Why Do CNC Machines Not Have a Z Axis?
Every machining center has a Z axis. When the machine behaves as if it does not, the fault is almost always in the reference, the setup or the control offset, not in the machine design. This page is for engineers and buyers who see Z faults on the floor and need to know what to check first.

In this article
- 1
- 2
- 3
- 4
- 5
- 6
CNC machines Z axis faults: symptom, cause, action
Match the symptom you see on the machine to the most likely cause before you touch a parameter.
| Symptom | Likely cause | What to do |
|---|---|---|
| Z travel stops short of the part | Work offset or tool length set wrong | Re-measure tool length and re-probe Z zero |
| Z reads a value but never moves | Z axis servo or drive in fault | Read the drive alarm code, clear and re-home |
| Z moves only in one direction | Soft limit or travel limit active | Check travel limits against the program Z range |
| Deep pocket floor is off by 0.1 mm | Tool pull-out or thermal growth | Re-touch the tool, add a spring pass |
| Angled face cannot be reached | 3-axis setup cannot index the part | Move the job to a 4-axis or 5-axis center |
| Z drifts after a long run | Spindle and ballscrew thermal growth | Warm up 20-30 min, re-touch, keep coolant steady |
| Finished floor has a visible step | Z backlash or lost motion | Check backlash, inspect thrust bearing preload |
Fix the reference before you change the machine
Nine out of ten Z faults come from tool length, work offset or thermal drift. Measure those three first, and only then look at the axis hardware or consider a different machine.
Why do CNC machines not have a Z axis is the wrong question
Almost every vertical machining center sold today has three linear axes: X, Y and Z. On a VMC the spindle moves up and down along Z while the table carries the part in X and Y. On a horizontal machine the Z axis usually drives the column toward the fixture instead. The axis is there. The question comes up because the Z axis is the one you cannot see moving in a clean, obvious way.
What people actually mean when they ask why do CNC machines not have a Z axis is that the machine is not moving in Z, or moves less than expected, or the display shows a Z value that does not match the part. Those are faults. They have causes you can find in an hour with a dial indicator, a probe and the alarm log.
The Z axis also differs from X and Y in one important way. It carries the cutting load in the direction of the spindle thrust, so any error in tool length, thermal state or preload shows up directly in the floor depth of a pocket. X and Y errors tend to shift a feature sideways. Z errors change how deep it cuts.
That is why the same fault looks dramatic in Z and mild in X or Y. A 0.05 mm error in X on a slot is invisible. The same 0.05 mm in Z on a sealing face is a leak.
- 1Z is present on 3, 4 and 5-axis machinesRotary axes are added on top of the three linear axes, they do not replace Z.
- 2Most Z faults are reference faultsWork offset, tool length and probe results are the first three things to check.
Reference, tool length and work offset errors
The most common reason a machine appears to have no Z axis is that Z zero is in the wrong place. If the operator touches off on the top of a vise jaw instead of the part face, every Z move in the program shifts by that difference. The machine still moves. It just cuts air or cuts too deep.
Tool length offsets are the second trap. A tool measured 2 mm short will drive the spindle 2 mm lower than the program expects. On deep pockets with long reach tools the error grows with every re-touch, because the pull-out in the holder is not the same each time. Re-measure the tool in the presetter, not by eye against the part.
Work offset and tool length must agree on the same datum. Mixing a G54 set from the vise with tool lengths measured from the spindle gauge line is a classic recipe for a scrap part. Pick one datum, write it on the setup sheet, and keep it for the whole run.
On a probe-equipped machine, verify the probe stylus length after any crash or tip change. A bent stylus still triggers, but it triggers at the wrong height, and the error is often between 0.05 mm and 0.3 mm.
- 1Check the datum twiceWork offset and tool length must both reference the same surface.
- 2Re-measure after every crashEven a light bump can shift a pull-stud or a stylus tip.
Backlash, thermal growth and spindle drift
If the Z axis moves but the depth drifts over a long run, the cause is usually thermal. A spindle that has run for 30 minutes is longer than a cold spindle. On a machine cutting aluminium at high rpm the Z position can shift by 0.02 mm to 0.06 mm between the first part and the twentieth. Warm up the machine for 20 to 30 minutes at the running spindle speed before you touch off.
Backlash is the second mechanical cause. A worn thrust bearing or a loose locknut lets the Z axis lose motion when it reverses direction. Check with a dial indicator: approach a stop from below, zero the indicator, then approach from above. The difference is lost motion. Anything above 0.01 mm on a finishing axis needs attention.
Tool pull-out looks like an axis fault but is not. A long end mill in a collet holder can move 0.02 mm to 0.1 mm under heavy axial load. The machine holds position. The tool slides. Use a hydraulic or shrink-fit holder for deep axial cuts and keep the flute length as short as the feature allows.
On machines with a counterbalance cylinder, a slow gas leak changes the load on the Z servo. The axis may still position, but it will fault under acceleration or drift when the brake releases. Check counterbalance pressure against the maintenance chart.
- 1Warm up before touching off20 to 30 minutes at cutting speed removes most thermal drift.
- 2Measure lost motionApproach from both directions and compare indicator readings.
When the limit is the machine, not the axis
Sometimes the Z axis works perfectly and the part still cannot be cut. A 3-axis machine cannot reach a face that points sideways without a re-fixture. The Z stroke is fine. The setup cannot present the feature to the tool. That is a machine capability limit, not a Z fault.
If a part has features on four or five faces, deep cavities with undercuts, or a free-form surface that must be cut in one pass, a 3-axis setup will force multiple operations. Each re-fixture adds stack-up error. Moving the job to a 4-axis or 5-axis center removes those setups and keeps the datum intact.
At GreatLight we run 16 simultaneous 5-axis machining centers, 12 four-axis mills and 27 three-axis machines, so the choice is made on geometry rather than on what is free. A simple plate with one flat face belongs on a 3-axis machine. An impeller or a medical implant with compound angles belongs on a 5-axis center.
Z travel also sets the limit. Our large machine has a travel of 4,000 × 400 × 150 mm, which suits long, shallow parts. The medium platform at 750 × 1,150 × 550 mm and 600 × 600 × 600 mm covers most enclosure and manifold work. If the part is taller than the Z stroke, no parameter change will fix it.
- 1Count the setups firstIf a part needs three or more fixtures, check 4-axis or 5-axis instead.
- 2Match Z stroke to part heightA part taller than the Z travel cannot be cut on that machine.
Step by step: diagnose a Z fault
Work in this order. Stop as soon as the indicator or the probe disagrees with the control.
- 1Read the alarm log firstOpen the alarm history and note the servo, drive and overtravel codes. Clear them, then re-home Z and watch the reference return. A repeat alarm points to hardware, not to the program.
- 2Verify the reference returnHome the machine and check that the Z reference value matches the last known value within 0.01 mm. A shift here means the encoder, the dog or the grid offset has moved.
- 3Re-measure tool lengthMeasure every tool in the set on a presetter, not against the part. Compare with the stored offsets and correct anything off by more than 0.01 mm.
- 4Confirm the work offset datumTouch off Z zero on the same surface the setup sheet names. Check G54 through G59 for stale values left by a previous job.
- 5Check lost motion with an indicatorApproach a stop from below and from above. A difference above 0.01 mm means backlash or a loose thrust bearing.
- 6Run a warm-up cycleSpindle on at cutting speed for 20 to 30 minutes. Re-touch the first tool afterwards and compare the Z offset to the cold value.
- 7Cut a test feature and measureFace a small pad or a shallow pocket, then measure the floor depth. Compare with the nominal and adjust the tool offset by the difference.
- 8Log the resultWrite the corrected offsets and the warm-up time on the setup sheet so the next run starts from a known state.
Common questions
Do any CNC machines really have no Z axis?
A 2-axis lathe has X and Z but no Y, so the naming can confuse people. Some flat-bed machines are described by the axes they control, not by a fixed three-axis rule.
Every machining center that cuts a pocket or a face has a Z axis. If a machine has no Z, it is a saw, a laser or a waterjet, not a milling center.
Why does my Z axis move but not reach the programmed depth?
Check the tool length offset first, then the work offset. If both are correct, measure lost motion with a dial indicator and inspect the thrust bearing and locknut.
Thermal growth can also account for 0.02 mm to 0.06 mm over a long run. Warm up the spindle and re-touch before blaming the control.
Can I add a Z axis to a 2-axis machine?
No, not in any practical sense. Adding a controlled Z axis means a new column, a new drive, a new control configuration and a new safety assessment.
It is almost always cheaper to buy a machine that already has the axis than to retrofit one onto a lathe or a router.
What tolerance can a healthy Z axis hold?
On a well-maintained machining center we hold ±0.005 mm on critical Z dimensions, with surface finish between Ra 0.8 μm and 1.6 μm on a standard machined face.
Tighter finishes down to Ra 0.2 μm are possible on specific features after finishing passes and polishing when the drawing calls for it.
Does a 5-axis machine remove all Z setup error?
No. It removes the error from repeated re-fixturing, because the part stays in one setup. The tool length and thermal checks still apply.
The gain is in geometry access and datum stability, not in a magic correction of offsets.
How do I know if the fault is the machine or the part?
Cut a test pad on a scrap blank with the same tool and the same offsets. If the test pad meets nominal, the machine is fine and the part setup is the problem.
If the test pad is off by the same amount, the fault is in the machine or in the offsets, and the part drawing is not the cause.
Send us the part and the tolerance
Upload the drawing and we return a quotation with a free DFM analysis within 12 hours. Production can start within 24 hours, and parts ship in 3–5 days.
12-hour quote100% inspectionNo minimum order quantityNDA on request