SigTrap Home Success Stories Technical Your Project About IMPACT Contact Us

The SigTrap System provides an automated real-time objective noise testing system for motor and component noise quality decisions. A typical system is shown below with a description of the key components and their relationships. We are experts at configuring and integrating a system that will best fit your needs on your production floor.

Typical SigTrap System Diagram

sensor on a motorSensors - are used to acquire critical vibration and acoustic information from the component being tested. The most common application in a production environment is to use one or more magnetically attached accelerometers strategically placed on the component under test. Magnetically attached accelerometers offer a good compromise between signal fidelity and isolation from the environmental noise common in the production environment. Accelerometers can be manually or automatically placed (via a one axis pick and place unit) on the component under test or in some cases, on the fixture holding the component. Microphones have also been successfully used in some applications on the factory floor. The strategic placement of the sensors and correlation of the factory-floor measurements to subjective ground truth is a difficult and application specific problem which IMPACT Engineering and its customers have successfully solved in numerous applications.

NVClient HMI screenThe SigTrap NV+ Signal Processor- receives signals from the Sensors and the Fixture Controller controlling the testing fixture. The accelerometer or microphone signals are powered and conditioned directly by the SigTrap NV+. The Fixture Controller provides information about the component type, serial number or bar-code and test start and stop signals. The SigTrap NV+ Signal Processor acquires the real-time signals, measures and analyzes key statistics and metrics and completes a classification of component quality (Good/Bad). Complete component quality test results are sent to the Fixture Controller for feedback to the Operator or direct action from the production line equipment. The SigTrap also automatically performs diagnostic analysis on the sensor signals to identify Bad Cable, Mis-Located Sensor and Dead Sensor faults for production maintenance.

The NVClient Software - operates locally on a Windows PC to provide data logging, operator information display, system setup and production status and historical information. Typically, the local PC will be accessible via a touch screen interface with full security modes. Multiple SigTrap systems can be networked via a single NVClient PC and the PC itself is commonly interfaced into the plant network so that information collected via the SigTrap system can be integrated into plant production statistics databases and tools.

Signal Processing Algorithms, Methods and Expertise

First Pass Yield resultsThe SigTrap System provides a wide range of unique and powerful noise analysis algorithms that are put into effect as necessary to analyze and classify your component noise. Each of the following methods have been successfully used for one or more of our customers:

First Pass Yield, Detectability and False Alarm Rate Analysis - IMPACT provides off-line tools to analyze first pass yield, detectability rates and false alarm rates. Classification methods can be evaluated and tested off-line using real world production data and special known-noise components and special case golden sample components.

Adaptive Harmonic Analysis and Order Tracking - Small motors and rotating components are notorious for speed variations which smear normal spectral (FFT) measures and prevent accurate order readings with traditional methods. IMPACT uses an adaptive harmonic analysis tracking algorithm, which works on the data itself, to track the speed of the motor and un-wrap the smearing effect of the  estimated order strengths. The results are highly accurate order strength and frequency measures.

Signal Rotation Signature Analysis - The SigTrap can extract a 'Single Rotation Signature' which is an averaged time-series pattern representing a single rotation of the main rotating shaft of the motor. This signature is valuable in identifying undesirable brush / commutator bar interactions that typically result in sharp ticking or buzzing sounds. Extracting this response pattern can be key to classifying and understanding the root cause of the noise problems.

Transient noise exampleTransient Signal Analysis and Detection (Flutter, Pop, Click, Thump, Rattle)
- Transient signals are difficult to characterize and detect. IMPACT has put a number of methods and algorithms to work capturing and classifying these noises in our customer's real world production systems.

Zwicker Sound Quality Measurements: Loudness, Specific Loudness and Fluctuation Strength
- Zwicker Psycho-acoustic measures for Loudness, Specific Loudness and Fluctuation Strength have been implemented in SigTrap solutions. The Loudness and Specific Loudness methods include a configurable transfer function used to convert the measurements of the production line to the equivalent of those measured on the components in the Acoustic sound lab.

Root Cause Problem Classification with Adaptive and Customer Configured Rules
- SigTrap systems can provide more than a Pass / Fail decision. A number of motor customers have the capability to identify root cause problems and actually configure root cause rules for special parts or special causes as they see fit.

WOW variation exampleWOW (Rotational Speed Fluctuation) Measurement - The amount of allowable speed variation during steady state operation can be accurately measured and analyzed with the SigTrap system.

Transient Analysis and Detection of Gear Mesh Noise - Components with gearing systems present special noise problems relating to gear mesh errors and variations. SigTrap systems have been used to identify, detect and quantify rumble, 'helicopter' and other transient noises sourced at the gear mesh interface of the components.

Adaptive Fixture Brake Control - One SigTrap system uses long-term adaptive feedback control for particle brake settings and effectively compensate for the set-point drift inherent in the brakes.

Automated Sensor Diagnostics Algorithms - We provide automated tools to detect and identify common system faults such as bad cables, a sensor which has been mis-located and a sensor which is not connected (either the sensor or the cable is fatally broken).

End to End System Calibration (1Grms)
- We provide a simple, end-to-end calibration method which assures that your measurements are accurate and calibrated to 1 Grms or 1 µPa. This is particularly important for accelerometers which vary from part to part.