Simon Kristiansson

Substrate resistance modeling for noise coupling analysis

Accurate substrate modeling is of utmost importance for substrate noise coupling analysis in mixed-signal circuits. In this paper we present a two-port Z-parameter model based on a physical description of the substrate surface potential. The Z-parameter model is expressed using a one-port semiempirical resistance model. This resistance model accurately describes the observed initial increase in resistance followed by the observed saturation as the contact separation increases. The Z-parameter model was compared to measurement data obtained from a set of CMOS test structures. The model fits measured results well, in contrary to when resistive networks are used to represent the substrate. Furthermore, we show that the substrate coupling between a digital circuit and an analog circuit does not have to become zero as the distance between the circuit blocks increases, instead the coupling between the circuits approaches a constant non-zero value.

A surface potential model for predicting substrate noise coupling in integrated circuits

Caution must be taken when designing circuits so that noise injected to and transmitted through the substrate does not reach and degrade the performance of sensitive circuitry present on the chip. In this paper we present a simple analytic substrate model for evaluating substrate noise coupling. The model can handle an arbitrary number of aggressor and victim devices on a multi-layered substrate with either biased or floating backside. The model has been validated by finite element calculations and measurements on test structures manufactured in a 0.35 /spl mu/m CMOS process, and it is shown that the model gives an accurate description of the substrate noise coupling. For example, the noise suppressing properties of guard rings have been evaluated.

Compact Spreading Resistance Model for Rectangular Contacts on Uniform and Epitaxial Substrates

We present a compact analytical spreading resistance model for substrate noise coupling analysis. The model can handle rectangular contacts on uniform substrates of finite thickness with a grounded backplane. In contrast to previously published compact models, the model does not require extraction of fitting parameters. The model is also scalable with the resistivity and thickness of the substrate, and with the contact size. The model is verified with extensive finite-element calculations, and the accuracy is shown to be good. We also show that the model can predict the spreading resistance on epitaxial substrates.

Properties and modeling of ground structures for reducing substrate noise coupling in ICs

Substrate noise can seriously degrade the performance of system-on-chip designs containing sensitive analog circuits together with noisy digital circuits. To reduce the coupling guard bands or guard rings are often used for protecting the analog circuitry. However, the efficiency of these ground structures is often not known on beforehand. Therefore, in this paper we investigate the noise reducing properties of different ground structures. It is shown that distributed ground structures are typically more efficient than localized ground structures due to smaller inter-contact influences. We also introduce two convenient measures of grounding efficiency, the global and local substrate potentials, and we present compact analytical models for these. The models are verified with numerical calculations and the accuracy is shown to be very good

Modeling of rectangular contacts for noise coupling analysis in homogeneous substrates

A major concern when designing mixed-signal integrated circuits today, is to avoid unwanted signal interaction between digital and analog circuit blocks present on the same silicon substrate. Part of this interference is conducted through the substrate. To minimize this substrate noise coupling, simple models are needed which can predict the noise and guide the IC designer in the layout of the mixed-signal chip. In this paper we present simple models for the resistance and surface potential of rectangular contacts based on approximating the rectangular contacts with elliptic contacts. The accuracy of the approximate model is validated with finite element and boundary element calculations.

Resistance Modelling in 1D, 2D, and 3D for Substrate Networks

Normally, circuit model resistor networks are used to describe the substrate when noise coupling analysis of mixedsignal circuits is performed. In commercially available tools the substrate is discretized into a large number of resistors for detailed postlayout analysis. However, circuit designers need to perform quick calculations for rough estimates early in the design process, even before layout, and must therefore catch the behavior of the substrate in a lumped model with only a few resistors. For correct prediction of substrate noise coupling these resistors must be accurately modelled and their geometry dependence understood. In this work a comparative study between one-, two-, and three-dimensional substrate models is presented and it is shown that the resistive properties of the substrate can only be correctly captured using a three-dimensional analysis. One- and two-dimensional models presented in the literature are simply not good enough since they, for instance, do not capture the saturation effects found in threedimensional models.

Evaluation of active cancellation of substrate noise in mixed-signal ICs

In mixed-signal integrated circuits interference between noisy and sensitive circuits is a serious problem. Active cancellation of this substrate coupled noise was proposed about a decade ago and has been claimed to be more efficient than passive decoupling (DC/capacitive grounding). Most of the previously presented papers have investigated active decoupling methods using simple, less accurate, resistive network models as the common substrate. In this paper two common active decoupling circuits (feedforward and feedback) have been investigated using an accurate substrate model based on z-parameters. Three different configurations of contacts have been investigated: 1. all the contacts are close to each other; 2. just guard bands are close to each other; and 3. when guard rings are implemented. Conclusions have been reached with the aid of theoretical and simulation based approaches. When distances between contacts are small, the level of suppressed noise highly depends on the type of active decoupling circuit. In this case, if backplane is floating there is an optimum op-amp gain which should be used in order to gain the minimum coupled noise level. Conversely, when distances are large, none of the mentioned effects have been noticed. When guard rings are in use instead of guard bands, regardless of the fact that overall noise level is better, simple DC grounding is preferred.

Noise Interaction Between Power Distribution Grids and Substrate

We have investigated the interaction between power delivery and substrate coupling in terms of noise. From our results, we identify that an increased density of substrate contacts does not to any significance decrease noise on the power supply lines. However, the current injected into the substrate is highly dependent on higher-level grid/package inductance and substrate contact density. We have derived statistically that substrate noise variations could be related to these two design parameters to 69.75%. Based on linear fitting, a model that describes the injected current as function of substrate contact density and power delivery inductance is developed.

Evaluation of using active circuitry for substrate noise suppression

The performance of system-on-chips can be severely degraded if noisy circuits interfere with sensitive circuits through the common silicon substrate. Many methods have been proposed to suppress such substrate noise, ranging from designing circuits that generateless noise to using guard bands to prevent noise from reaching the sensitive parts. In this paper we have investigated a proposed method for substrate noise suppression using active circuitry, and compared it with two passive methods; resistive grounding and capacitive decoupling. Published results indicate that active noise suppression outperforms passive noise suppression. However, these results are based on simulations using over simplified substrate models, like single node or resistor T-models. In this paper we confirm these previous results, but we also show that when using a more realistic substrate model these advantages disappear! We claim that there is no noticeable difference between active noise suppression and DC grounding; these two methods are comparable and better than capacitive decoupling. The substrate resistivity in system-on-chip solutions typically result in large substrate resistances that decrease the efficiency of active decoupling compared to simple DC grounding.

A High-Frequency Extension of a Surface-PotentialBased Substrate Model for Noise Coupling Analysis

In this paper we present a high-frequency extension of our surface-potential-based substrate model for predicting substrate noise coupling in integrated circuits. The model handles an arbitrary number of aggressor and victim devices on a multi-layered substrate with either biased or floating backside. We show that the dielectric properties of the substrate are easily included in this model for providing a more accurate description above GHz frequencies. Finite element calculations are used for validating the model