A 2D imager for X-ray FELs with a 65 nm CMOS readout based on per-pixel signal compression and 10 bit A/D conversion

详细信息    查看全文

• 作者：L. Ratti^a ; ^f ; ^{lodovico.ratti@unipv.it" class="auth_mail" title="E-mail the corresponding author} ; D. Comotti^a ; ^f ; L. Fabris^b ; ^f ; ⁱ ; M. Grassi^a ; ^f ; L. Lodola^a ; ^f ; P. Malcovati^a ; ^f ; M. Manghisoni^b ; ^f ; V. Re^b ; ^f ; G. Traversi^b ; ^f ; C. Vacchi^a ; ^f ; G. Rizzo^d ; ^g ; G. Batignani^d ; ^g ; S. Bettarini^d ; ^g ; G. Casarosa^d ; ^g ; F. Forti^d ; ^g ; M. Giorgi^d ; ^g ; F. Morsani^g ; A. Paladino^d ; ^g ; E. Paoloni^d ; ^g ; L. Pancheri^e ; ^h ; G.-F. Dalla Betta^e ; ^h ; R. Mendicino^e ; ^h ; G. Verzellesi^c ; ^h ; H. Xu^e ; ^h ; M.A. Benkechkache^e ; ^h
• 关键词：Low noise amplifiers ; CMOS technologies ; Non-linear circuits ; Low power ADC ; X-ray imaging
• 刊名：Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment
• 出版年：2016
• 出版时间：21 September 2016
• 年：2016
• 卷：831
• 期：Complete
• 页码：301-308
• 全文大小：1171 K

文摘

A readout channel for applications to X-ray diffraction imaging at free electron lasers has been developed in a 65 nm CMOS technology. The analog front-end circuit can achieve an input dynamic range of 100 dB by leveraging a novel signal compression technique based on the non-linear features of MOS capacitors. Trapezoidal shaping is accomplished through a transconductor and a switched capacitor circuit, performing gated integration and correlated double sampling. A small area, low power 10 bit successive approximation register (SAR) ADC, operated in a time-interleaved fashion, is used for numerical conversion of the amplitude measurement. Operation at 5 MHz of the analog channel including the shaper was demonstrated. Also, the channel was found to be compliant with single 1 keV photon resolution at 1.25 MHz. The ADC provides a signal-to-noise ratio (SNR) of 56 dB, corresponding to an equivalent number of bits (ENOB) of 9 bits, and a differential non linearity DNL<1$\mathrm{DNL}<1$ LSB at a sampling rate slightly larger than 1.8 MHz.