jonwvara 

I just serviced the bottom bracket on my Raleigh Sprite, and reassembled it with new cotters--1970s vintage NOS Sturmey-Archer cotters; the old ones had been hammered by a previous owner and were toast.
Pressed 'em in with the Bikesmith Designs cotter press, using the correct amount of pressure. (I have some previous experience with cottered cranks, though not a huge amount.) The crankarms ended up at about 180 degrees to each other. I think so, anyway: depending on how I look at them, I can convince myself that they're a tiny bit off one way or the other, which is probably a pretty good indication that they're more or less centered.
But what DID happen is that the inner chainring is now about 1 to 1 1/2 mm further inboard than it was at first. I can tell because the heads of the chainring bolts were just barely clearing the rear derailleur cable (which is routed over the bottom bracket) before I started, and now they catch on the cable when I turn the crank.
It's not a big problem, since the chainline is still about dead-on--maybe slightly better than it was before, in fact--and I can always reroute the cable in a housing under the bottom bracket. And there's still plenty of clearance for the chainrings. But I'm not sure what changed. Has anyone else experienced this?
Before anyone suggests it, I didn't unintentionally reverse the spindle. The side-to-side difference is so great that would have been hard to do without noticing.

sailorbenjamin

The notches that the cotters sit in on the spindle are a little wider than the cotter allowing a little side to side play.

randyjawa 

Yup, you have nothing to worry about, in my opinion. If it is working, let it be. You might want to check the other side to see if it is set at the same position.

Also, I have found that it is often necessary to file cotter flats to get 180 intervals on the cranks. This is a pesky task but worth the effort, in my opinion.

jonwvara 

That makes sense. So if I understand you correctly, I could return the crank to its original position relative to the centerline of the bike by pressing the cotter drive-side back out, sliding the crankarm a little further outboard on the spindle (which could be gauged by the position of the chainring bolt heads relative to the cable) and pressing the cotter back in place?
Excellent! There are so many ways to screw up that I despair of finding them all--but I do find a lot of them.

cudak888 

Correct.

-Kurt

ftwelder

Rather than positioning it where it was before, consider where it should be. Check chain line and shoot for the middle of the stack, match the end of the cylindrical part of the spindle to the face of the arm for aesthetic points or whatever floats yer boat!

jonwvara 

Great, thanks everyone, I fixed it. But to hijack my own thread--which i guess is okay, since the subject line is pretty elastic--what's a good way to tell if the cranks are in in fact opposed by exactly 180 degrees? The crank spider doesn't line up with either crank, which introduces a source of visual confusion when you just try to eyeball it.
Furthermore, how close is close enough? If I were to determine that my crankarms were at, say, 178 degrees to one another, I doubt that I could improve on that by filing the cotters. I'd be at least as likely to make it worse, I think.
There's a Yiddish term that a friend of mine used to use--which I would spell phonetically as "farpotchket"--which according to her means, approximately, "something broken as a result of an attempt to fix it."
My instinct is to say that if it eyeballing it doesn't indicate pretty clearly that it's off one way or the other, it's probably close enough.

sailorbenjamin

Yeah, I think the old globo del ojo should be good enough here

rhm

One way to tell if the arms are properly aligned would be to check how well it's balanced. Take the chain off and give it a spin! See where it stops. If it always stops in the same place, that might indicate the arms are both on that side of the center of gravity.

John E

1) Hang a plumb bob from one pedal spindle.
2) Set that crank pointing straight upward, so that the string of the plumb bob runs through the center of the crank spindle.
3) The plumb line should now also be even with the center of the opposite side pedal spindle.

jonwvara 

Have you done this? Not sure I can picture it.
But I am convinced that my cranks are a little bit off--just a few degrees. What I have never understood is how one files the cotters to correct for that small amount of error. Presumably in this case you just file one of them. Which one? How much do you file it? The goal would be to slightly change the angle of the flat, I'm guessing, rather than keeping the angle constant and just filing the flat deeper, which would let you press it in further but would not change the relative positions of the arms. Or would it? This makes my head hurt.

Grand Bois

The flats on the spindle are at 180 degrees from each other. If the angles on the cotters are identical, the cranks can't be anything but 180 degrees from each other. You don't file cotters to change the angle. You file them so they seat at the correct depth.

jonwvara 

Right, that's what I had always thought. So what does it mean when the crankset is assembled with two apparently identical cotters (good-quality 70s-era NOS Sturmey-Archer cotters than came in a still-sealed blister pack) which seem to press in correctly, but your alignment is still off? What's the recourse then? Keep trying different pairs of cotters until you get it right?
It is possible, I suppose, that one of the crankarms (the non-drive side) was slightly buggered by the previous owner. That original cotter showed signs of bing hammered, and the replacement on that side had to be pressed in significantly further than the drive-side cotter to get the same amount of torque (not that I used a torque wrench, but by feel). The hole where the fat end of the cotter protrudes is also a little bit ovalized. Maybe I need a new crankarm?

Buikema

This whole filing business has also always confused me. I have installed cotters and ended up w/ not 180 degree opposite cranks, which must mean that they (w/ pounding and/or cotter press) were set w/ different amounts of pressure, and there fore different depths? This also makes my mind bend.

Also, whenever I pull cotter they seem to have a serious notch worn in them, and I envision that correctly seated cotters would sit with the flat of the spindle to the flat of the cotter, no notching. Help me on this one.

Grand Bois

Different amounts of pressure won't change the angle. Whatever the angle is, the crank will turn on the spindle to match it. If the angle is the same on both cotters, the arms will be at 180 degrees. I suppose that an ovalized hole in an arm could cause the arm to sit crooked, but I've never seen that. They're normally made of hard stuff.

The notch is from pedaling.

dbakl

Assuming the cotters are oriented correctly: I believe they need to be inserted opposite each other, not both from the same direction.

Right, the filing is also to get them to insert the correct amount. Almost all are oversize to be a more universal uset and always require filing to fit correctly. Be glad they're the soft ones of today and not the hard steel ones of the 50s!

Yes, that's what I'd guess, pressure from the edge of the spindle cutout.

Grand Bois

Obviously.

I'd like to know if there is a convention for installing them. Nut up or down on the drive side with the arm forward? It wouldn't make any functional difference, but there is a convention for everything else on bikes.

rhm

It is also possible for cotters to be correctly oriented, but not correctly seated. I don't know how this happens, but it happened to me only a few weeks ago; and as a result the crank arms didn't align quite right. I pressed the cotters out, filed the flats a little, greased them a little, and pressed them back in place. This solved the problem. Needless to say I couldn't see exactly what was wrong in there, but I assume one of the cotters was a little twisted or something like that.

clubman 

I've heard nut down so it won't catch the trailing cuff of your favourite cycling "pantalons"

Road Fan

I have vivid memories of my grandmother telling me "quit potchkying with my ______!" when I was just looking at it to figure out what it did. So my interpretation of "farpotchket" would be "it was messed around with" or "some hack farted around with it." Not as polite as your interpretation ...

But, I think your eye will be the best gauge of straightness. I don't know of a tool that would help, though I could imagine designing one. I would say it needs to be aligned well enough that you aren't irritated with how it looks.

Oldpeddaller

JohnE's plumb line idea sounds like the perfect test. I just check visually using a metal ruler as a straight edge to ensure that when viewed from the side, both cranks look like a single steel bar. If one cotter seats deeper into the crank than the other (ovalised hole, worn spindle flat, etc.) then I file some material off the other so it seats to the same depth to bring the crank into line with the other side.

Back in the 1970's I used to do this several times a week but since then only a handful of times. Now for the first time ever, I've come across a pair of cotters I can't shift and will have to drill out (Gazelle ladies single speed). Just shows it doesn't matter how good you think you used to be!

Convention is with the pedal at the top of the stroke the nut should face forwards - no idea why, it makes no difference at all.

jonwvara 

Oldpedaller, I recognize the voice of experience and believe you, but I can't get my head around how filing some material off on flat--while presumably maintaining the original angle--can possibly change the alignment of the crank arms. Do you have to believe that it will work in order for it to work? "Inscrutable" is the word, I guess.

jonwvara 

Update: Crankarms were a few degrees out of alignment. I pressed the pins out, shifted the drive-side pin to the non-drive side and vise-versa. Also reinstalled them as Oldpeddaller described, so the nut faces forward when the pedal is at the top of it's stroke--the opposite of how I'd installed them previously. (That method is consistent with the prevent-the-nut-from-snagging-your-trousers-cuff concept that clubman mentioned.)
The result is that the cranks are now opposed by what appears to be 180 degrees. I'm not sure what conclusions can be drawn from this other than that if you fiddle with things enough they sometimes work out.