Not related to bearings, but frame flex, one of the biggest issues with Thompsons Pacifics was frame flex around the cylinder area due to traction forces, the result being broken bolts on the smokebox saddle, looking through photos there's plenty of evidence of these engines with one, two or even three bolts missing in service. The flexing of frames also had an effect on valve timing and uneven wear in bearings and to some extent even Gresley engines suffered the same.
I have seen some references and I think drawings that allow for sideplay in engine frames, mostly eight and ten coupled, I'm sure the 9F has side play designed into the middle axle and I'm also sure the 8F has side play designed into the rear axle and some NER eight coupled freight engines as well, possibly also the WD's, being at work I cannot confirm any or either.
Where side play is designed in there also has to be corresponding alterations to the knuckle joint in the coupling rods, some having semi spherical bushes / bearings as opposed to straight or flush.
I accept there is going to be some working clearance but the gap on the Britannia cannon boxes visually appear much larger than just running clearances, especially when compared to the bogie which show virtually no gap between the cannon box and horn guide face. I didn't have a chance last night to dig out all of my Britannia notes and references but it'd be nice to know if it is tabulated somewhere and what the tolerance actually is. There may also be more information in Wild Swan on the Princess Coronation engines 6256 and 6257 as well as the 8F and I think the roller bearing Black 5's are covered in one of the sections as well in the Wild Swan Black 5 books.
I have never seen the cannon boxes modeled on any Britannia or Standard five model, real or simple psuedo covers. They are certainly visible on the leading bogie axle, but accept that on the drivers they are less visual and of course you cannot add one to the models driven axle.
Apologies to Ken for the distraction and diversion.
MD