Resistor Geometry Comparison With Respect to Current Noise and TrimSensitivity

Abstract

The paper deals with laser trimming of film resistors. The finite-difference solution of the electric field calculated by a digital computer gives the total resistance and the inhomogeneous current distribution within the resistor area. The current noise voltage was found to be proportional to the sum of reciprocal subareas wherein the current density is approximately constant. Resistor trimming is modelled by finite increments of cut length to obtain the resistance trim rate and the trim sensitivity.Known and new resistor geometries and cut configurations are compared (based on the same resistor area and the same initial tolerance of ± 20%) to find the optimal geometry depending on the ratio of the nominal resistance value to the sheet resistivity. Diagrams are given leading to practical design rules not only for resistance trimming but also for deterministic and functional tuning.

Noise spectroscopy of tunable nanoconstrictions: molecule-free and molecule-modified

“…For simplicity of resistance and noise calculations we will consider our system as a 2D resistor. For the 2D case the power dissipation can be calculated using following equations: 31 The precise calculation of the integrals in 2D is often done by solving finite-difference equations 32,33 or by conformal mapping. 34,35 An analytical solution based on approximated equipotentials has an advantage, especially for the relative noise SR / R 2 , because errors in SR and R are partially cancelled out.…”

“…It turns out that C us /R sh ≈5×10 −21 cm 2 /Ω for uniform gold layers and poly-silicon of different thickness [31]. The precise calculation of the integrals in 2D is often done by solving finitedifference equations [32,33] or by conformal mapping [34,35]. An analytical solution based on approximated equipotentials has an advantage, especially for the relative noise S R /R 2 , because errors in S R and R are partially canceled out [26].…”

Comment on ‘Comparison of excess 1/fnoise spectra in trimmed and untrimmed thick film resistors’

“…If the trimming process does not damage the material, there is no reason for an increase in ¬. Moreover, Ulbrich (1977) has already reported that noise increase in laser-trimmed resistors can be attributed to the more inhomogeneous current distribution. Figure 4 of Peled et al (2000) shows a surprising result where a reduction in the volumetric percentage of bismuth ruthenate grains leads to an obvious increase in resistance but surprisingly enough a decrease in relative noise C. This can be explained in the framework of multi-spot contact behaviour.…”