fatigue/vibration loads.

Having threads in bearing is not good for any of the loading conditions, and probably not for SCC (thought fortunately I don't have to worry about SCC in composite parts).

'Thin sheet metal' for removable panels is pretty sketchy as far as load transfer.

However, for GA/homebuilt Acft structural doors, infrequently removed [~1-time per year, for inspection], then layout the pattern in the door with pilot holes and drill-to-size thru the mating door-doubler flange... and very lightly deburr. Drill for a class 1 or step-ream for a close-reamed fit. See NAS618.

Mount the nut plates directly in-line with the bolt hole using a stub-screw to lock it down for drilling the [2] adjacent NP rivet holes... or using a specialized NP drill aligned to the hole to drill these holes. Then assign a door number and witness mark to ID the install/orientation and ensure re-alignment for swift install [no mix-up with 'identical' looking doors].

NOTE1. For thin sheet metal, subject to being 'double-dimpled'... there is potential for added strength. IF appropriate dimpling, drilling and fastening is used in conjunction with countersunk [holes] in the fay-side of the NPs. Double dimpled construction can be especially 'strong' due to the install of the flush-head screw/bolt in the matched/inter-locking/line-drilled holes/dimples.

On my dad's homebuilt T18, he installed 'flush doors' [double-dimples as described above] in each of the lower wing skins/bays 'opening reinforcement-doublers' for interior PS-890 sealing [fillet, packing, over-coating, etc] to ensure fuel tightness of the wings... then installed the doors 'wet with sealant' to attain a 'permanent install'. He performed a low fluid water-pressure check to validate perfect-seal of the integral tanks... and he NEVER disturbed these 'structural doors'. EVER. I presume that the retired aircraft still has these doors in-place... probably never opened... even by the current museum-owner.

PS: my sister was the only one in our family with fore-arms/hands small-enough to fit/thru/maneuver around the integral-tanks 'bays' for sealing corners, ribs/spars-edges... and over-coating every rivet bucked-tail. SOOOO she was voluntold to do this 'job'. To this day, she vividly remembers the unique stink of the polysulfide sealant PS-890... and how it took weeks to remove the stuck/cured sealant residues and the 'stink'.

Regards, Wil Taylor
o Trust - But Verify!
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation, Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", HBA forum]
o Only fools and charlatans know everything and understand everything." -Anton Chekhov