A Method for Multi-parameter Robust Stability Distribution Analysis of Analog Circuits

Abstract

The invention belongs to the field of circuit analysis and relates to a method for analyzing distribution of robust stability of multiple parameters of an analog circuit. The method for analyzing distribution of robust stability of the complete analog circuit in multi-parameter space is combined with an analog circuit symbol emulation technique, stability judgment methods such as Routh Table in control theories, an interval calculation method, a high dimension segmentation technique and an importance sampling monte carlo method. According to the method, defects of prior art are overcome, a stability conclusion obtained by using the interval calculation method is a definitive conclusion; by means of a parameter axis correlation preferential segmentation technique, segmentation efficiency of high-dimensional parameter space is effectively improved; and stability percentage of a circuit parameter vector collection of which stability can not be judged is estimated by using the importance sampling monte carlo method based on volumes. The method is capable of solving the problem of stability of the analog circuit in the parameter space and the statistical problem of the stability percentage.

Description

technical field [0001] The invention belongs to the field of circuit analysis methods, in particular to a method for analyzing multi-parameter robust stability distribution of analog circuits. Background technique [0002] In the integrated circuit manufacturing process disclosed in the prior art, the relative error between the geometric and electrical parameters of the device obtained by tape-out and the design nominal value is not large, but as the integrated circuit semiconductor manufacturing technology moves to 65nm / 45nm / 32nm process With the development of nodes, complex processes such as sub-wavelength lithography under nanotechnology, Damascus copper interconnection and chemical mechanical polishing (CMP) are widely used, which makes the deviation between the actual silicon wafer pattern and the design layout increasingly serious[1, 2]. The interference and diffraction effects of light waves in the photolithography process cause severe distortion of the layout on th...

AI Technical Summary

Problems solved by technology

Therefore, for characteristic polynomials larger than the fourth order, the analytical expression of stability cannot be obtained from the transfer function, that is, it is difficult to apply symbolic methods to the stability analysis of high-order circuits

Second, even if the symbolic expression of the stability is obtained, it is not easy to count its stable percentage in the high-dimensional parameter space

Unfortunately, since this method does not introduce a space segmentation strategy, it cannot be directly applied to the stability distribution analysis

Ratschek proposed segmentation strategies such as slice, sub-slice, and bisection to solve a feasible solution of a 9-dimensional circuit design problem [20], but this method cannot be directly applied to solve the stability distribution problem of the entire parameter variation space

[0010] In summary, for the problem of robust stability distribution analysis in the parameter space of analog circuits, the existing cybernetic methods cannot solve the problem of nonlinear correlation between parameters; the Monte Carlo method has a large amount of calculation and is non-deterministic The method cannot judge the stability of the region, and can only solve low-dimensional problems; the symbolic simulation and sensitivity analysis methods can only deal with the stability of low-order circuits; there is no directly available method for interval calculation

Images