IC693APU301 Low-Speed Encoder Jitter: Tuning the Sample Rate Time Window for Stable Velocity Feedback
Engineers commissioning GE Series 90-30 motion axes with the IC693APU301 often report severe velocity display oscillation below 10 RPM. This article explains how the speed calculation window interacts with quantization error. It also offers practical configuration guidance for the Proficy Machine Edition environment. Field data shows that extending the sample window from 100 ms to 500 ms cuts low-speed ripple by a factor of five. Moreover, the trade-off between display stability and control responsiveness is quantified below.

Why Low-Speed Measurement Becomes Unstable on the APU301
The IC693APU301 serves as a single-axis positioning module within the Series 90-30 platform. Its encoder interface accepts A-quad-B signals up to 200 kHz. The module derives velocity by counting encoder edges within a fixed time base. At high speeds, this method delivers excellent resolution. However, below 10 RPM, the pulse count per window becomes critically small.
Consider a 1024 PPR encoder running at 5 RPM. The pulse frequency is roughly 85 Hz. Within a 100 ms sample window, the module counts only 8.5 pulses. Therefore, a single edge of quantization error translates directly to a velocity error of about 0.6 RPM. This error appears as visible display jitter. In addition, the module’s 1 ms servo loop update rate does not automatically solve this quantization problem. Engineers must configure the sample window for velocity derivation independently.
Quantifying the Resolution Trade-Off Across Window Settings
Rockwell Automation published a reference table that illustrates the relationship between sample window duration and velocity resolution. For a 1024 PPR encoder with 4x interpolation, a 100 ms window yields a resolution of 0.146 RPM. Extending the window to 500 ms improves resolution to 0.029 RPM. Moreover, a 1000 ms window achieves 0.0146 RPM resolution.
These figures assume ideal signal conditions. In practice, mechanical vibration and electrical noise introduce additional variance. A 500 ms window provides a practical balance for most low-speed applications. It reduces quantization noise by a factor of five compared to the default 100 ms setting. The resulting display update rate of 2 Hz remains acceptable for operator monitoring. However, for control loop feedback, a 500 ms delay may introduce unacceptable phase lag.
Configuring the Sample Window in Proficy Machine Edition
The sample rate parameter resides within the APU301 channel configuration in Proficy Machine Edition. Navigate to the Hardware Configuration for the target rack. Open the APU301 module properties. Then select the Channel Parameters tab for the encoder input in use.
Locate the Velocity Averaging Time or Sample Rate field. The default value typically ships at 100 ms. For low-speed stability, increase this value incrementally. Begin at 250 ms and observe the display behavior. If jitter persists, proceed to 500 ms. Values beyond 1000 ms rarely provide additional benefit and degrade responsiveness. Document the final value in the configuration backup. The module stores this parameter in non-volatile memory and applies it after each power cycle.

Separating Display Smoothing from Control Feedback
A single velocity value cannot serve both display and control optimally. The display requires heavy smoothing for readability. In contrast, the control loop requires minimal lag for stability. Industrial practice recommends a dual-routine architecture. Routine A uses a short window for under-speed trips and interlocks. Routine B uses a long window with exponential moving average filtering for HMI display.
For the APU301, implement Routine A with a 100 ms sample and a 200 ms debounce timer. This prevents nuisance trips from single-sample dropouts. Implement Routine B with a 500 ms sample. Apply an EMA filter with α = 0.2 on the PLC side if the module does not natively support averaging. As a result, the effective time constant becomes approximately 2.5 seconds. This yields a visually stable display that still tracks actual speed changes within acceptable limits for operator awareness.
Verifying Signal Integrity Before Adjusting Parameters
Parameter changes cannot compensate for corrupted encoder signals. Before extending the sample window, verify the A/B/Z waveforms with an oscilloscope. The APU301 requires 5 VDC differential signals. A 24 VDC encoder connected directly to this module will not function and may damage the input buffers. Therefore, confirm the input voltage selection matches the encoder output type in the software configuration.
Check the shield termination. The encoder cable shield should land at one end only to avoid ground loops. Route encoder cables at least 12 inches from VFD power conductors. Cross power cables at right angles when separation is unavoidable. These physical measures eliminate the high-frequency noise that no sample window can filter effectively. After signal integrity is confirmed, the sample window adjustment will produce predictable and stable results.
Recommended Configuration Values for Common Low-Speed Scenarios
For a 1024 PPR encoder operating between 1 and 10 RPM, set the sample window to 500 ms. This yields a resolution of approximately 0.03 RPM. The display update rate of 2 Hz provides adequate operator feedback. For a 2500 PPR encoder in the same speed range, a 250 ms window achieves comparable resolution due to the higher pulse density per revolution. Set the window to 250 ms to preserve faster display response.
For applications requiring both low-speed stability and rapid fault detection, configure the under-speed trip with a 100 ms window and a 500 ms debounce. Configure the HMI display with a 500 ms window and EMA filtering. This dual-path approach satisfies both requirements without compromise. Record all final parameters in the maintenance documentation for future reference.
Author Insight: Why Sample Window Tuning Matters in Factory Automation
In industrial automation, low-speed axis stability remains a common commissioning challenge. Many engineers focus on the PLC scan time or the servo update rate. However, the velocity sample window often determines the actual display quality. From a practical standpoint, the APU301 rewards careful parameter tuning rather than hardware replacement. Moreover, this lesson applies broadly to control systems that derive speed from pulse counts. DCS and PLC platforms alike benefit from separating display filtering from trip logic. Therefore, commissioning teams should treat the sample window as a first-class tuning parameter.
Application Case: Stable Low-Speed Feedback on a Packaging Line
A factory automation team commissioned a Series 90-30 axis on a low-speed packaging conveyor. The HMI showed ±0.6 RPM jitter at 5 RPM, which confused operators. The team verified the 5 VDC differential encoder signals first. They then extended the display sample window to 500 ms and added EMA filtering in the PLC. As a result, the display stabilized at 2 Hz updates. Meanwhile, the under-speed trip kept a 100 ms window and a 500 ms debounce. This dual-path setup eliminated nuisance alarms and improved operator confidence in the control system.

Frequently Asked Questions
What causes velocity display jitter on the IC693APU301 at low RPM?
At low RPM, the encoder produces very few pulses per sample window. A single quantization edge therefore creates a large velocity error. Extending the sample window reduces this error.
Should I use the same sample window for trips and HMI display?
No. Use a short window with debounce for under-speed trips. Use a longer window with EMA filtering for HMI display. This dual-path approach balances safety and readability.
Can a 24 VDC encoder connect directly to the APU301?
No. The APU301 requires 5 VDC differential signals. A 24 VDC encoder may damage the input buffers. Always confirm the voltage selection in the software configuration.
Does a longer sample window improve control loop performance?
Not necessarily. A longer window adds phase lag, which can destabilize the control loop. Use long windows for display only and keep feedback paths short.
What sample window suits a 2500 PPR encoder at low speed?
A 250 ms window usually provides comparable resolution to a 500 ms window on a 1024 PPR encoder. The higher pulse density per revolution preserves faster display response.



