How to use this calculator
- Indicate the edges. TIR at the flutes, real holder, working stickout — rotate slowly and read the swing.
- Compare to the programmed feed. The share f_z ± TIR tells you the tooth-to-tooth load spread; the 10% line is the common target.
- Fix runout before feed. Clean the taper and collet, shorten stickout, or step up the holder class — feed changes only redistribute the symptom.
- Re-check after thinning compensation. HSM paths run small actual chips; the compensated programmed feed is the right denominator.
How it works
Runout offsets every flute's effective radius, so the programmed feed per tooth is only an average. The bound is plain arithmetic:
f_max = f_z + TIR · f_min = f_z − TIR (floor 0) · guideline: TIR ≲ 0.1·f_z
The high tooth wears at the rate of its chip, not the average — which is why runout shortens tool life out of proportion to the numbers. The programmed feed itself comes from the chip load calculator, gets compensated at light stepovers by the radial chip thinning calculator, and turns into spindle terms via the SFM to RPM converter.
Worked example
Verified against the live calculator
A 4-flute programmed at 0.002 in/tooth, indicating
0.0005 in TIR at the edges:
share = 25% · chips swing 0.0025 / 0.0015 in · budget at 10% = 0.0002 in
Half a thou sounds respectable, but it is 2.5× the guideline at this feed: one flute carries 25% extra load all day while another rubs 25% light. The fix is in the holder stack — a budget of 0.0002 in points at precision collets or shrink/press-fit holders, not at the feed override.
Frequently asked questions
How much runout is acceptable on an end mill?
The commonly quoted guideline is TIR under ~10% of the feed per tooth. At 0.002 in/tooth that is a 0.0002 in runout budget — tighter than many drill chucks and worn collets deliver, which is why holder choice shows up directly in tool life.
What does runout do to chip load?
It splits the feed unevenly: the high flute cuts up to f_z + TIR while the low flute gets f_z − TIR. A "good" half-thou of TIR on a 0.002 in/tooth program swings tooth loads ±25% — one flute does 0.0025 in of work while another rubs at 0.0015.
What happens when runout exceeds the chip load?
Low flutes stop reaching the work entirely. The cutter becomes effectively single-tooth: one edge carries multiple feeds, wears fast, sounds loud once per revolution, and then hands its overload to the next edge as it dulls — the classic runout wear cascade.
Where should runout be measured?
At the cutting edges, in the actual holder, at working stickout. Spindle-taper readings flatter the setup: collet, cleanliness, gripping length and stickout each add their share, and tilt-type runout grows with overhang.
Method & assumptions
- Worst-case bound: f_z ± TIR is exact for 2 diametrically opposed flutes and conservative for higher flute counts under pure eccentricity.
- Tilt-type runout (growing along the flute length) at long stickout concentrates on the tip and can exceed this screen's bound locally.
- The ~10% figure is the widely quoted shop guideline, not a standard — tool makers' own runout limits govern where published.
- Surface-finish effects (one-flute witness lines, per-rev patterns) follow the same arithmetic but are not quantified here.