Advertisement

How much automation are current electronic design automation (EDA) tools providing? While the RF computer aided engineering (CAE) industry has a lot to offer in the way of automated circuit and system analysis, what about automated circuit synthesis? The circuit design project that starts with one of the typical simulation programs on the market today probably isn’t benefited by design automation as much as it could be. The reason for this is that many of these software programs are simulation focused rather than design focused. In other words, these programs can do a good job of analyzing or even optimizing an existing circuit design, but do little or nothing to aid the engineer in coming up with the circuit topology in the first place. The result is that the design engineer is forced into a kind of backward process of trial and error in developing an initial circuit design or prototype. Moreover, when only simulation and optimization are employed, the design process can fail completely.

Some simulation programs attempt to assist the designer by providing example schematics of common circuits. However, there is no guarantee that the circuit topology provided will satisfy a new set of specifications, even after employing optimization. Consider, for example, the output matching network for an RF amplifier. One of the most economical circuits for narrow-band matching is the two-element L network. There are eight different ways an inductor or capacitor can be combined to form an L network. At least one of these L networks is capable of matching any (real or complex) source to any other (real or complex) load. While there may be several L configurations that will provide a match, not all eight will work for a given source and load impedance. Therefore, if one of these L configurations is borrowed from an existing design and applied to a new device or load it is likely to fail to provide an impedance match, even after lengthy attempts to use an optimizer. The designer may not realize the futility of the attempt until a great deal of time has been expended.

The optimizer in the simulation program fails because it cannot know beforehand which topology will solve the problem. In contrast, a circuit synthesis program can select the correct topology and provide the exact component values.

Another situation in which the optimizer is sure to fail is when the goal is to provide a conjugate match to both ports of a potentially unstable device. A simultaneous bilateral conjugate match is not possible unless the potentially unstable device is first rendered unconditionally stable at the design (match) frequency. Resistive loading at the input or output terminals of the device or various forms of feedback can usually stabilize the device. However, neither the simulator nor the optimizer is equipped to alter the circuit topology appropriately.