Spindle troubleshooting guide

How to Diagnose Spindle System Failure on a Twin Spindle Machining Center

Two spindles share one control, one hydraulic circuit and one thermal environment, so a fault on one head often shows up on the other. This guide walks through the checks we run in our own shop, in the order that finds the cause fastest.

Get Instant Quote Talk to an Engineer

6 diagnostic checksStep-by-step orderSymptom-to-cause table

Diagnosing spindle system failure on a twin spindle machining center

On this page

1
What to settle before you touch the spindle
2
Why spindle system failure looks different on two heads
3
Reading the first symptom correctly
4
The four circuits that cause most stops
5
Six checks, in the order that saves the most time
6
Symptom, likely cause, and the check that confirms it

Key takeaways

What to settle before you touch the spindle

Split the two heads firstRun the spindles independently. If only one head faults, the fault is local, not shared.

Thermal drift mimics bearing failureA cold spindle and a warm spindle give different readings. Measure at the same temperature.

Most faults are not the bearingsClamping, drive tuning and sensor wiring account for the majority of stops we see.

Record the alarm code and the axis positionThe code narrows the circuit. The position narrows the mechanical cause.

Pull the spindle lastRemoval costs days. Rule out the drive, encoder and hydraulic first.

How a twin spindle machine fails differently

Why spindle system failure looks different on two heads

On a single-spindle machine, a fault has one path. On a twin-spindle machining center, two heads sit on the same bed, draw from the same hydraulic power unit and breathe the same air. That shared hardware is what makes diagnosis faster and also what makes it misleading.

The first question is never what broke. It is whether the fault follows the head or follows the machine. Run spindle A alone at 6,000 rpm for ten minutes with spindle B stopped. Then swap. If the alarm stays with the same head, the cause is local: bearings, clamp, encoder, tooling. If the alarm follows the operating condition instead, look at the shared supply.

We see this most often with hydraulic clamping. One power unit feeds both drawbars. A slow pressure build shows up as an unclamp alarm on whichever head happens to be changing tools when pressure dips. Operators blame the spindle. The gauge on the power unit tells the real story.

Temperature is the second shared variable. Two spindles in one enclosure heat each other. A head that runs fine alone can drift past its thermal limit when both run together for an hour. Log spindle growth at 15-minute intervals, not just once at start-up.

1
Local faultStays with one head regardless of load or duty cycle.
2
Shared faultMoves between heads with hydraulic, pneumatic or thermal conditions.
3
Log both spindlesRecord growth and current draw at the same intervals.

Symptoms and what they point to

Reading the first symptom correctly

Spindle system failure rarely announces itself as one clean symptom. It arrives as a cluster: a noise, a surface finish change, a current spike, then an alarm. The order matters. Whatever appeared first is usually closest to the cause.

A growl that starts at 3,000 rpm and fades above 8,000 rpm points to bearing preload or a worn ball track. A growl that gets louder with speed points to something rotating out of balance, often a holder or a damaged taper. Same noise, different repair.

Surface finish is the quiet indicator. If Ra drifts from 0.8 μm to 1.6 μm on the same program, same tool, same material, the spindle is moving more than it should. Check runout at the taper first. A reading above 0.005 mm TIR is enough to explain the finish.

Current draw on the spindle drive is the cheapest instrument you have. A bearing on its way out pulls more amps at the same cut. Trend the drive load over a week rather than reading it once.

1
Speed-dependent noiseBearing preload or ball track wear.
2
Load-dependent noiseBalance, holder or taper damage.
3
Finish driftRunout above 0.005 mm TIR.
4
Rising ampsFriction building before an alarm trips.

Where to look first

The four circuits that cause most stops

In our shop, spindle stops cluster into four circuits: the drive and its tuning, the encoder and feedback cable, the hydraulic clamp, and the bearings themselves. Bearings come last more often than people expect.

Drive tuning drifts after a motor or cable change. If someone replaced an encoder cable and did not re-run the auto-tune, the spindle can fault under acceleration while running perfectly at steady speed. The alarm looks mechanical. The fix is a parameter set.

Feedback cable damage is easy to miss because it is intermittent. A cable that rubs against a moving cover only faults at certain axis positions. Wiggle test the cable while watching the position error display. If the error jumps, you found it without touching the spindle.

Hydraulic clamp pressure should sit inside the range the machine builder specifies, commonly 4.5–6.0 MPa for a 40-taper drawbar. Below that, tools release slowly and the unclamp sensor trips late. Above it, seals wear early. Check both heads at the same gauge point.

Only after these three circuits read clean do we pull a spindle. On a twin-spindle machine, pulling one head means realigning two, and that is a two-day job with a laser and a test bar.

1
DriveRe-tune after any motor, cable or parameter change.
2
FeedbackWiggle test the encoder cable across the full axis travel.
3
ClampVerify 4.5–6.0 MPa at both drawbars.
4
BearingsPull only after the other three read clean.

Step by step

Six checks, in the order that saves the most time

1
Log the alarm and the machine stateWrite down the alarm code, the axis position, spindle speed, tool number and whether both heads were running. Do this before clearing anything. The code narrows the circuit; the position narrows the mechanical cause.
2
Run each spindle aloneRun head A at 6,000 rpm for 10 minutes with head B stopped, then swap. Note the exact speed where noise or vibration starts. A fault that stays with one head is local. A fault that follows the duty cycle is shared.
3
Measure runout and taper contactIndicate the taper with a test bar at 50 mm from the gauge line. Anything above 0.005 mm TIR needs attention. Blue-check the taper contact; below 80 percent contact area, re-grind before blaming the bearings.
4
Wiggle test the feedback and drive cablesWith the machine in jog, flex the encoder and power cables along the full axis travel while watching position error. A jump in error identifies the cable, not the spindle. This test takes ten minutes and saves a teardown.
5
Check clamp pressure at both drawbarsFit a gauge at the drawbar inlet. Expect 4.5–6.0 MPa and a release time under 0.5 s. Slow release on one head with normal pressure points to a worn seal or a sticking piston in that head.
6
Trend spindle current and temperatureRecord drive load and spindle housing temperature every 15 minutes for one hour, both heads running the same program. A head that climbs steadily while the other flattens out has a friction problem building.
7
Pull the spindle only when the data agreesIf runout, taper contact, cable test, clamp pressure and current trend all come back clean and the fault remains, the bearings or the rotor are the remaining cause. At that point removal is justified, not before.

Symptom to cause

Symptom, likely cause, and the check that confirms it

Use this table when the alarm code alone does not point anywhere.

Symptom	Likely cause	Confirming check
Noise only at low rpm	Bearing preload loss	Run at 3,000 rpm, listen, then 9,000 rpm
Noise rising with speed	Holder or taper imbalance	Balance the holder, blue-check the taper
Unclamp alarm on one head	Low or slow clamp pressure	Gauge at the drawbar inlet, expect 4.5–6.0 MPa
Position error at one axis point	Damaged feedback cable	Wiggle test across full travel
Fault only under acceleration	Drive auto-tune not re-run	Re-run auto-tune after cable or motor change
Finish drifts over a shift	Thermal growth	Log runout cold and after one hour
Both heads fault together	Shared hydraulic or air supply	Check the power unit gauge, not the spindles
Amps climb week over week	Friction or lubrication loss	Trend drive load daily, check oil lines

FAQs

Questions we get from maintenance teams

How long should a spindle run before I trust a runout reading?

Take the reading cold first, then again after 30 minutes at the cutting speed you normally use. Thermal growth of a few microns is normal. What matters is the difference between the two heads on the same machine.

If head A grows 8 μm and head B grows 3 μm over the same period, head A has a lubrication or preload problem even if both readings are inside tolerance cold.

Can a bad encoder look exactly like a bad bearing?

Yes, and this is the most common misdiagnosis we see. An encoder losing counts at one axis position produces vibration that feels mechanical and often shows up in the finish.

The wiggle test separates them. Flex the feedback cable through the full travel while watching position error. A bearing will not care where the axis is. A cable will.

Should I replace both spindles when one fails?

Not automatically. It depends on running hours and whether the two heads have shared the same duty cycle.

If both heads have logged similar hours on the same programs, replacing one usually means the second follows within a few months. If one head runs a light finishing cycle and the other roughs, their wear rates differ and you can replace them independently.

What clamp pressure should we hold on a 40-taper drawbar?

Most builders specify 4.5–6.0 MPa at the drawbar inlet, with tool release under 0.5 s. Hold both heads at the same point on the range.

Running at the low end saves seal life but risks slow release and false unclamp alarms. Running above the range shortens seal life and can distort the drawbar. Pick one value inside the range and keep both heads there.

What data should we log so the next fault is easier?

Speed at which noise starts, runout at the taper, clamp pressure, drive load and housing temperature, all taken at the same intervals on both heads.

A single reading tells you almost nothing. A week of the same reading tells you which head is drifting, and drift is what turns into a failure.

When is pulling the spindle the right call?

When the drive, feedback cable, clamp circuit and thermal behavior all test clean and the fault still repeats at the same speed and load.

On a twin-spindle machine, pulling one head means realigning two. That is a two-day job with a test bar and a laser, so it should be the last step, not the first.