I have a carbon bike frame with a PF30 bottom bracket. The bottom bracket suffers from creaking when applying high pedaling forces (high for me that is).

I removed the crank to check the state of the bearings, and to see if the crank axle has been under-rotating in the inner bearing races. As usual the crank needed to be tapped out of the bearings with a soft-faced hammer. Unfortunately, the axle stuck in the drive side bearing and the plastic PF30 bearing cup and bearing came out of the frame.

Obviously the force was applied to the inner race, but I was not tapping the axle very hard, so I don't think the bearing was damaged.

Can I press the cup back into the frame, perhaps using some retaining compound, or should I replace the cups (which will mean replacing the bearings I believe)? My concern is that the plastic cup has been deformed by being extracted and is now undersized.

  • 2
    The PF30 cup getting extracted with that little effort does not sound good. Do you have a pair of calipers accurate to 2 decimal places? I suspect your shell has gotten oversized. I can't find the exact specification for the PF30 shell, but I believe the actual spec for the internal diameter might be something like 45.96 - 45.98mm - that's under-sized relative to the nominal spec of 46.00mm, but that would deal with the variation in cup sizes. If you measured your shell at several places, you might be able to confirm.
    – Weiwen Ng
    May 26, 2020 at 19:15
  • @WeiwenNg Unfortunately I do not have a good set of calipers. May 26, 2020 at 19:22
  • Can’t you upgrade to a bottom bracket where both bearing halves screw into each other? Or T47. I’m not sure about the compatibility.
    – Michael
    May 27, 2020 at 5:21
  • @Michael you can re-thread a PF30 BB on a metal frame to T47. You absolutely can’t do that to carbon, as it won’t hold threads.
    – Weiwen Ng
    May 27, 2020 at 9:25
  • @WeiwenNg: I thought there are versions of T47 without threads which just screw into each other? Or maybe that was for PF30.
    – Michael
    May 27, 2020 at 9:30

2 Answers 2


Once you get the bearing and cup off the spindle, you can assess the state of the bearings a little. Turn them with your fingers while pushing hard in both the radial and lateral directions. This kind of test isn't perfect because the loads applied in use are so much higher, but sometimes it can reveal notchiness or roughness that lets you know you'd be better off replacing them.

Unfortunately, between the creak and the ejection of the cup, it's likely that the bore on the frame is either oversize, roundness-challenged, or both. You could check it with calipers if you want. Any such issues can have an adverse effect on bearing life, as the creak indicates the cup is moving a little which in turn means the bearing alignment won't be ideal. If there's no or minimal wear or roughness detectable, on a personal bike I think it's fine to just re-use it and see what life one can get out of it. While it's convenient I would lift the bearing seal and degrease and re-lubricate it, but that's optional. One nice thing about doing that is once you get it clean and re-check how it feels, you can get a better read on whether any roughness or notchiness was caused by contamination versus damage to the races.

On the other hand, it wouldn't take much perceptible wear for me to just replace it while it's out.

Retaining compound is a good idea and hopefully it will eliminate the creak.

  • Thanks. NDS bearing is smooth, DS has some slight roughness detectable. It's a second bike that I may sell so I'll re-use the cup with RC for now. If I sell it I'll divulge what I've done to the BB. At least I know now that the crank axle is not worn from under-rotation. May 26, 2020 at 19:26

Elaborating a bit on Nathan's answer, I believe that while the nominal specification for PF30 bottom bracket shells is a 46mm internal diameter, the actual engineering specification is likely to fall within a range that's slightly smaller than nominal. For example, the BB30 spec is a 42mm shell, but this blog post by the Accidental Randonneur says that the actual engineering spec is 41.96-41.98mm. This under-sizing means that the shell provides a firm hold on the cups (if they're properly sized themselves), but it's possible the shell was manufactured or became out of tolerance.

It might also be good to measure the cups, since it's possible that the cup was under-sized rather than the shell being over-sized. I do not know the cups engineering specifications, but I would assume that if the cup measures smaller than the frame's BB shell, the problem was with the cup. However, if the cups are plastic, they may have deformed slightly if they were in-spec and installed into an in-spec shell. I would assume they are not usable after being removed. NB: it's possible that nylon cups might also deform over time from pedaling forces, and these items are a possible argument for using PF BBs with alloy cups (e.g. White Industries, Wheels Mfg.) Naturally, if you reinstall the cup and disclose that, the buyer could price in the fact that they should install a new BB.

To confirm shell size, you could use an accurate pair of calipers to measure the diameter of the drive-side shell at several locations. If any one of the measures is clearly over 46.00mm, then your shell has clearly become out of specification, and the bottom bracket cup would inherently fit loosely in it. Creaks are caused by loose fits. If the cup slid easily out of the frame, then that's another indicator that the shell is out of specification. If a new cup slides easily into the hole, then again, the shell is likely out of spec (or perhaps the cup was). In any case, the frame could have been made out of spec, or the crankset could have been poorly installed and it could have worn the shell out of spec. If the frame's shell clearly measures between 45.96 and 45.98mm at several locations, then it seems likely a new cup could fix the creak (NB: I can't locate the proper engineering specs for the PF30 shell, so I'm assuming that the BB30 spec extrapolates to the PF30 system; this engineering diagram from Hambini looks consistent with my assumption, but I can't confirm.)

I lack personal experience in what to do if the BB shell is too big. I initially wrote that I didn't expect retaining compounds to work. However, note the comments by Nathan; Locktite should be able to fill a gap in size of up to 0.1mm. Loctite 660 appears rated to fill a 0.5mm gap (listed as diametrical clearance maximum under the spec sheet), and it would be a solution if the bore is not round as well as oversized. However, the fix is now reliant on the strength of the retaining compound, and the BB cup will need to be centered correctly in the shell lest it incur off-axis loads.

In theory, if you found a PF30 BB cup that was intentionally made slightly oversized, that could be a fix. For example, Trek makes custom bearings that are oversized by just 0.1mm. This covers the case of a BB90 frame that is just out of specification. However, I can't find slightly oversized PF30 cups. Anyway, it would be best practice to confirm that the BB shell was oversized uniformly. What if the BB shell is out of round but not uniformly out of specification? I'm not sure.

If you're the original owner and a warranty applies, I'd assume that the manufacturer would probably inspect the frame with calipers to confirm that the shell is out of spec.

For future readers, what if you have the opposite problem, i.e. the cups are too hard to press into the frame? If the cups are within the engineering specification, then the frame's shell is likely to be too small. Happily, this can be reversed, albeit only by a shop with the specialized and expensive equipment necessary to face carbon frames (credit to James Raison for the article and deep dive). Also, if the bearing bores were not aligned properly (e.g. the surfaces are not parallel to each other), it's possible facing could address that.

Unfortunately, as Raison discusses, facing equipment is a poor return on investment for shops. This is unfortunate, because many bikes would benefit from some facing work; in the experience of the shop he interviewed, it seemed that only high-end custom bikes were consistently faced when the shop received them. Moreover, it doesn't seem likely that this would help Argenti. His frame would have to be in spec but the cups too small.

  • 1
    Each retaining compound has a gap fill number provided by the manufacturer, which is given as the diameter or total difference between the two parts. For Loctite 609, probably the most widely used for bikes, it's 0.15mm, which is a large number considering it's typically trying to correct a gap where there was supposed to be a 0.1mm interference fit. So it takes a pretty messed up frame to exceed what regular 609 can do, and you could move to a different compound if that did happen - they go to to about 0.5mm. May 26, 2020 at 20:27
  • Also, retaining compounds and oversize bearings don't address exactly the same problems. RC can fix out-of-round bores. Putting an oversize bearing in in the same situation could cause other problems, like shortening bearing life or in an extreme case cracking the shell. I'm a little down on oversize bearings because I think they're really only good at correcting frame bores that are just plain badly made, and the problem is for all their cost, that kind of frame is also likely to have side-to-side bore alignment problems they don't address that will make the bearing life terrible regardless. May 26, 2020 at 20:33
  • Thanks for extra info. The cup do not slide out 'easily', it did require some stout-ish tapping with a nylon faced hammer, probably about as much as I'd expect if I were intentionally extracting the cup. It probably took more force than I should have expected it would take to slide the axle out of the bearing, which should have made me realize something was going wrong May 26, 2020 at 20:37
  • @NathanKnutson I did not know this, and I will amend the statements in the answer that deal with gaps. Can you confirm that the fillable gaps for compounds go up to 0.5mm, though? That does seem big relative to the tolerance in what I think is the BB spec.
    – Weiwen Ng
    May 26, 2020 at 20:50
  • Loctite 660 has a gap fill number of .02in/0.5mm. That's an extreme example of course. I should also note that retaining compounds often have some some degree of warnings in the fine print against using them on plastics, which technically includes carbon. 609 is pretty widely used without issue on carbon frames but I wouldn't necessarily promise any of the other ones can be used with impunity. Again, it takes a lot to go beyond what 609 can do. I've seen the more extreme ones used with success on antique bike type situations where the bore spec was a little pre-modern. May 26, 2020 at 21:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.