According to teaching and experimental requirements of Optoelectronic information science and Engineering, in order to consolidate theoretical knowledge and improve the students practical ability, the Fourier transform spectrometer ( FTS) experiment, its design, application and improvement are discussed in this paper. The measurement principle and instrument structure of Fourier transform spectrometer are introduced, and the spectrums of several common Laser devices are measured. Based on the analysis of spectrum and test, several possible improvement methods are proposed. It also helps students to understand the application of Fourier transform in physics.
CMOS APS sensors suffer from the SNR reduction caused by the decrease in integration time, especially under the case of low light condition and high relative velocity between objects and sensors. In this paper, a new APS circuits architecture by which TDI function is implemented in the analog voltage domain is proposed to overcome this problem. Time delay integration operation is a straightforward method comes from good old CCD sensors in which signals are added in charge packages with low noise. In this design TDI is mainly achieved by the switch capacitor accumulators and corresponding timing sequences implemented by digital logic on chip. The main noise source especially come from analog sample hold capacitors and OPA are analyzed. The offsets and FPNs are reduced by auto-zero operation and double difference sampling. A 16x128 3T TDI pixel arrays has been taped out by CSMC 0.5 um technology. And finally from the simulation results, a 10dB/dec SNR increase in terms of TDI level has been achieved under low light condition.
Nowadays, CMOS sensors still suffer from the problem of low SNR, especially in the stage of low illumination and high relative scanning velocity. Lots of methods have been develop to overcome this problem. Among these researches, TDI (Time Delay Integration) architecture is a more natural choice, which is natively supported by CCD sensors. In this paper a new kind of proposed current-mode sensor is used to achieve TDI operation in analog domain. The circuit is composed of three main parts. At first, a current-type pixel is proposed, in which the active MOSFET is operated in the triode region to ensure the output current is linearly dependent on the gate voltage and avoid the reduction of threshold voltage in the traditional voltage mode pixels, such as 3T, 4T which use the source followers as its active part. Then a discrete double sampling (DDS) unit, which is operated in the form of currents is used to efficiently reduce the fixed pattern noise (FPN) and make the output is independent of reset voltage of pixels. For accumulation, an improved current mirror adder under controlled of timing circuits is proposed to overcome the problem of saturation suffered in voltage domain. Some main noise sources, especially come from analog sample and holds capacitors and switches is analyzed. Finally, simulation results with CSMC 0.5um technology and Cadence IC show that the proposed method is reasonable and efficient to improve the SNR.