Machine tool metrology

How to Use a Laser Doppler Interferometer as a Warranty Method for CNC Machine Tools

A laser doppler interferometer turns machine geometry into numbers you can compare against a spec. This guide is for maintenance engineers and quality leads who need to prove a machine still holds tolerance after a move, a crash or a rebuild. Read it and you will know which checks to run, what limits to compare against, and when the method is the wrong tool.

Get Instant Quote Talk to an Engineer

Linear positioningStraightnessSquarenessBacklash

Laser doppler interferometer setup for checking CNC machine tool accuracy

On this page

1
Key takeaways
2
What a Laser Doppler Interferometer Measures on a Machine Tool
3
When a Laser Doppler Interferometer Beats a Test Cut
4
The Five Checks to Run on a Machine Tool Network
5
Setup, Warm-Up and Temperature Rules
6
How to Read the Report Without Overreacting
7
Step by Step: Measuring an Axis
8
Typical Limits and What They Mean

Quick answer

Key takeaways

It measures distance, not surface finishThe beam tracks axis travel and reports error as a number.

Five checks cover most warranty claimsPositioning, repeatability, straightness, squareness and backlash.

Compare against the machine spec sheetA generic tolerance is useless without the axis and travel length.

Temperature decides the verdictA 1 °C drift over 4,000 mm can look like a real fault.

What the instrument actually does

What a Laser Doppler Interferometer Measures on a Machine Tool

A laser doppler interferometer splits one beam into a reference path and a measurement path. The measurement path hits a retroreflector mounted on the moving part of the machine. When that part moves, the return beam shifts in frequency, and the instrument counts that shift to work out how far the part actually traveled.

The output is distance versus commanded position on every axis. If the control says move 300 mm and the optics say 300.035 mm, the difference is the axis error. That is the whole point of the tool. It does not tell you why the error exists, only where it is and how big it is.

Modern systems also take angular optics, straightness optics and a squareness kit. Swap the lens and you can check yaw, pitch, roll and the perpendicularity between two axes without moving the machine to another setup. That is why a single visit can cover a whole machine tool network.

It is a warranty method because the numbers are traceable. When a machine is delivered, an axis fails after six months, or a rebuild is signed off, both sides argue from the same interference data instead of a test cut.

1
Positioning accuracyDifference between commanded and actual travel along one axis.
2
RepeatabilitySpread of results when the axis returns to the same point.
3
StraightnessSideways drift of the axis over its full stroke.
4
SquarenessPerpendicularity error between two axes, in arc seconds.

When it is the right tool

When a Laser Doppler Interferometer Beats a Test Cut

A test cut shows the finished result. Interferometry shows the cause. If a part is out of tolerance, a test cut tells you the machine is wrong but not which axis, how much, or whether the error is linear or periodic. The interferometer breaks that down axis by axis.

Use it when a machine has just been installed, moved between buildings, rebuilt, or after a collision. Use it when a warranty dispute needs evidence, and when two parties disagree about what a machine is capable of. Also use it before a long production run on a tight part, so you know the machine is sound before you cut 500 pieces.

There are limits. The instrument measures the machine, not the part. It cannot see spindle thermal growth under load, tool deflection, fixture movement or material springback. A machine can pass every linear check and still cut a bad part because of the spindle, the tool or the setup.

It also needs a clear optical path. On a machine with bellows, way covers or a trunnion in the way, the beam gets blocked and you lose part of the stroke. That is a setup problem, not a measurement problem, but it limits what you can prove on a fully enclosed machine.

1
Good fitPost-install, post-move, post-crash, rebuild sign-off, warranty disputes.
2
Weak fitChasing surface finish, chatter or a single bad feature.
3
Needs clear beam pathWay covers and trunnions block part of the stroke.

The five checks

The Five Checks to Run on a Machine Tool Network

Positioning accuracy is the first check. Run the axis over its full stroke in small increments, usually 10 to 25 points per meter, and record the error at each point. Compare against the machine spec. A typical machining center spec sits near ±0.005 mm over the travel length, but the number only means something with the axis name and travel attached.

Repeatability is the second check. Go back to the same points from both directions five to seven times. The spread tells you whether the control and the drive can hold a position, not just reach it once. A machine with good accuracy and poor repeatability will scrap parts unpredictably.

Straightness and squareness follow. Straightness measures sideways drift over the stroke. Squareness measures how far two axes are from 90°, reported in arc seconds. A 20 arc second error over 500 mm is roughly 0.05 mm of drift, which is enough to matter on a long part.

Backlash is the last check. Command a small move in one direction, then reverse. The gap between the two readings is lost motion in the screw, the coupling or the thrust bearing. If backlash grows after a crash, the mechanical damage is usually in the bearing or the nut, not in the control.

1
Log every axis separatelyA single average hides which axis is drifting.
2
Do a warm-up cycle firstRun the spindle and axes 20 to 30 minutes before measuring.
3
Keep the reportIt becomes the baseline for the next check.

Setup and environment

Setup, Warm-Up and Temperature Rules

Set the instrument on a tripod or a magnetic base that is rigid and off the moving part. If the optics move while the machine moves, the reading is worthless. Point the beam at the retroreflector on the axis under test and align until the return signal is strong.

Warm the machine up before you measure. A cold machine and a warm machine do not have the same geometry. Run the spindle and the axes for 20 to 30 minutes, then let the machine sit for 10 minutes so the structure settles. Measure the same way every time, because the baseline is only useful if the conditions match.

Temperature is the biggest single source of argument. Steel grows about 11 μm per meter per °C. Over a 4,000 mm axis, a 1 °C drift is roughly 44 μm of apparent error. Log the shop temperature, the machine temperature and the material, and compensate if the instrument supports it.

Air and floor vibration matter on long travel. Do not measure next to a running grinder or a stamping press. If the readings jump by more than a few microns between passes with no machine movement, the environment is the problem, not the machine.

1
Rigid mountOptics must not sit on the moving axis.
2
Warm-up20 to 30 minutes of axis and spindle motion.
3
Log temperatureNote shop, machine and part temperature on the report.
4
Avoid vibrationStay away from presses and grinders during the run.

Reading the result

How to Read the Report Without Overreacting

A single failed point is not a failed machine. Look at the shape of the error curve. A linear slope over the whole stroke points to pitch error or thermal growth. A repeating wave points to screw or gear error. A random scatter points to vibration, a loose mount or a dirty optic.

Compare the run to the last baseline. If the numbers moved by a few microns and the machine still cuts good parts, watch it. If the numbers moved by tens of microns after a crash, pull the machine out of production and inspect the mechanics before you cut anything else.

Remember that interferometry measures unloaded motion. A machine that passes cold can still fail under cutting load because of spindle growth, tool push-off or fixture flex. Use the report as one input, not the only one, and back it with a test cut when the part tolerance is tight.

Finally, keep the machine class in mind. A 4,000 mm gantry and a 500 mm machining center do not share limits. Judging both against ±0.005 mm will either pass a bad machine or fail a good one.

1
SlopePitch error or thermal growth.
2
WaveScrew or gear error.
3
ScatterVibration, loose mount or dirty optic.

How to run it

Step by Step: Measuring an Axis

1
Clean the machine and clear the pathWipe the ways, open the covers you can open, and check that the beam can reach the full stroke. A blocked beam means a partial measurement, which is not evidence.
2
Warm up the machineRun the spindle at a moderate speed and move the axes for 20 to 30 minutes. Stop, let it settle for 10 minutes, then start.
3
Mount the opticsPut the interferometer on a rigid tripod off the machine. Mount the retroreflector on the axis under test. Align the return beam until the signal strength is high and stable.
4
Set the programUse 10 to 25 measurement points per meter of travel. Run both directions. For a 500 mm axis that is 5 to 12 points each way; for a 4,000 mm axis it is 40 points or more.
5
Run the measurementKeep the shop door closed and stop other machines if you can. Note any reading that jumps without a command. Repeat the whole run if the spread looks wrong.
6
Note backlash separatelyCommand a 0.05 mm move in one direction, then reverse. Record the difference. Do this at three or four points along the axis.
7
Repeat for every axisRun X, Y and Z with the same warm-up and the same point spacing. Then run straightness and squareness with the matching optics.
8
Write the reportInclude axis, travel, point count, temperature, date and the pass or fail against the spec. Keep a copy for the next check.

Pass and fail

Typical Limits and What They Mean

Limits depend on the machine class and travel length. Use these as a starting point, then compare against the machine spec sheet.

Check	Typical limit	Likely cause if it fails	First thing to do
Positioning accuracy	±0.005 mm over travel	Screw wear, pitch error, thermal drift	Re-run warm and cold to separate drift
Repeatability	Under 0.003 mm spread	Loose coupling, encoder fault, stick-slip	Check coupling and encoder signal
Straightness	Under 0.01 mm per 500 mm	Way wear, misaligned rail, foundation	Check rail alignment and level
Squareness	Under 20 arc seconds	Frame twist, collision damage, settling	Re-level the machine and re-check
Backlash	Under 0.005 mm	Thrust bearing, nut wear, loose coupling	Inspect bearing and nut preload
Thermal drift	Under 0.02 mm after warm-up	Spindle heat, coolant, ambient swing	Log temperature over 4 hours

The short version

Use a laser doppler interferometer when you need traceable axis data for a warranty claim, a rebuild sign-off or a post-move check. Use a test cut when the problem is the part, not the machine. Know which question you are answering before you set up the optics.

Get Instant Quote Talk to an Engineer

FAQs

Frequently Asked Questions

Can a laser doppler interferometer replace a ballbar test?

No. They answer different questions. A ballbar sweeps a circle and shows the combined contour error of two axes, which is fast and good for spotting servo and geometry problems.

An interferometer measures one axis at a time with traceable distance data. Use the ballbar for a quick health check and the interferometer when you need numbers to compare against a spec or a warranty claim.

How often should we check a machine?

After installation, after any move, after a crash, and after a rebuild. In normal production, once a year per machine is a common baseline.

If a machine runs three shifts on tight parts, check it twice a year. The cost of an hour of measurement is small next to a scrapped batch.

Do we need to measure every axis every time?

At minimum, measure the axes used for the tightest features. On a 5-axis machine that usually means all linear axes plus the rotary table.

If time is short, run positioning and repeatability on all axes, and add straightness and squareness on the axes that carry the critical dimensions.

What temperature should the shop be at?

A stable temperature matters more than a cold one. A shop that sits at 20 °C all day is easier to measure than one that swings from 15 °C to 28 °C.

Log the actual temperature at the start and end of the run. If it moved more than 1 °C, treat the numbers with care and repeat the measurement.

Why do the numbers look worse after a warm-up?

Because a warm machine has grown. If the machine was leveled and set cold, the geometry can shift once the spindle and ways heat up.

Measure warm, and if the cold and warm runs disagree, the difference is the thermal error you need to compensate or control.

Can we do this in-house or should we call a service?

In-house works if you have the optics, a trained operator and a stable shop. The instrument is only half the job; alignment and warm-up habits decide the result.

For warranty disputes, an independent measurement with a dated report carries more weight than an internal one. Either way, keep the raw data.

Send Us the Drawing and the Tolerance

We machine to ±0.005 mm with 100% inspection before shipment, and we share inspection reports on request. Upload a drawing and get a quotation with free DFM analysis within 12 hours.

12-hour quote100% inspectionNo minimum order quantity

Get Instant Quote Talk to an Engineer