← Yarn Evenness and Imperfection Testing

Spectrogram and Variance-Length Curve Analysis

topic
Spectrogram analysis applies Fast Fourier Transform (FFT) to the yarn mass variation signal from the evenness tester, displaying power spectral density versus wavelength (1 mm–1,000 m range on logarithmic scale) to identify periodic defects invisible in time-domain CV% measurement. A periodic defect creates a 'chimney' peak on the spectrogram — peak wavelength identifies fault source: 3.5–4.5 cm wavelength chimneys indicate ring frame top roller eccentricity (roller diameter × π = wavelength); 80–120 cm peaks indicate roving frame bobbin winding pitch or main drafting zone roller eccentricity; 3.0–8.0 m peaks indicate card cylinder eccentricity; 15–25 m peaks indicate drawing frame delivery roller periodic slip. Diagnosing a 5 cm chimney on Ne 30 ring yarn: wavelength 5 cm in ring frame drafting (total draft 25×) corresponds to a periodicity of 5 cm × 25 = 125 cm in the roving or roving frame roller eccentricity — maintenance engineer checks 40 mm diameter delivery rollers on roving frame for worn bearings. Variance-length (V-L) curve (Uster Tester 6 software): plots CV% contribution from increasing cut lengths from 1 cm to 1,000 m — short-wave contribution CV%sw (1–50 cm) diagnoses carding and ring frame drafting quality; medium-wave CV%mw (50 cm–5 m) diagnoses roving frame and drawing autoleveller; long-wave CV%lw (>5 m) diagnoses bale laydown and blow room blending uniformity. Target values for Ne 30 export-quality ring yarn: CV%sw < 8%, CV%mw < 2.5%, CV%lw < 1.5% — exceeding any threshold triggers specific machine investigation before full-scale production. Wavelength amplitudes >3× baseline on spectrogram = significant periodic fault requiring immediate correction to prevent systematic fabric defects.

Role

Spectrogram and variance-length analysis are the most powerful diagnostic tools in spinning mill quality management, converting complex multi-source mass variation patterns in yarn into specific machine fault signatures that production engineers use to direct targeted maintenance — reducing defect diagnosis time from 2–4 days of systematic machine investigation to 15–30 minutes of spectrogram interpretation.

Explore "Spectrogram and Variance-Length Curve Analysis" on the interactive map →