Development of SMD 32.768 kHz tuning fork-type crystals

Abstract

Abstract Taguchi' design of experiment method and statistical F-test were used to optimize design of tuning forks for use in Qualcomm® mobile station modem (MSM)-3000TM series central processing unit chips of code division multiple access, personal communication system, and global system for mobile communication systems. Design parameters for motional capacitance C1 was further optimized after much trial and error so that as-fabricated resonance frequency values of a tuning fork blank remain almost the same prior to final frequency adjustment plating of the mounted crystals, regardless of packaging pressure change from ambient to designated 6 × 10– 5 Torr. Finite element method modeling was used in conjunction with Taguchi' design of experiment method and, after much trial and error, this enabled selection of essential design parameters to effectively manufacture tuning fork test samples using photolithography with subsequent photoresist spray coating. The tuning fork test samples thus fabricated were characterized and analyzed using Taguchi' method and statistical F-test to evaluate sensitivity of motional capacitance C1 and as-fabricated resonance frequency for various design parameters at ambient to designated packaging pressure of 6 × 10– 5 Torr. Motional capacitance C1 is most sensitive to tine width and tine surface electrode width. Unlike as-fabricated resonance frequency values, motional capacitance values do change with vacuum level of the package. However, tine width asymmetry, electrode thickness and tine tip electrode thickness could be freely adjusted for crystal impedance level without affecting resonance frequency and motional capacitance value. Other design parameters could be adjusted to best suit the manufacturing process.