How to Tell a Bottom Bracket Creak From a Pedal Creak or a Derailleur Cage Rub

Three of the most common drivetrain noises sound almost identical from the saddle. Here's a simple isolation test that tells them apart in under five minutes.

You've already lubed the chain and checked for loose bolts, and the creak is still there. That's the frustrating middle ground: the noise clearly has a drivetrain source, but you can't pin it down by ear alone from the saddle.

Bottom bracket creaks, pedal creaks, and a derailleur cage rubbing all live in roughly the same part of the bike and can sound almost identical while you're riding, which is exactly why so many riders end up on forums asking how to diagnose a bike creak with no clear answer. The fix isn't a better ear. It's a structured isolation test that removes variables one at a time, and it takes less time than reading through a dozen forum threads guessing at the same problem.

Why this is genuinely hard to diagnose by ear alone

All three noises happen once per pedal stroke or continuously while pedaling, all three are subtle enough to blend with road noise and tire hum, and all three can be made worse or better by standing versus sitting. Sound also travels strangely through a bike frame: a creak that's actually coming from the seatpost can feel like it's radiating from somewhere near the bottom bracket by the time it reaches your ears.

The isolation test: one leg at a time, standing vs seated

Do this in a quiet space, ideally on a bike stand or turned upside down so you can pedal by hand, though it also works ridden slowly and carefully in an empty parking lot. Pedal seated, then pedal standing. Coast without pedaling, then pedal without coasting, to rule out wheel or hub noise entirely. Pedal with one leg only for ten or fifteen revolutions, then switch legs and repeat. Shift through several gears while pedaling at a steady, light effort, noting whether the noise changes.

A creaking noise only when standing is one of the most useful single data points in this whole process. Write down exactly when the noise appears and disappears through each change, because that pattern is the actual diagnosis, not a guess based on how the sound feels.

What a bottom bracket creak feels and sounds like

A true bottom bracket creak is rhythmic and tied directly to crank rotation: one creak per full pedal revolution, present from both legs equally, and almost always louder standing than seated because standing puts more lateral load through the bottom bracket shell and bearings. It typically won't change when you shift gears, since gear position doesn't affect the bottom bracket bearings at all. A creak that changes character when you shift from the small to the big chainring is a strong sign it's not actually the bottom bracket, whatever it sounds like.

What a pedal creak feels and sounds like

A pedal or crank-arm creak is also rhythmic, but it's usually isolated to one leg, meaning it disappears or gets noticeably quieter when you pedal with only the other leg during the one-leg isolation test. It comes from dry pedal threads, a loose crank bolt, or worn pedal bearings, and the single most useful elimination step is pulling that pedal, cleaning and greasing the threads, and reinstalling it. If the creak follows the pedal when you swap it to the other crank arm, you've conclusively found it: that swap test alone rules out the crank arm and bottom bracket completely.

What a derailleur cage rub sounds like

A rubbing derailleur cage doesn't care about crank position at all. It's a constant sound while pedaling in a specific gear or two, not a once-per-revolution click, and it usually changes or disappears entirely when you shift to a different gear. This derailleur cage rubbing noise is typically the chain rubbing against the front derailleur cage in cross-chained gear combinations (big chainring paired with the largest rear cogs, or small chainring with the smallest rear cogs), or a rear derailleur that needs a minor indexing adjustment so the cage sits centered over each cog rather than slightly offset.

That covers the three main suspects on their own. Put together, they line up into a quick decision table you can run through in order.

A simple decision table to work through in order

If the noise happens once per revolution and is louder standing than seated, suspect the bottom bracket first, then check crank arm bolt torque as a cheap second step. If it happens once per revolution but only on one leg during the isolation test, suspect that side's pedal threads or pedal bearings, and pull that pedal to grease and re-test. If it's continuous, gear-dependent, and disappears when you shift, suspect the derailleur cage or a need for a minor indexing tune, which usually takes under five minutes with a barrel adjuster.

If none of the above isolates it cleanly, the bearings inside the derailleur's jockey wheels are the next place to check. A worn jockey wheel bearing can produce a noise that mimics all three of the above depending on chain tension in a given gear.

Why terrain and riding position change what you hear

A creak that's obvious on a smooth trainer or turbo session can disappear entirely on rough outdoor pavement, simply because road buzz and tire noise mask a subtle creak that stood out in a quiet indoor setting. The reverse also happens: a creak tied to standing climbing effort might never show up on a flat, seated commute at all, since it depends on load patterns that don't occur during easy, seated spinning. If you're struggling to reproduce a noise a friend or mechanic described hearing, try to match the exact riding position and effort level where it was first noticed, rather than testing under easier conditions and concluding it's gone.

Other suspects worth ruling out before you conclude it's the bottom bracket

Two components outside this core trio commonly get blamed for the same symptoms and deserve a quick check first. A dry or slightly loose seatpost can produce a creak that syncs with pedaling because your weight shifts subtly in the saddle with each stroke, and it's often mistaken for a bottom bracket creak specifically because both are louder standing.

Cleats on your shoes, if you ride clipless, can also creak against the pedal body itself: a noise that's easy to blame on the bottom bracket since it happens once per revolution just like a genuine BB creak. Removing and re-greasing the cleat bolts, or simply trying a different pair of shoes for one ride, isolates this quickly. Once you've narrowed things down using the tests above, the remaining questions riders usually have are about the process itself rather than the mechanics.

How long should this isolation process actually take?

Budget ten to fifteen minutes for a first pass through the full isolation test: seated versus standing, one leg at a time, coasting versus pedaling, cycling through a few gears. That's genuinely enough time to narrow the noise down to one of the three main categories in this article for the vast majority of cases.

What if the isolation test rules out all three?

If pedaling standing, seated, one-legged, and through several gears all produce the same unchanged noise, look further afield than the drivetrain. A dry rear suspension pivot on a full-suspension mountain bike, a loose rack or fender mount, or even a water bottle cage bolt vibrating against the frame can all produce a noise that seems to track with pedaling simply because the whole bike is moving and flexing together as you ride.

Does the type of pedals I use (clipless vs. flat) change this diagnosis?

It adds one more variable rather than changing the core method. Clipless pedals introduce the cleat-creak possibility mentioned above, plus a cleat with worn or loose retention tension can itself produce a faint click as your foot rocks slightly with each stroke, easily confused with a genuine bottom bracket or pedal-body creak. Flat pedals remove that specific variable but don't eliminate the core three-way isolation problem this article addresses, since the bottom bracket, crank bolts, and derailleur cage all creak the same way regardless of pedal type.

Should I record the exact test results somewhere, or just remember them?

Writing it down, even briefly, is worth the thirty seconds it takes. "Standing only, both legs, gone when shifted to small chainring" is a specific, useful note that a mechanic, or your future self, can act on immediately. Compare that to trying to recall the exact pattern from memory a week later, once the noise has possibly changed character or a second noise has joined it.

This is a small habit that pays off disproportionately given how much time a poorly described noise can waste in diagnosis, and it's exactly the kind of thing worth carrying over once you've actually fixed the problem, not just while you're chasing it.

Once you've isolated and fixed it, that specific noise-and-fix pairing is worth keeping somewhere you'll actually find it again. Log it as a service event in Tiiks and the next time something similar shows up, you're not starting the isolation test from zero: you already know what it turned out to be last time, on this exact bike.