Tool changers on CNC machining centres: failure patterns and early signs

A tool changer works through the same sequence of movements thousands of times. When it fails, the machine stops, often in the middle of a program and not infrequently with a tool that is seated properly neither in the spindle nor in the gripper. Many failures, however, have a history: isolated fault messages that disappear after acknowledgement, a change time that slowly gets longer, a noise that was not there before. This article describes the design of automatic tool changers (ATC) on CNC machining centres, the typical failure patterns and the early signs you can observe with simple means.

How the tool magazine and changer are built

The system consists of two parts that have to work together.

The tool magazine holds the tools ready. Common designs are disc or carousel magazines, chain magazines for larger numbers of tools, and rack or cassette magazines. A drive positions the required pocket at the transfer point. Depending on the design, a swivelling or tilting movement of the tool pocket is added.

The changer brings the tool from the magazine into the spindle. Two basic forms are common:

• Double gripper: a gripper arm grips the old tool in the spindle and the new one at the transfer point at the same time, withdraws both, rotates through 180 degrees and inserts them again. The motion sequence is often generated by a cam mechanism driven by a motor. The cam defines stroke and rotation mechanically.
• Pick-up principle: the spindle itself moves to the magazine pocket, puts down the old tool and picks up the new one. There is no separate gripper arm; instead the axes and the magazine perform the movement.

The drives vary by manufacturer and year of manufacture. There are changers with a servo motor, whose movement the CNC control regulates like an axis, variants with an induction or geared motor and limit switches, and pneumatically or hydraulically actuated movements. Which variant your machine has is stated in the machine builder's documentation. This determines which signals are available at all.

The spindle side is also part of the change: the clamping system in the spindle holds the tool by spring force, usually via a disc spring stack. To release it, a hydraulic or pneumatic release cylinder pushes against it. During the change, an air blast cleans the taper. Many machines then check via an air pressure sensor whether the tool is seated flat.

The change cycle step by step

For troubleshooting it helps to see the change as a chain of individual steps. Each step has a condition that must be met before the next one begins. The CNC control usually monitors this via limit switches, sensors and time limits.

Typical failure patterns on the tool changer

The following failure patterns are found on many designs. Your own fault history is the best guide to which ones are common on your machine.

Strictly speaking, some of these failure patterns concern the spindle: clamping force, taper and face contact. In practice, however, they are reported as a "tool changer problem" because they become noticeable during the change. When looking for the cause, it is worth considering both sides.

Change time as a simple indicator

The simplest early sign is the duration. A healthy changer works very consistently, because the cam mechanism and the control sequence are fixed. If a step becomes stiff, a pressure takes longer to build up or a switch responds with a delay, the change takes longer. Often the duration also starts to fluctuate more before its mean value rises.

For meaningful observation you should separate two things:

• Total chip-to-chip time: includes axis movements, spindle run-up and program logic. It is important for production, but too coarse for diagnosis.
• Time of the actual change: from the start to the end of the change movement, ideally per step. This is where a mechanical change shows most clearly.

A comparison is only fair if the conditions are the same. Depending on the machine, heavy and long tools require a slower movement, and the travel in the magazine depends on which pocket is being fetched. It therefore makes sense to compare within comparable changes, for example separately by tool group or only for the change movement without the magazine search.

The timestamps can be obtained in various ways: from signals of the machine's PLC, from the tool change command in the program sequence or from state changes that the CNC control reports via an interface. Which signals are available on your machine is best clarified with the machine builder or by looking at the interface documentation.

Motor current per change cycle as a principle

If the changer is driven by a closed-loop controlled motor, its current during a change contains an image of the mechanics. The current is approximately proportional to the torque. If friction in the cam mechanism rises, for example due to a lack of lubrication, or if a movement jams, the motor needs more torque in that section.

The principle is simple: for each change, record the current profile over time or over the rotation angle of the drive and compare it with the profile of a healthy change. The peak values at start-up, the mean value in the uniform sections and the shape of the profile are meaningful. A change that builds up over many tool changes is an indication. A single outlier is not yet one.

In practice there are two hurdles. First, a change is short. Standard interfaces such as OPC UA supply values at intervals of tenths of a second to whole seconds. With these you capture a few points per change at best. For a real current profile you need the trace function of the CNC control or a separate measurement. Second, not every changer has a drive whose current the control knows. With pneumatics or simple motors with limit switches, times, pressures and switch states remain the available signals.

Maintenance, lubrication and alarm history

Most changer problems can be prevented or found early with consistent maintenance. Intervals and lubricants are specified by the machine builder. Regardless of the design, these points have proven themselves:

• Lubrication: lubricate the cam mechanism, gripper and magazine guides and chain drives as specified by the manufacturer. Check gearbox oil levels.
• Cleanliness: remove chips and coolant residue from grippers, magazine pockets and limit switches. Dirty switches are a frequent cause of sporadic faults.
• Gripper: visually inspect gripper fingers, springs and locking mechanisms for wear and play.
• Pneumatics and hydraulics: check supply pressure, air service unit, filters and leaks. Check that the air blast at the spindle taper works.
• Spindle clamping system: measure the clamping force regularly with a clamping force gauge and document the value. Keep the taper and tool holders clean and undamaged.
• Transfer position: after a collision or an intervention, have the alignment of spindle orientation and change position checked.

An underrated source is the alarm history of the CNC control. Changer fault messages are often acknowledged and forgotten. If you evaluate them regularly, you see patterns: the same time-out in the same step, clustered at a particular magazine pocket or after the weekend. An alarm that rises from once a month to several times a week is an early sign, even if each individual one looks harmless. SINUMERIK and HEIDENHAIN TNC keep alarm and error logs for this, which can be viewed on the control and in some cases exported. Also note what was done after a fault. Only the combination of message and action makes the history useful.

What we offer today

The tool changer is on our list because the signals for it are in principle available in many cases: timestamps, states and fault messages from the CNC control, and with closed-loop drives also the load. However, we will only offer monitoring once we have shown on real machines that the assessment is reliable. Until then it is not part of our offer. In any case, our access to the control remains strictly read-only.

Today we monitor the spindle and its bearings. How we go about this is described in the guide to spindle monitoring. How faults on individual assemblies affect availability is covered in the article on the main causes of low OEE. The current scope of our service is described on the page Spindle monitoring for CNC milling centres.

Frequently asked questions