What went wrong with my bottom bracket spindle length calculation?

Question

Here's the facts:

• My bike frame requires a 70 mm Italian threaded bottom bracket
• I'm going to use a Nuvinci N360 IGH which requires a chainline of 49 (± 0.5) mm.
• I have Andel fluted cranks, which according to at least couple of sellers should require a 110 mm bottom bracket with a JIS taper. And at least this seller says that 110 mm will produce a chainline of approximately 42 mm.
• I have a 3/32" chainring

I decided that I would buy a Shimano UN55 bottom bracket, as it was quite widely praised as a good basic bottom bracket.

Here's the assumptions I made:

• Shimano bottom brackets are symmetrical
• I need +7 mm on the drive-side, so I need 14 mm longer bottom bracket spindle

I ended up buying a UN55 70x122 bottom bracket. However after crank installation I noticed that the combination gave me a chainline on 53.5 mm (chainring normally installed) or 46.5 mm (chainring installed on the "inside" of the crankarm).

So where did my calculations go wrong?

So far I've considered the following things:

• What if the expected chainline was 43.5 mm instead of 42 mm? But that would have still caused a smaller error than what I am seeing now.
• What if the 122 mm spindle length is something special reserved only for triple cranks? However, I've found no info that it should affect the situation this way.
• What if Italian threaded Shimano bottom brackets use ISO taper instead of JIS taper? I've found no info to back that one up, and in that case, according to Sheldon Brown, the effect should be the opposite than what I am seeing now.
• What if the cranks are really using ISO taper instead of JIS? That would explain the difference I'm seeing now. But on the other hand, quite many sites confirm that the cranks should be used with a JIS tapered bottom bracket.
• What if on non-triple cranks the chainline is always measured "from the middle", like on double-cranks. This one I'm actually not so sure about; but I though that on this kind of single-speed cranks it would be measured from the chainring instead.

Of course I could (and actually may) go with the easy route and just order a new BB with a shorter spindle and hope it fits. However, as I try to learn bicycle mechanics, I'd really like to understand the math around this issue. As an added bonus I'd be very glad if someone could also explain the optimal UN55 spindle length that I should try to find next.

Answer 1

I went to take pictures of the BB installation as @Batman suggested. After removing the crank arm, I actually noticed an offset of 1-2 mm on the drive side, like @mattnz anticipated. However, I also noticed that there was a gap of about 1 mm between the frame and the BB lid:

After this observation I googled a little bit, and learned from this forum thread that the drive side main body of the BB should be installed before the non-drive side cup. I had installed the parts in the wrong order. I loosened the non-drive side cup a little bit, I was able to screw the main body in a couple of turns more. This eventually resulted as a chainline of 50 mm, which I find totally acceptable, as my target was 49 mm. The non-drive side cup sticks now out about a 1 mm or so, but the above mentioned forum thread suggested that this is OK as well.

After the corrected installation, I can assure (for future reference), that at least this my UN55 70x122 bottom bracket does not have any offset at all; length from end of the spindle to the center of the frame is equal on both sides.

I think that I've found answer to the math side of my question as well:

By combining info from pages 16 and 34 of this PDF I was able to find out that 122 mm Shimano BBs should give a relative chainline of +8 mm compared to the 110 mm ones. So if 110 mm BB should give a chainline of 42 mm with my crankset (as was the assumption), 122 mm should give 42 mm + 8 mm = 50 mm which is exactly what I'm seeing now.