The argument about these horrible staking procedures as a means to prevent the outer race from falling out when installing timing spindles would be quite silly for a company in production business. Good logic would first shrink in components with highest temperature requirements like the timing side spindles at min. 180 degrees and only then shrink in mains outer races at around 100-120 degrees. So no, staking was a last step for somewhat securing the races from walking into the crank pin nut. Noway can you rely on shrink fits in serious engineering for components to stay put. You´d have to think about other ways to have them solidly positioned.
My guess is they had the "brilliant" idea of slimming down the outer races for no chamfer to move them as close as possible to the very wide twin crankshaft. So only choice was the roller bearing type with plain outer race as only with these you can put the races on a table grinder for the job. But then they were out of ideas how to stop them from walking, no lips on these. That is when the staking was entered - with troubles coming up onto later generations.
So I think you guys have to come over the religion of absolutely wanting plain outer races and rollers kept on the inner race. At the same time get over the imperial sizes of mains when repairing clapped out bearing seats, look at metric types with all sorts of lipped versions as you want them. You´ll be safe with them ever after and they are a LOT stronger than the originals. I think the Horner bros got 30 mm mainshafts today, was my choice certainly 30 years ago, by the way.
Sure, you cannot feed the oil pump scroll through the metric lipped roller bearing NJ 206 when it is dropped into the case already. So you either thread the mainshaft carefully through the already dropped in scroll plus key when lowering the crank into the case - or putting the case onto the crank below it. This is no miracle to do, easy compared to the string trick with rollers unwilling to enter the plain outer race without chamfer.
Alternatively, think about any other crank with mains ballbearings on, you just put on all components on mainshafts that have to be on place at that stage. Then get your electric heat gun and have the case up to min. 100 degrees and sink the crank down to stops - simple and very common elsewhere.
This works on the timing side and drive side as well. I got no ball bearings on the drive side, instead lipped rollers with oil seal between bearings. The outer roller bearing got a plain inner race so it slips through both roller bearings on assembly . So then the positioning of crank is done at two places:
Both inner roller races of d.s and t.s. are lipped certainly and set to near zero side play in cold case. That will expand a lot in operating condition.
So second place for sideplay setting is between lips at drive side, no ball bearing there, one direction already set by inner d.s. roller bearing type NJ 2206, with lips, plus the other direction is set by a special steel bush that carries the triplex sprocket. I did a fat hardened lip/top hat bush for providing the outer roller bearing lip and this sets side play better at hot engine. The NUP 2206 got a separate "lip" to go with its inner race shortened by the thickness of that extra lip. So then you can do the hardened top hat bush for the sprocket and ample faces to bear the ESA side loads. It seems I turned that single lip on the inner race of the NUP down for slipping the lot through all bearings of d.s. .
Next use of that top hat bush is a solid face taking the substantial side loads from the ESA springs: Standard was the sprocket loads the inner race side face of the ball bearing. But sadly there are two radii on that inner race and the minimal remaining face wears into the sprocket plus the inner race, throwing chain alignment out in time.
Did I say the 30-62-20/-16 mm NU or NJ 206 roller bearings are a LOT stronger than the old originals ? But after many decades with these old engines another factor is more important: Aluminium got a "memory" about earlier high load operatioins, meaning with a lot of hours of high stress it goes weak and soft and in places may do cracks and not reversible any more. Ask any mechanic about common aero engines, turbo types mainly, they hardly go more than 800-1000 hours before cracking cylinder heads in that dreaded large thread that holds the head on the steel cylinder, no studs on these. Or talk to dragster guys, some run alu conrods - but change them each year or even earlier, else they destroy the engine.
So that is why I will NEVER remove material around the mains bearing seats in old engines , places of high loads I´d think. And no screws in any adapters certainly, Greg will know very well. Does not matter when you have steel adapters for resizing correct fits in worn cases, this will not make the case stronger at all as you take out a few millimeters again - a no, no.
So another bonus with the smaller but even stronger metric roller bearings, you ADD alu adapters for them, from 63.5mm down to 62mm. When the old bearing seat is not too mangled you just turn an alu adapter 2 thou oversize, loctite all faces and press that in quickly. Line boring for correct shrink fit is exactly same job as with restoring the imperial sizes, same amount of efforts in either option.
There is the only extra part you got to do, the sleeve on the mainshafts from 1" up to 30 mm - no big deal I say and you can even care for some loose bearing fit on the old mainshaft, just get some Loctite 638 for the sleeve with push fit.
Between both roller bearings on d.s. I put a hard spacer for the shaft seal to act on for separating oils of engine and primary and have reliable venting of crankcase.
So what´s not to like about that metric mod, no weakening of bearing seats as no oversize boring for a sleeve is to be had. All sorts of lipped roller bearings avaliable from SKF or FAG top quality. I see no extra high demands on the skills of the mechanic with the metric mod but lots of snags with the standard stake-prone plain outer races. Same goes for Prof Neville´s steel threaded fiddly adapter, requires oversize boring the case - not my idea in old engines.
Vic
top hat bush on d.s. mainshaft against the outer bearing forming the outer "lip" and sprocket bush :