Tapered Bored Holes

title: Tapered and Bell-Mouthed Bored Holes: Diagnosis and Fixes category: troubleshooting tags: [boring, bored-hole, taper, bellmouth, deflection, spring-pass, stickout, L/D-ratio, boring-bar, chatter] compiled: 2026-04-11

Summary

A bored hole that's tapered or bell-mouthed is almost always a rigidity problem — tool deflection under cutting force. The boring bar bends away from the work under load, cuts undersize at depth, and the resulting hole measures bigger at the entry than the back. This article is the diagnostic ladder for killing taper in a bored hole, ordered from the easiest fix (spring pass — free, takes seconds) to the hardest (swap to a larger or damped shank). Work the list in order. Most tapered bores are fixed by step 1 or 2.

What "Tapered" Actually Means in a Bore

Three distinct failure modes that can look similar on a bore gauge but have different root causes:

1. Linear Taper

The hole is larger at the entry (front) than at the back (depth), with a roughly constant rate of change along the bore length. This is the classic deflection signature. The boring bar flexes away from the cutting force as it enters the bore. As the bar moves deeper, the unsupported length doesn't actually decrease (the bar is cantilevered from the holder, not from the bore wall), but the cut may stabilize as the chip load and engagement become consistent. The result: the first portion of the bore is oversize relative to the back. Typical magnitude on a marginal setup: 0.0005"–0.002" over a 3" bore depth. Measured with a telescoping gauge or bore gauge at 0.5" increments, you'll see a steady linear slope.

2. Bell Mouth

The hole flares at the entry — typically the first 0.100"–0.250" — but is cylindrical (or close to it) below that zone. Two common causes:

Entry impact: The cut is heaviest at the start before the chip establishes. The bar deflects maximally on entry, then settles as full engagement stabilizes the cut.
Retraction spring-back: The bar springs into the work on retraction (pulling out of the bore). If your retract move doesn't clear the bar from the wall before Z-axis withdrawal, the insert drags across the entry zone and opens it up. Fix: program a small radial retract (X-axis move away from the wall by 0.010"–0.020") before Z retract.

3. Barrel

The hole is wider in the middle than at either end. This is rare in single-pass boring. Usual causes: spindle runout that shifts the effective cutting diameter as the bar extends and retracts through resonance, a bent boring bar, or (on lathes) bed wear that creates a concave Z-axis travel path. Indicate the bar tip at multiple extensions to check for a bent shank. Check spindle runout at the nose — typical spec is under 0.0002" TIR; anything above 0.0005" needs attention.

The Diagnostic Ladder

1. Run a Spring Pass (Free Fix — Always Try First)

A spring pass is a second finish pass at the same commanded diameter — no depth change, just repeat the identical finish G-code line. On the first pass, the bar deflected away from the work by some amount (say 0.0004"). That 0.0004" of material is still standing. On the spring pass, the bar encounters that thin remaining stock. The cutting force is a fraction of the first pass, so deflection drops to nearly zero, and the bar cuts to commanded position.

Expected results:

For a 0.0008" taper: one spring pass typically reduces it by 50–80%.
Two spring passes usually bring you within 0.0002" total taper.
Cost: a few seconds of cycle time per pass. No tool wear worth measuring.

This is the single most underused fix in boring. Add it to every critical bore as standard practice on your finish pass. In CNC, it's literally duplicating one line of code. On manual machines, just run the bar through again at the same dial setting.

Programming note: Ensure the spring pass runs at the same feed and speed as the finish pass. Some programmers drop the feed further on the spring pass (e.g., 0.002 IPR instead of 0.004 IPR) — this can help but is usually unnecessary. The key is that the bar sees near-zero radial load.

2. Verify Stickout and L/D Ratio

Boring bar stickout (unsupported length from the holder face to the insert tip) is the dominant rigidity factor. Deflection scales with the cube of overhang length — double the stickout, get 8× the deflection. Rules of thumb by bar material:

Bar Shank Material	Max L/D Before Taper Problems	Example: 0.500" Ø Bar Max Reach
Steel shank	4:1	2.000"
Carbide shank	6:1	3.000"
Heavy-metal / vibration-damped (e.g., tuned-mass damper bars)	8–10:1	4.000"–5.000"

If your stickout exceeds the bar class limit, no amount of cutting-parameter adjustment will fully fix the taper. You must either shorten the stickout (flip the part, use a shorter bar, bore from both ends and blend) or move to a higher-class bar.

Practical check: Measure from the holder face to the insert tip with a caliper. Divide by the bar shank diameter. If the number exceeds the limit for your bar type, that's your problem. Fix the stickout before chasing anything else.

Also verify that the bar is fully seated in the holder — a bar that's slipped 0.25" out of the clamp has gained 0.25" of unsupported length you didn't intend.

3. Reduce the Cutting Load

Deflection is directly proportional to cutting force. Reduce force → reduce deflection → reduce taper. Targets for a finish boring pass:

Depth of cut (radial): 0.005"–0.010" per side for finish. If you're leaving 0.020" for finish, you're asking the bar to do too much work. Rough closer to size first.
Feed rate: 0.002"–0.005" IPR for finish boring. Feed has a strong, roughly linear effect on cutting force. Dropping from 0.006 to 0.003 IPR cuts force nearly in half.
Cutting speed: Increase slightly (10–15%) to reduce force per unit of chip area, but don't exceed insert grade limits. Speed has less effect on force than feed does.
Insert nose radius: This one is counterintuitive. A larger nose radius (e.g., 0.8 mm vs. 0.4 mm) spreads the cut over more edge and can improve finish, but it increases radial cutting force because more of the edge is engaged. On marginal-rigidity setups (long stickout, thin walls), a smaller nose radius (0.2–0.4 mm) often produces less taper despite a slightly rougher finish. Test both.
Insert geometry: Switch to a sharp, positive-rake insert (CCGT, DCGT, VCGT type designations in ISO) instead of a general-purpose negative-rake geometry (CCMT, DCMT, VNMG). Positive-rake inserts cut with substantially lower forces — typical reduction of 20–40% vs. negative-rake at the same parameters. This is critical for long-reach boring. The trade-off is that positive inserts are single-sided (no flip) and the edge is less robust, but for a light finish pass that doesn't matter [harvey_performance].

4. Check Runout at the Boring Bar Tip

Runout in the toolholder or spindle amplifies with stickout. A 0.0005" TIR at the holder face can become 0.001"+ at the insert tip on a 4:1 stickout bar. This creates an uneven cut per revolution — one side heavy, one side light — which produces both taper and roundness errors.

How to check: Mount an indicator on the turret (lathe) or table (mill). Indicate the bar shank near the tip. Rotate the spindle by hand (lathe: rotate the chuck by hand or use jog). TIR at the tip should be under 0.0003" for precision bores.

Common causes of excessive runout:

Dirty or burred toolholder bore or bar shank — clean both with a Scotch-Brite pad.
Worn or damaged collet/clamp. Replace.
Set screws that push the bar off-center — use a split-sleeve holder or hydraulic/shrink-fit holder for critical bores.
Spindle bearing wear — if runout is in the spindle nose, that's a maintenance issue, not a tooling fix.

5. Address the Machine Side

If everything above is clean and taper persists, the problem is the machine geometry:

Spindle-to-turret alignment (lathes): The spindle axis must be parallel to Z-axis travel. Misalignment produces taper on every bore regardless of tooling. Check with a test bar and indicator: mount a precision ground bar in the chuck, indicate it at two Z positions separated by 6"–8". Typical spec: under 0.0002" per 6" of travel for a production lathe. Correction is a machine leveling and alignment procedure — consult the machine manual.
Spindle-to-table squareness (mills): Tram the head. A head that's tilted will produce taper in a bored hole. Tram within 0.0002" over a 6" sweep for critical boring work.
Bed/way wear: On older flat-bed lathes, way wear in the most-used travel zone creates a dip that changes the effective alignment over Z travel. Run an indicator along the ways to check. This is a scraping/regrinding issue.
Thermal growth: A machine that's been sitting cold will shift alignment as the spindle and bed warm up. Run the machine at operating speed for 15–30 minutes before making finish boring cuts on critical parts. On long production runs, bore dimensions may drift as the machine reaches thermal equilibrium — monitor the first 5–10 parts.

Specific Case: The 3" Deep Bore in 4140 Prehard

A 3" deep bore in 4140 prehard (28–32 HRC), carbide boring bar, measuring 0.0008" taper — tight at the bottom, oversize at the entry. This is the textbook deflection-away-from-work signature: the bar is bending under radial cutting force, cutting less material at depth.

Fix sequence:

Spring pass — Run the finish pass again at the same commanded diameter. Expected result: taper drops from 0.0008" to ~0.0003".
Reduce finish feed — If running above 0.005 IPR, drop to 0.003 IPR. At 4140 prehard, use coated carbide at 400–550 SFM with a positive-rake CCGT or DCGT insert.
Check stickout — A 3" bore depth requires at least 3.25"–3.50" of bar reach (insert must clear the bore fully). On a 0.500" diameter steel shank bar, 3.5" reach is 7:1 L/D — far beyond the 4:1 limit. This is the root cause. Options: - Best: Move to a 0.750" diameter carbide shank bar. At 3.5" reach, L/D = 4.7:1, well within the 6:1 carbide limit. - Alternative: Use a 0.625" or 0.750" damped bar if available. At 3.5" reach on a 0.750" damped bar, L/D = 4.7:1 — the damping provides additional margin. - Budget option: If stuck with the 0.500" steel bar, minimize the finish depth of cut to 0.003"–0.005" radial, drop feed to 0.002 IPR, and run two spring passes. You may get within 0.0003", but you're fighting physics.
Final approach: Light finish pass at 0.003" radial depth, 0.003 IPR feed, followed by one spring pass. Expected result: straight bore within 0.0002"–0.0003" on the carbide shank bar.

Quick-Reference Checklist

Symptom	Most Likely Cause	First Fix
Linear taper, oversize at entry	Bar deflection (L/D too high or feed too heavy)	Spring pass, then check L/D
Bell mouth at entry only	Entry impact load or retraction drag	Program X retract before Z retract; reduce entry feed
Barrel shape (wide in middle)	Spindle runout, bent bar, or way wear	Indicate bar tip and spindle; check ways
Taper that changes part-to-part	Thermal drift or chip packing	Warm up machine; check coolant/chip evacuation
Taper plus chatter marks	Deflection AND vibration (compound problem)	Shorten stickout, reduce feed, switch to damped bar

[[chatter-vibration]] — Vibration vs. deflection diagnosis: if the bore has chatter marks AND taper, you have both problems. Fix deflection first.
[[speeds-feeds-fundamentals]] — Finish-pass parameter selection for boring.
[[boring-bars-catalog]] — Specific bar selection by L/D ratio and bore diameter.
[[workholding-rigidity]] — Part deflection can mimic tool deflection. Thin-wall bores need support.