Diagnosing Creaks: A Systematic Elimination Method
A creak shows up on a ride and by the time you're home, checking the bike on a stand, it's gone. Or it's there on the stand too, but only when you push down on one pedal, and you regrease the seatpost anyway because that's the easy thing to check, and the noise comes back on the next ride. Creaks are frustrating for one specific reason: hearing a noise tells you almost nothing about where it's coming from, because sound travels through a bike frame far better than most people expect.
A creak felt at the saddle can originate at the bottom bracket. A click that seems to come from the pedals can be a loose bottle cage bolt two hundred millimetres away. Chasing the sound by ear, checking whatever seems closest to where you hear it, is how people end up stripping and regreasing three different parts of the bike before finding the actual source, or worse, never finding it.
The fix isn't a better ear. It's a fixed test order that isolates one variable at a time, starting with the fastest checks and working toward the ones that need tools and time.
How do you find the source of a creaking bike?
Work through the likely sources in a fixed order, testing one at a time and confirming the creak is gone before moving on, rather than fixing several things at once and hoping. Start with the saddle and pedal interfaces, since they're both quick to check and responsible for most creaks, then move toward the bottom bracket, headset and cockpit only if the fast checks come up clean.
The reason order matters is that fixing two things in the same pass tells you nothing. If you regrease the seatpost and retorque the stem in the same session and the creak stops, you don't actually know which one solved it, which means the next time it happens you're back to guessing.
Why creaks are hard to diagnose by ear alone
Carbon and aluminium both carry vibration efficiently along the tube walls, so a noise generated at one bolted joint can resonate and seem to be coming from somewhere else on the frame entirely. Riders standing on the pedals often hear a creak "in the bottom bracket area" that's actually the saddle rails flexing under the extra body-weight load of an out-of-saddle effort.
Load direction changes what creaks too. Seated pedalling loads the bottom bracket and cranks in one pattern; standing loads the saddle, seatpost and headset differently as the whole bike rocks side to side under you. A creak that only appears out of the saddle is a genuine diagnostic clue, not just an annoyance, because it narrows the list of likely sources before you've touched a tool.
The 8 most likely sources, ranked and in test order
This order runs fastest and most common first. Work down the list and stop as soon as the creak is confirmed gone over the same test stretch you used to reproduce it.
|
Test order
|
Creak source
|
Likelihood
|
30-second test
|
|
1
|
Saddle rails & clamp
|
Very common
|
Rock the saddle side to side while seated static; remove and regrease or repack if suspect
|
|
2
|
Pedal / cleat / crank interface
|
Very common
|
Coast with pedals unweighted, then stand and load one side hard; swap which foot you push
|
|
3
|
Seatpost-to-frame interface
|
Common
|
Remove the post, clean the interface, apply the correct paste, reinsert to the marked height
|
|
4
|
Bottom bracket
|
Common
|
Remove the chain, rotate the cranks by hand feeling for grit; load-test on a trainer
|
|
5
|
Headset
|
Occasional
|
Front brake held on, rock the bike fore-aft, feel and listen at the headset
|
|
6
|
Disc rotor bolts / thru-axle
|
Occasional
|
Check bolt torque, spin the wheel freely and listen for a rhythmic tick
|
|
7
|
Cockpit bolts (stem / bar clamp)
|
Occasional
|
Torque-check faceplate bolts to spec, then sprint gently out of the saddle
|
|
8
|
Bottle cage / accessory bolts
|
Rare, easy fix
|
Remove bottles and cages, ride the same test stretch to confirm or rule out
|
Diagram of a bike showing 8 common creak sources ranked by likelihood and test order
The elimination method, step by step
-
Find a short stretch of road, a speed bump, a rough patch, a short out-of-saddle sprint, that reliably reproduces the creak. You need this as your control test.
-
Work down the table above in order, testing and fixing one item at a time.
-
After each fix, ride the exact same test stretch before touching anything else.
-
If the creak is gone, stop. If it's still there, undo nothing, move to the next item on the list, and test again.
-
Log what you changed and in what order. If the creak returns weeks later, you'll already know where to start.
Saddle, seatpost and pedal interfaces first
These three account for most creaks because they carry direct body weight through a small number of contact points, and they're the parts most likely to have been touched last, a saddle height adjustment, a pedal swap, a new set of cleats.
Metal seatposts want grease at the frame interface. Carbon posts, and carbon frames with metal posts, need a carbon-specific assembly paste instead, because grease actively reduces the friction the clamp relies on to hold the post in place without over-tightening. Using grease on a carbon interface is a common cause of both creaking and slipping saddle height.
Pedal spindles typically torque to around 35 Nm into the crank arm, and cleat bolts to roughly 5-6 Nm, though the numbers stamped on your specific pedals and cleats always take priority over a general figure. A loose cleat bolt is one of the most common creaks in cycling and one of the easiest to miss, because the play is small enough that the cleat still clips in and out normally.
Bottom bracket, headset and cockpit
If the fast checks come up clean, the bottom bracket is next. Water and grit work into the bearing interfaces over time regardless of bottom bracket style, and a creak here often shows up as a rhythmic tick matching cadence, once per pedal revolution, that gets louder under load.
Headset creaks tend to show up as a knock felt through the bars under braking rather than a tick tied to cadence. Cockpit bolts are usually the last stop, since an integrated cockpit like an H-series two-bolt faceplate setup rarely creaks if it was torqued correctly during assembly, but a bar or stem that's been removed and reinstalled without a torque wrench is worth checking before you assume the frame itself is the problem.
Worked example: a creak that only shows up out of the saddle
A rider notices a creak at 3,200 km that only appears standing on climbs. Saddle rock test: silent, ruled out. Pedal test standing and loading one side: creak persists. Pedal spindle torque checked against the 35 Nm spec: correct, not the cause. Cleat bolts checked next: one bolt reads roughly 2 Nm against a 5-6 Nm spec, clearly under-torqued. Retorqued to spec, same climb ridden again: silent.
The whole process took under fifteen minutes because it followed the list in order instead of starting with the bottom bracket, which would have meant pulling the cranks for a fault that was actually two bolt-turns away.
Frequently Asked Questions
Why does my bike only creak when I'm out of the saddle?
Standing changes both the load path and the load direction through the bike. It puts extra weight through the pedal-crank interface and rocks the frame side to side, which loads the saddle rails, seatpost and headset in ways seated pedalling doesn't. A creak tied specifically to standing efforts usually points to one of those three areas first.
Can a creak mean something is about to break?
Usually not. Most creaks come from dry or slightly dirty contact between two bolted surfaces, not structural failure, and resolve with cleaning, the correct paste, and correct torque. A new creak that appears immediately after an impact or crash is different and should be treated as a structural check, not a maintenance one.
Should I use grease or carbon assembly paste on a seatpost?
Use grease on a metal post in a metal frame. Use a carbon-specific assembly paste wherever carbon is one of the two mating surfaces, since grease lowers the friction the clamp needs to hold position, which can cause both creaking and the saddle slowly dropping under load.
Why did the creak come back after I regreased everything?
Regreasing several parts in one session without testing between each one means you likely masked the real source temporarily rather than fixing it, or contamination worked back into the joint faster than expected. Go back through the elimination list one item at a time and test after each single change.
Share:
Wrapping Bar Tape on an Integrated Cockpit
Chain Wear: How to Measure It and Why It Saves Your Cassette