As of July 1, 2021, Google will discontinue the RSS-to-email subscriptions service.
Hence, the email alerts will be impacted while we explore other options. Please stay tuned for further communication from us.
Cadence offers multiple electromagnetic (EM) extraction technologies to model the parasitic effects of interconnect and passive component geometries (see Virtuoso Meets Maxwell: Layered Electromagnetic Modeling For Sufficient Accuracy for details).
One of these EM extraction technologies is Quantus RLCK extraction, which IC designers at Rohde & Schwarz have been using with great success for more than a decade now.
Rohde & Schwarz is a global company headquartered in Munich, Germany, with more than 12,000 employees in over 70 countries. They are among the technology and market leaders in wireless communications, RF test and measurement as well as other related fields. Because ICs with the required performance for their products are often not available from semiconductor vendors, they have their own department for developing high-performance high-frequency ICs like mixers, variable-gain amplifiers, analog/digital converters etc.
Many of you are probably familiar with parasitic RC extraction, which has become essential for designs in all current semiconductor processes. Tools like Quantus calculate the parasitic resistances and capacitances of all interconnects and add them to the circuit netlist in order to represent the circuit performance in a more realistic way.
But Quantus cannot only do RC extraction. It can also extract parasitic inductance (L) and mutual inductance (K), which represents the magnetic coupling between any pair of parasitic inductors. With such a full RLCK extraction, high-frequency electro-magnetic effects can often be modeled with pretty good accuracy. In order to add even more precision, an option to also do a full RLCK extraction of "open" connections (stubs) has recently been added to Quantus.
When Rohde & Schwarz started using Quantus more than a decade ago, this was simply the only way to simulate the electro-magnetic effects in larger circuit layouts in an efficient way. The simulated and measured performances usually matched pretty well, the results were generally well within the range expected due to the variability of the active and passive devices in the circuit.
But Quantus also has other advantages. Electro-magnetic simulators like Clarity, AXIEM, and EMX generally produce S-parameter data, which represent the properties of the interconnect network in the frequency domain. Using these data in a time-domain simulation is not without challenges, especially if the network has a large number of ports. The network is also a black box, so that only voltages and currents at its ports are accessible for probing.
With Quantus, the parasitic devices causing the electro-magnetic effects are added directly to the circuit netlist. This makes it possible to probe voltages and currents inside the interconnect networks. With this approach, the IC designers at Rohde & Schwarz have been able to examine parasitic oscillations for example in bias networks, which may occur when long connecting lines act as unintended resonators.
To be fair, it should be mentioned that fully RLCK extracted netlists present their own set of challenges to a circuit simulator. For a normal circuit, the circuit matrix is usually sparse, which means that it only has a low number of non-zero elements. Circuit simulators are optimized to take advantage of this fact. For an RLCK extracted netlist, however, the matrix is no longer sparse, because there is a mutual inductance between any pair of parasitic inductors, which corresponds to a very large number of elements. This may lead to slow simulations or even to convergence problems.
The straightforward approach of setting all mutual inductors below a certain threshold to zero does not work very well, because it may numerically turn the interconnect network into an active circuit, leading to convergence problems and runaway signals. SpectreRF R&D has been working closely with Rohde & Schwarz to address these challenges.
We thank Frank Wiedmann from Rohde & Schwarz for contributing to this blog.