How to True a Bicycle Wheel: Lateral, Radial, Dish, and the Nipple-Turn Rule

Why one nipple turn moves the rim two ways at once, why the two sides don't move equally, and why a good wheel barely needs truing — with links to the original posts.

Truing looks simple — turn the nipple toward the side that's rubbing — and that's exactly why so many wheels come back from a quick truing slightly worse than they went in. The nipple does more than one thing at a time, and on a rear wheel the two sides don't respond equally. This guide lays out the mechanics one workshop has written about, with links to the source. "Right" means the drive (freehub) side.

1. The three things you're correcting

A wheel can be out in three independent ways, and you correct them in this order on a truing stand:

They interact, so you work in small increments and recheck — but the reason they interact is worth understanding, because it's where truing goes wrong.

2. One nipple turn moves the rim two ways at once

For lateral wobble, tightening and loosening look the same — either pulls the rim toward the spoke you adjusted. For radial, they are not the same: tightening pulls the rim inward (toward the hub), loosening lets it move outward. So a nipple turn moves the rim along the combination of both — sideways and radially at once.

This is the trap. If you fix a lateral wobble by cranking the single nipple at the worst point, you'll pull a radial flat spot in at the same time. The fix is to spread the correction: adjust the nipples on either side of the worst point as well, so their radial pull cancels out while the lateral correction adds up. Truing a wheel and bringing a half-laced wheel to "done" are, in the end, the same skill.

Source: Nomu Lab #939

3. The nipple-turn rule: the two sides don't move equally

On a dished rear wheel, the same nipple turn produces a different amount of lateral rim movement on the left than on the right — because the two sides' spokes meet the rim at different angles. From the workshop's own measurement (loosening every nipple exactly one turn and watching the rim shift):

Non-drive side lacingRoughly equal lateral movement
Tangential (crossed)drive 1 turn ≈ non-drive 3/4 turn
Radialdrive 1 turn ≈ non-drive 2/3 turn (or just over 1/2)

Exact figures depend on hub dimensions, but the direction is fixed: the non-drive side moves the rim more per turn, and a non-drive radial side moves it more still (a radial spoke has no angle loss). The practical upshot when correcting dish or a one-sided wobble: don't turn both sides the same amount — the non-drive side needs less.

Source: Nomu Lab #3926 — Lateral rim movement per nipple turn

4. Bladed spokes: don't let them twist

Aero/bladed spokes are where careless truing does visible damage. As you turn the nipple, the spoke wants to wind up — twist along its length — and a spoke left twisted, or with its blade turned sideways instead of edge-on, is a botched job. Hold the blade with a bladed-spoke holder while you turn the nipple. A spoke that has wound up but not permanently deformed can be gripped by the blade and untwisted back; one whose blade is set sideways needs its tension released, the blade re-aligned, and the nipple re-tightened.

Some bladed spokes have anti-rotation ribs under the head (four ribs at 90°). At low tension you can feel them click into place every quarter-turn; once tensioned they won't rotate, so set the orientation before you bring the tension up.

Source: Nomu Lab #6714

5. A well-built wheel barely needs truing

When a couple of spokes are replaced on a wheel that was built well, you can usually true it back to round, straight and centred by adjusting only the nipples you disturbed — the rest of the wheel hardly needs touching. If a wheel fights you everywhere during a simple repair, the build underneath was the problem, not the repair. Good truing starts with a good build, and the inverse — a wheel that comes back from a shop with bladed spokes twisted and turned sideways — is a sign the work was done without the right care, not a sign the wheel was hard.

Source: Nomu Lab #6714

6. Centre drift over time, and what it tells you

A dished rear wheel that was built dead-centre will, after a year or a few thousand kilometres, sit very slightly toward the drive side. This isn't nipples unwinding — it's the spokes elongating under tension, and because of the angle difference the same tiny elongation shifts the rim toward the drive side. A track hub with no dish doesn't drift; a disc front (little dish) barely does.

This is a diagnostic. A small drive-side drift on a used wheel is normal and forgivable. A drift toward the non-drive side is not something use produces — it means the wheel was built off-centre to begin with. That's why an experienced builder tolerates a hair of drive-side drift but calls non-drive drift "built wrong." (Non-drive radial wheels drift fastest, because angle — not spoke length or tension — is the dominant factor.)

Source: Nomu Lab #3926

Truing is the last 10% of building

The same understanding that builds a wheel keeps it true. Start from the Wheel Building Guide, see why the two sides differ in the spoke tension guide, and get the dish right with the spoke length deep dive. Every claim here links into the Nomu Lab Diary archive.