Yaskawa YRC1000 Interference Area Resolving TCP-Only Safety

1. Overview

The Yaskawa Motoman YRC1000 and the compact YRC1000micro controllers both use a Cartesian interference area primitive to define volumes the robot must not enter while an external axis is moving. The limitation that consistently surfaces in field deployments is that the controller evaluates interference area penetration against the Tool Center Point (TCP) only. Any other point on the arm — the wrist housing, elbow link, base shoulder, or a secondary tool body — is invisible to the geometric check.

This article documents the underlying architecture, the practical workarounds used on production cells, and how the optional Functional Safety Unit (FSU01 / FSU02) package on the YRC1000 closes the gap with safety-rated monitoring. The intended audience is robot integrators, controls engineers, and functional-safety specialists working with Yaskawa manipulator cells that include linear rails, positioners, or ganged servo axes.

2. Interference Area Architecture in YRC1000

The YRC1000 maintains interference areas as named Cartesian primitives stored in the controller's collision-detection database. Up to 32 simultaneous areas can be active per job, drawn from a pool of three shape types:

Each primitive is evaluated at the interpolation cycle of 4 ms (standard YRC1000 servo update). The check is performed by sampling the active TCP pose at the end of every interpolation cycle and asking whether that single XYZ point falls inside any active volume. If the answer is yes, motion is clamped to a stop at the boundary and alarm 4107 "INTERFERENCE" is raised.

The relevant job-control instructions are IMON (interference monitor on), IMONOFF, and the AREA family that toggles named areas from the pendant or from INFORM jobs. The I/O handshake used to coordinate with an external PLC or external axis drive is the dedicated interference area output group (system outputs #50080–#50127), where each bit is the live state of one defined area.

3. The TCP-Only Monitoring Limitation

The geometric engine treats the manipulator as a single point. A typical six-axis arm sweeping a path will, at some poses, place the TCP well clear of a conveyor-mounted nest while the elbow link passes within 20–40 mm of the nest wall. Because the engine never samples the elbow or forearm, the alarm does not fire and the collision is mechanical.

The condition is most acute in three cell topologies:

1. Tool handoff to a linear axis — the robot holds a part above a cup, then the linear slide moves under it. TCP clearance is fine, but the upper arm sweeps the slide's vertical column.
2. Overhead gantry with risers — the TCP moves in a plane below a structural column that the elbow crosses.
3. Positioner + manipulator — the wrist housing and the positioner tailstock occupy overlapping arcs even when the TCP stays in the safe pocket.

The native robot model does not provide per-link sampling. The DX200 carried a similar limitation; the YRC1000 retains the same architecture for backward compatibility with INFORM jobs and only the FSU option adds per-axis safety monitoring.

4. Workarounds for Arm-Wide Interference Checking

Five techniques are used in production. They are not mutually exclusive and most cells combine two of them.

4.1 Inflate the Interference Volume

The most common fix is to grow the cube or sphere so the worst-case envelope of the entire arm fits inside. Determine the maximum distance from TCP to any link along the planned path using a Monte-Carlo sweep of the joint envelope (2500 sampled poses is normally sufficient), then set the cube half-extents or sphere radius to that distance plus a margin. The penalty is a large dead zone in the workspace; for a GP12 the worst-case TCP-to-elbow distance is approximately 520 mm, which can shrink the usable reach by 18–25%.

4.2 Use Multiple Offset Cubes

Define two or three co-located cubes whose origin is the TCP and whose dimensions correspond to the wrist segment, the forearm segment, and the upper arm. Each is monitored independently. The dead zone is smaller than a single over-sized volume but alarm logic becomes more complex because multiple interference bits may assert for the same collision.

4.3 Servo Tracking of the Elbow

Use the ELSF elbow-sphere function (available on YRC1000 firmware YAS2.81 and later) to define a virtual TCP at the elbow. Place a separate sphere with center at the elbow virtual TCP and a radius equal to the link cross-section. The same approach works for the wrist with WTSF. Note that ELSF/WTSF are display features; they do not automatically drive the interference engine, so they must be combined with a second area definition.

4.4 Path Restriction via Cubic Interference

Where the external axis only intrudes at known poses, store the offending path segments as CUBIC interpolation blocks and gate them with an IF … INTERFERENCE() … THEN test that compares the current external-axis encoder reading against a safe window. The robot refuses to commit the path if the linear slide position falls inside the unsafe window, even if the TCP is in the clear.

4.5 Hand Off to the Functional Safety Unit

For a safety-rated solution that does not shrink the workspace, install the FSU01 option board and define ROBOT RANGE monitoring against the elbow or wrist joint values. The FSU samples each joint every 2 ms, is rated to PL d / Cat 3 per ISO 13849-1:2015, and drives a hardwired safe stop that is independent of the YRC1000's job-level interference engine. This is the only approach that is acceptable for a cell that requires third-party safety certification.

5. Functional Safety Unit (FSU) Overview

The FSU is a Yaskawa option module that mounts inside the YRC1000 backplane and is configured through the pendant under SETUP → FUNCTION SAFETY. There are two hardware revisions:

The FSU has its own processor running a certified safety logic core. It reads the joint encoders directly through a redundant path that bypasses the standard servo CPU. The safety core is rated to SIL 2 per IEC 61508 and to PL d, Category 3 per ISO 13849-1. It cannot be re-programmed by INFORM or by the standard job interpreter — the safety configuration is a separate, signed parameter set loaded by the safety engineer.

6. FSU Safety Functions in Detail

The FSU implements five core monitoring functions, plus a configurable safe-stop output.

6.1 AXIS RANGE

Monitors a single joint against a configurable minimum and maximum angle. Up to four ranges can be defined per joint. Hysteresis (typically 0.5°) prevents chatter. The function is useful for safety-rated soft limits that complement the job-level pulse limits in S1CFX files. Alarm on violation: 4421 "FSU AXIS RANGE".

6.2 ROBOT RANGE

Monitors a Cartesian volume defined by a cube or sphere. Unlike the job-level interference area, ROBOT RANGE samples the TCP and can be configured to sample a secondary point such as the elbow or wrist by selecting a virtual tool frame. Two ranges are permitted simultaneously. This is the function that resolves the TCP-only limitation described in Section 3. Alarm on violation: 4422 "FSU ROBOT RANGE".

6.3 AXIS SPEED MONITOR

Monitors per-joint angular velocity against a threshold. Typical use is limiting axis 3 (wrist) to 250°/s during collaborative operation. The monitor is independent of the programmed override and of the job-level OVRD setting; it can only be relaxed by an authorized safety-mode change. Alarm: 4423 "FSU AXIS SPEED".

6.4 SPEED LIMIT

Monitors the TCP Cartesian speed against a threshold. Used in collaborative mode to enforce the 250 mm/s limit of ISO/TS 15066:2016. Note that SPEED LIMIT and AXIS SPEED MONITOR can be active concurrently — both must pass for motion to be allowed.

6.5 TOOL ANGLE MONITOR

Compares the active tool orientation (Z-axis of the tool frame) against a configured reference direction. The function is intended for collaborative applications where the tool must remain pointed away from the operator. Tolerance band is configurable down to 0.1°. Alarm: 4425 "FSU TOOL ANGLE".

6.6 Safe-Stop Output Behavior

On any violation the FSU drives the safe-stop output (default: SF-OUT1 on the FSU connector). The default response is SS1 (Safe Stop 1) — controlled deceleration along the YRC1000 deceleration ramp, followed by removal of the enable signal to the servo pack once speed is below the threshold. SS2 (Safe Stop 2) — stop with drives energized, awaiting an external reset — is selectable per output. The stop output is a redundant pair of safety contacts rated at 24 V DC / 2 A.

7. PLC and Fieldbus Integration

The FSU is the safety source, but the application side of the interlock — for example, gating the external-axis drive — is typically handled by a safety PLC. Three integration patterns are common.

7.1 Hardwired Safety I/O (FSU01)

The FSU exposes 8 safe inputs and 4 safe outputs on a dedicated 24-pin Harting connector on the backplane. Outputs are sourced by the FSU's own 24 V supply; inputs are sinked from the safety PLC. The PLC's safety task reads the FSU outputs (in-area, overspeed, range violation) and decides whether to release the contactor that energizes the external axis.

7.2 PROFIsafe over PROFINET (FSU02)

The FSU02 board carries an additional PROFINET interface (part HW1481020-A) that maps the safety I/O onto PROFIsafe profile v2.4. Up to 32 safety slots are available, with the FSU acting as a PROFIsafe device (slaved to the safety PLC master such as a Siemens F-CPU or a Phoenix Contact RFC). The PROFINET cycle time is 1 ms, deterministic. Slot assignments are configured in Yaskawa's FSPRO tool before being loaded into the FSU.

7.3 CIP Safety on EtherNet/IP (FSU02 with adapter)

Some North American cells use the ODVA CIP Safety profile. The FSU does not include a native CIP Safety stack, but a third-party gateway (Rockwell 1734-IE4S or HMS AB7648) bridges PROFIsafe to CIP Safety. The gateway must be listed in the cell's safety validation report because it sits in the safety chain.

8. Configuration Procedure

The step-by-step below configures ROBOT RANGE on the FSU to monitor a second point on the arm — the elbow — so that interference checking is no longer limited to the TCP.

8.1 Prerequisites

• YRC1000 controller with FSU01 or FSU02 option enabled.
• Firmware YAS2.81 or later. Verify under MENU → SETUP → FUNCTION SAFETY → VERSION.
• Pendant programmer mode (key switch in TEACH or PLAY with the safety password entered).
• Joint map of the elbow TCP offset, obtained with the standard three-point method and stored as tool frame U2.

8.2 Step-by-Step

1. On the pendant, navigate to SETUP → FUNCTION SAFETY → ROBOT RANGE.
2. Select range index 1, set SHAPE = CUBE.
3. Enter corner 1 and corner 2 in base frame, sized to enclose the external-axis swept volume plus 25 mm of safety margin.
4. Under REFERENCE TOOL, select U2 (the elbow frame). The FSU will now sample the elbow position rather than the active tool.
5. Set HYSTERESIS to 5 mm to prevent oscillation at the boundary.
6. Assign an output: RESPONSE = SS1, output SF-OUT1.
7. Write the configuration to the FSU's signed parameter set. The pendant prompts for the safety password.
8. Reboot the controller. FSU parameters are loaded only at boot.
9. Run a validation sweep with all axes at 10% override, intentionally driving the elbow into the configured volume to confirm the safe-stop output asserts within 30 ms.

8.3 Verification

Confirm the FSU's safe-stop output by measuring the contact transition with an oscilloscope on SF-OUT1A and SF-OUT1B. Both contacts must open within 25 ms of the threshold being crossed. The safety validation report must record this measurement per ISO 13849-1 §4.5.4.

9. Safety Standards Compliance

The YRC1000 with FSU is designed to support cells rated to the following standards. The integrator is responsible for the overall risk assessment and validation; the FSU contributes only the safety functions it implements.

10. Commissioning and Verification

A cell that relies on ROBOT RANGE for collision avoidance must be commissioned with the following checks, recorded in the cell's safety validation report.

1. Geometric verification — measure the elbow TCP offset with the three-point method and compare against the saved tool frame. Tolerance: ±1 mm and ±0.5°.
2. Boundary trace — drive the elbow along the boundary of the configured cube at 100 mm/s. The safe-stop output must assert at every point of contact, with no more than 5 mm of overshoot.
3. Response time — measure the time from threshold crossing to safe-stop contact opening. Requirement: ≤ 30 ms total.
4. Fault injection — disconnect one encoder channel at the FSU connector and confirm the FSU raises alarm 4401 "FSU ENCODER FAULT" within 5 ms and enters the safe state.
5. Diagnostic coverage — verify the FSU's diagnostic test cycle by powering the controller for 24 hours. The built-in self-test fires every 8 hours; record each test pass in the maintenance log.

11. Troubleshooting Matrix

12. Selecting the Right Approach

Use the matrix below to choose between the workarounds discussed in Section 4.

Frequently Asked Questions