P!N20 wrote:I had the opposite experience when getting new ball bearings for my Campy hubs. Campy specifies ball bearings in metric to 1000th of a millimetre, but I could only find the corresponding sizes in imperial.
Ok. So you mean they specified the imperial sizes (like 1/4") but stated them in metric units like 6.350 mm. Well yeah, that's not really the "opposite" is it. It's actually exactly what I'm talking about.
I was just a bit miffed that when I enter "bicycle ball bearings" into the ebay search that I overwhelmingly got hits for sizes of 4mm, 5mm and 6mm in loose ball bearings (even when I specifically searched for BICYCLE ball bearings). Hence the question of "have they recently changed over" to these (whole number) metric sizes?
Anyway I've found a source of what appears to be good quality 1/4" and 3/16" from the US. So I will give those a try.
From memory two sizes were bang on but the third was 1/1000th of a millimetre out. Hub still spins smooth.
Which really leads to my other question about how much small size discrepancies actually matter. If you think about how a cup and cone works I think it should cope ok with small size discrepancies. I've never even considered replacing 1/4" (6.35mm) with 6.0mm for example, however I have seriously considered replacing 5/32" (3.969mm) with 4.0mm.
Remember that there are two types of discrepancies. One where all of the balls are the same size as each other, but all sightly different from the correct nominal size. And the other where the balls have poor size uniformity within the set. This latter type of variation is a bigger concern in my opinion.
Torana68 wrote:.... and cheap arsed bearings may have a large size variation, avoid cheap eBay bearings unless they have the quality advertised ie ANSI/ABMA Standard 10A
Yep thanks Torana68. That's the sort of size discrepancy that I want to avoid.