A deep hole drilling machine that produces accurate hole depth on the first part of the morning and progressively deeper holes as the day warms up has a spindle growth problem. The spindle shaft expands as it heats — the drill tip moves axially relative to the workpiece — and the CNC does not know the spindle has grown. The result: holes that are consistently deeper than programmed, tool over-travel that can damage the drill or workpiece, and cycle time variation as operators adjust offsets to compensate. Spindle growth compensation corrects this — automatically adjusting the Z-axis position for thermal expansion.
Thermal Growth Fundamentals
Sources of Heat
| Heat Source | Location | Temperature Rise | Contribution to Growth | Control Method |
|---|
| Front spindle bearing | Bearing race — rolling elements | 10–30°C above ambient | 40–60% of total growth | Bearing preload — lubrication — cooling |
| Rear spindle bearing | Bearing race — rolling elements | 5–15°C above ambient | 10–20% of total growth | Bearing preload — lubrication — cooling |
| Motor (integral spindle) | Rotor — stator | 20–40°C above ambient | 10–20% of total growth | Motor cooling — thermal insulation |
| Coolant passing through spindle | Coolant union — spindle bore | 5–20°C above ambient (coolant temp) | 5–15% of total growth | Coolant temperature control — chiller |
| Ambient temperature | Machine environment | Diurnal variation — seasonal | 5–10% of total growth | Climate control — machine location |
| Friction from drilling | Cutting zone — drill | Localized — does not conduct far up spindle | Negligible direct effect | Coolant flow — cutting parameters |
Growth Characteristics
| Parameter | Typical Value | Notes |
|---|
| Growth rate — initial warm-up | 0.01–0.03 mm per minute (first 30 minutes) | Fastest growth occurs in first 30–60 minutes — thermal equilibrium takes 2–4 hours |
| Total growth — cold to equilibrium | 0.05–0.20 mm (varies by spindle size) | Larger spindles grow more — high-speed spindles grow more |
| Growth direction | Axial (Z-axis) — spindle extends | Spindle shaft expands toward workpiece — all growth is in one direction |
| Time to 90% of equilibrium | 60–120 minutes | Depends on spindle design — cooling system — operating speed |
| Cooling contraction rate | 0.005–0.015 mm per minute (first 30 minutes after stop) | Spindle contracts as it cools — growth curve is not symmetrical with cooling curve |
| Repeatability — daily | ±0.01–0.03 mm (if warm-up is consistent) | Inconsistent warm-up produces inconsistent growth — same warm-up produces repeatable growth |
| Effect of speed changes | Higher speed = more growth — faster growth rate | Growth rate is approximately proportional to spindle speed |
| Effect of coolant temperature | Cooler coolant = less growth — longer warm-up | Coolant temperature should be stabilized — ±1°C |
Measurement Methods
| Method | Accuracy | Resolution | Cost | Setup Complexity | Best For |
|---|
| Dial indicator — mechanical | ±0.005 mm | 0.001 mm | Low | Low — mount indicator on spindle face against fixed reference | Quick check — occasional measurement — verification |
| Dial indicator — digital | ±0.003 mm | 0.001 mm | Low–Moderate | Low — same as mechanical | Production measurement — data logging capable |
| Laser displacement sensor | ±0.002 mm | 0.0001 mm | Moderate–High | Moderate — laser alignment required | High-accuracy measurement — continuous monitoring |
| Capacitance probe | ±0.001 mm | 0.0001 mm | High | High — requires clean environment — stable mounting | Highest accuracy — research — calibration reference |
| LVDT (linear variable differential transformer) | ±0.002 mm | 0.0001 mm | Moderate | Moderate — contact measurement — spring-loaded probe | Continuous measurement — integration with compensation system |
| Eddy current probe | ±0.003 mm | 0.0001 mm | Moderate–High | Moderate — non-contact — insensitive to coolant | Coolant environment — non-contact measurement |
Measurement Procedure
| Step | Action | Detail |
|---|
| 1 | Mount measurement device | Dial indicator or laser on spindle face — reference against fixed machine structure (not workpiece) |
| 2 | Establish cold reference | Machine off for minimum 8 hours — record indicator zero at spindle face — record machine temperature |
| 3 | Record baseline | Temperature at spindle housing — coolant temperature — ambient temperature — indicator reading |
| 4 | Start spindle at operating speed | Use typical drilling speed — no cutting load |
| 5 | Record growth at intervals | Every 5 minutes for first hour — every 15 minutes for next 2 hours — every 30 minutes thereafter |
| 6 | Continue until equilibrium | No measurable growth for 30 minutes — typically 2–4 hours |
| 7 | Record data | Time — spindle speed — indicator reading — bearing temp — housing temp — coolant temp |
| 8 | Stop spindle | Allow to cool — record contraction data at same intervals |
| 9 | Repeat for different speeds | If machine operates at multiple speeds — repeat procedure for each speed |
| 10 | Generate growth curve | Plot growth vs time — determine time constant — equilibrium value — warm-up rate |
Compensation Models
Model Comparison
| Model Type | Accuracy | Complexity | Sensors Required | Adaptability | Best For |
|---|
| Linear model | ±0.02–0.05 mm | Low | One temperature sensor (bearing or housing) | Low — fixed relationship | Machines with predictable growth — single operating speed |
| Multi-sensor model | ±0.01–0.02 mm | Moderate | 2–4 temperature sensors | Moderate — accounts for different heat sources | Machines with varying speed — multiple operating conditions |
| Lookup table | ±0.01–0.03 mm | Low–Moderate | Speed and time inputs | Low — fixed table | Machines with consistent warm-up cycle — repeatable operation |
| Adaptive model | ±0.005–0.015 mm | High | Temperature sensors + periodic verification | High — self-correcting | Precision drilling — varying conditions — frequent speed changes |
| Real-time measurement | ±0.002–0.010 mm | Very high | Non-contact displacement sensor | Maximum — measures actual growth | Highest precision — research — critical applications |
| Thermal network model | ±0.005–0.015 mm | High | Multiple temperature sensors | High — physics-based model | Complex spindles — varying loads — development work |
Implementation Comparison
| Implementation Method | CNC Integration | Update Rate | User Interface | Cost | Typical Application |
|---|
| CNC macro program | Direct — runs on CNC | Per cycle or per part | CNC display | None (software only) | Simple compensation — one speed — consistent conditions |
| CNC axis compensation parameter | Direct — built-in function | Continuous (per servo cycle) | CNC parameter screen | None (parameter setting) | Siemens — Fanuc — Heidenhain — built-in thermal comp function |
| External compensation unit | Via analog or fieldbus | Continuous | Separate display | $1,000–5,000 | Machines without built-in compensation — retrofit |
| PLC-based compensation | Via machine PLC | Per scan cycle | HMI display | $500–2,000 (PLC programming) | Machines with PLC control — custom implementation |
| Real-time sensor feedback | Direct — sensor to CNC | Continuous | CNC display | $3,000–15,000 | Highest accuracy — critical applications |
Calibration Procedure
| Step | Action | Detail | Verification |
|---|
| 1 | Machine cold soak | Minimum 8 hours off — measure baseline temperature | All temperatures stable — within ±0.5°C over 1 hour |
| 2 | Mount measurement system | Dial indicator or laser on spindle face — reference against fixed machine structure | Indicator zeroed — no drift — reference stable |
| 3 | Record cold reference | Spindle position — indicator reading — all temperatures | Baseline established |
| 4 | Program warm-up cycle | Spindle speed = typical drilling speed — no cutting — run 30 minutes | Record data at 5-minute intervals |
| 5 | Continue warm-up | Run until thermal equilibrium (no growth for 30 min) | Typically 2–4 hours — record data per measurement procedure |
| 6 | Generate compensation curve | Plot growth vs time — fit curve — determine time constant | Curve fit R² > 0.95 |
| 7 | Implement compensation | Set CNC parameters or program compensation | Per compensation model selected |
| 8 | Verify compensation — cold start | Run verification cycle from cold — measure actual vs programmed position | Error < 0.01 mm with compensation active |
| 9 | Verify compensation — warm | Run after thermal equilibrium — measure actual vs programmed position | Error < 0.01 mm at all temperatures |
| 10 | Verify across speeds | Repeat at all operating speeds — modify compensation if needed | Error < 0.015 mm across speed range |
| 11 | Document | Record compensation parameters — curve — verification results | Document for future reference — recalibration schedule |
Verification and Validation
| Test | Method | Acceptance Criteria | Frequency |
|---|
| Cold accuracy | Drill test part from cold start — measure hole depth | Hole depth within ±0.02 mm of programmed depth | After calibration — monthly |
| Warm accuracy | Drill test part after 2 hours operation — measure hole depth | Hole depth within ±0.02 mm of programmed depth | After calibration — monthly |
| Growth tracking | Measure actual spindle growth vs compensation command | Growth error < 0.01 mm at all temperatures | After calibration — quarterly |
| Temperature sensor correlation | Compare sensor readings to spindle growth | Correlation R² > 0.95 | After calibration — annually |
| Repeatability | Run 3 identical warm-up cycles — compare growth curves | Variation < 0.01 mm between cycles | After calibration — annually |
| Speed variation test | Run at different speeds — verify compensation accuracy | Error < 0.015 mm at all speeds | After calibration — after speed range change |
Troubleshooting
| Problem | Symptom | Likely Cause | Corrective Action |
|---|
| Compensation under-corrects | Holes deeper than programmed when warm | Growth model underestimates actual growth — sensor reading low — model coefficients wrong | Verify growth measurement — recalibrate — check sensor location |
| Compensation over-corrects | Holes shallower than programmed when warm | Growth model overestimates — sensor reading high — model coefficients wrong | Verify growth measurement — recalibrate — check sensor contact |
| Compensation works intermittently | Hole depth varies — sometimes correct — sometimes not | Temperature sensor loose — sensor wiring intermittent — electrical noise | Check sensor mounting — check wiring — shield sensor cable |
| Compensation slow to respond | Holes at start of warm-up correct — middle incorrect — end correct | Time constant in model incorrect — thermal mass not accounted for | Adjust model time constant — verify warm-up curve |
| Compensation drifts over weeks | Compensation was correct — now incorrect | Sensor calibration drift — machine mechanical change — coolant temp change | Recalibrate — check sensor — verify machine condition |
| No compensation effect | Compensation active but hole depth does not change | Compensation applied to wrong axis — scale factor zero — parameter not active | Verify compensation parameters — check CNC parameter documentation |
| Compensation oscillates | Z-axis position oscillates — hole depth varies | Sensor feedback loop unstable — update rate too fast — gain too high | Reduce gain — slow update rate — add damping |
FAQ
What is spindle growth and why does it matter in deep hole drilling?
Spindle growth is the thermal expansion of the spindle shaft as it heats up during operation — the spindle shaft increases in length as its temperature rises. In a deep hole drilling machine, the spindle shaft is oriented horizontally (typically) or vertically — growth occurs axially, moving the drill tip relative to the workpiece. A typical deep hole drilling spindle grows 0.05–0.20 mm from cold to thermal equilibrium — this change in Z-axis position means the drill tip is 0.05–0.20 mm closer to the workpiece than when the machine was cold. Spindle growth matters because it directly affects hole depth accuracy: on a cold-start first part, the hole depth may be correct (or slightly shallow if the operator compensates for growth). After 1–2 hours of operation, the spindle has grown — the drill tip is now closer to the workpiece — the next holes are drilled deeper than programmed. The error accumulates as the machine warms — the maximum error is at thermal equilibrium. For deep hole drilling applications, where hole depth tolerance may be ±0.05 mm or tighter, the 0.05–0.20 mm growth error is significant — it can push hole depth out of tolerance. Additionally, the changing Z position changes the effective feed rate and drilling conditions as the drill progresses through the workpiece — inconsistent conditions lead to inconsistent hole quality. Spindle growth compensation — using temperature sensors and a correction model — automatically adjusts the Z-axis position to cancel the effect of thermal growth — maintaining consistent hole depth from cold start through full operating temperature.
How do I measure spindle growth on my deep hole drilling machine?
To measure spindle growth on your deep hole drilling machine: mount a dial indicator (mechanical or digital) against the spindle face or a fixture mounted to the spindle face — the indicator must reference against a fixed point on the machine structure that does not move with the spindle (the headstock casting, the machine base, or a separate stand — not the workpiece or workholding that may have its own thermal movement). Set the indicator to zero with the machine cold (machine has been off for at least 8 hours — record the spindle housing temperature, coolant temperature, and ambient temperature). Start the spindle at the operating speed you typically use for drilling — run at this speed continuously (no cutting load — just the spindle rotating). Record the indicator reading and temperatures at regular intervals: every 5 minutes for the first hour, every 15 minutes for the next hour, and every 30 minutes thereafter — continue until the indicator reading stabilizes (no change for 30 minutes — typically 2–4 hours). The indicator reading at each interval is the spindle growth at that time — the difference between the cold reading (zero) and the final stabilized reading is the total spindle growth. Plot the data as growth vs time — this is the spindle growth curve for that spindle speed. Repeat the measurement at each spindle speed used in production — growth varies with speed. The total growth value and the shape of the growth curve are the foundation for the compensation model. For higher accuracy measurement: use a laser displacement sensor or capacitance probe — these provide better resolution and can be logged automatically. For repeatable results: run the same warm-up cycle each time — consistent warm-up produces consistent growth.
What is the best compensation model for spindle growth?
The best compensation model for spindle growth depends on the machine, the operating conditions, and the required accuracy. For most deep hole drilling machines, a multi-sensor model with 2–4 temperature sensors provides the best balance of accuracy, complexity, and cost. The model uses temperature sensors at: the front bearing housing (the primary heat source for most spindles — accounts for 40–60% of total growth), the rear bearing housing (secondary heat source — accounts for 10–20%), the spindle housing (measures bulk temperature), and the coolant at the spindle inlet (accounts for coolant temperature effects). The model calculates growth as a weighted sum of the temperature rises at each sensor: Growth = K1 × ΔT_front + K2 × ΔT_rear + K3 × ΔT_housing + K4 × ΔT_coolant. The coefficients (K1, K2, K3, K4) are determined during calibration — they represent the contribution of each temperature to total growth. This model is accurate to ±0.01–0.02 mm across varying operating conditions. For machines with a single dominant operating speed and consistent warm-up cycle, a simpler linear model (single temperature sensor at the front bearing housing) may be adequate — accuracy ±0.02–0.05 mm — with lower sensor and implementation cost. For the highest precision requirements (±0.005 mm or better), real-time measurement with a non-contact displacement sensor provides direct growth measurement — the sensor output is fed back to the CNC as a real-time Z-axis offset — this is the most accurate but most expensive approach. The key to any compensation model: the temperature sensors must be located where they accurately represent the thermal state of the spindle — sensors too far from the heat source have a slow response — sensors placed where coolant splashes read coolant temperature — not spindle temperature.
How do I implement spindle growth compensation on my CNC machine?
Spindle growth compensation implementation depends on the CNC control type. For modern CNCs (Siemens 828D/840D, Fanuc Oi/30i/31i, Heidenhain TNC 7): use the built-in thermal compensation function — these controls have a dedicated thermal compensation parameter set — input the temperature sensor locations, growth coefficients, and time constants per the control manufacturer's programming manual — the control applies the compensation automatically to the Z-axis position. For older CNCs or controls without built-in thermal compensation: implement via CNC macro program — measure temperature sensor inputs through the CNC's analog inputs or fieldbus — calculate growth per the compensation model — apply the calculated correction as a Z-axis offset (work offset shift or tool offset adjustment). The macro is called at the start of each drilling cycle to update the offset. For machines with a PLC (programmable logic controller): implement the compensation in the PLC — read temperature sensors — calculate growth — send the correction to the CNC via the axis offset word or fieldbus — the PLC can update the compensation continuously (every PLC scan cycle) — providing smoother compensation than per-cycle macro updates. For machines without any built-in compensation capability: install an external compensation unit — a standalone controller that reads temperature sensors, calculates growth, and outputs a correction signal (analog voltage, fieldbus, or digital offset) to the CNC — this is a retrofit solution that can be added to any machine. Verification after implementation: run the verification procedure from the calibration section — drill test parts from cold through warm — measure hole depth — confirm the compensation keeps hole depth within the target tolerance (±0.01–0.02 mm for most applications). The most common implementation error is incorrect sensor placement — sensors that respond too slowly or read the wrong temperature produce compensation that does not track actual growth.
How often should spindle growth compensation be recalibrated?
Spindle growth compensation recalibration frequency: recalibrate after any spindle maintenance (bearing replacement, preload adjustment, lubrication change — any change to the spindle's thermal behavior changes the growth characteristics — recalibrate after every spindle service event). Recalibrate after any significant change in operating conditions (if the machine's operating speed range changes significantly — if the coolant system is modified — if the machine is moved to a different location with different ambient temperature conditions). Verify accuracy monthly (drill a test part from cold start — measure hole depth — if depth is within tolerance, the compensation is still accurate — if depth has drifted, recalibrate). Recalibrate annually as a minimum (even without any changes, sensor drift and mechanical wear can affect compensation accuracy — annual recalibration captures these subtle changes before they cause out-of-tolerance parts). When to verify more frequently: machines that run critical parts (aerospace, medical, precision — verify weekly or before each critical production run). Machines with large ambient temperature swings (seasonal changes of more than 10°C can affect compensation — verify at the start of each season). Machines that run intermittently (frequent start-stop operation creates inconsistent thermal conditions — verify more often than machines running continuously). The recalibration procedure is the same as the initial calibration: mount measurement device — run warm-up cycle — record growth — verify compensation model — update parameters. Keep a log of calibration results — trend data shows when the spindle's thermal behavior is changing, which may indicate developing bearing problems.
Spindle thermal growth is a predictable, measurable phenomenon that directly affects hole depth accuracy in deep hole drilling. Measure the growth curve for each operating speed using a dial indicator or laser — data log temperatures at the front bearing, rear bearing, housing, and coolant. Implement compensation using the CNC's built-in thermal compensation function, a macro program, or an external unit. Verify compensation by drilling test parts from cold through warm — hole depth should be consistent within ±0.02 mm across all temperatures. Recalibrate after spindle maintenance — verify monthly — recalibrate annually. Proper spindle growth compensation eliminates a major source of dimensional variation — delivering consistent hole depth from the first part of the morning through the last part of the shift. This article reflects industry practice as of 2026.
