Designing for Stability in High Frequency Circuits

Application Notes

This application note will help engineers understand how instabilities fundamentally arise in their circuits and illustrate how to troubleshoot and resolve these issues upfront in the design process before manufacturing. This not only requires an understanding of classic theory and techniques but also practical knowledge to apply these efficiently using modern design tools such as a novel new impedance probe called the WS-Probe.

Abstract

Why should high-frequency circuit designers consider stability early on in the design process? Isn’t there enough to worry about just making the circuit function at the fundamental frequency? In the past, Microwave Engineers used to solve stability problems in the lab, perhaps adding bypassing or loss in a strategic location to stabilize their circuits. Stability was viewed as too complicated to model or predict, and the problems were usually easy enough to solve in the lab anyway. But things are changing. Across the entire wireless communications industry, standards are moving higher in frequency and systems are getting more complex. Instability arises from a combination of gain and feedback. In today’s circuits, the gain is higher due to increasing device fT’s, and feedback is more prevalent because features are more compact and resonate more easily with signals that have smaller wavelengths. At the same time, advanced packaging technologies make the internals of the circuit less accessible than in the past, meaning things are harder to fix after the fact in the lab, even with the most apt technicians. To make circuits that meet the needs of modern communications systems, designers need to master stability by truly understanding the root causes of problems in the circuit before building a design. The problem: stability is a very complex topic. Most high-frequency design engineers use only the classic “Rollett stability factor” to assess circuits, but this technique is based on assumptions that may not be valid for modern circuits. Besides, K-factor only applies to a two-port network at the external I/O’s, while the circuit inside could be very complex and hinder visibility from the outside. There are many alternative techniques in the literature, but they are sometimes difficult to apply correctly, and furthermore, it’s not clear which one is best for any given application. This Application Note will help designers understand how instabilities fundamentally arise in their circuit and illustrate how to troubleshoot and resolve these issues upfront in the design process before manufacturing. This requires not only an understanding of theory and classic techniques but also a practical knowledge of how to apply these techniques efficiently using modern design tools. This paper starts by reviewing the theory, discussing concepts like loop gain, return difference, and driving point impedance, and then expands to build a framework for applying these techniques to modern circuit design. The key is to use a new probe, called the WS-Probe, which has recently become available in Keysight’s PathWave Advanced Design System (ADS), to derive the necessary stability measures quickly and efficiently. The probe allows the application of multiple stability analysis techniques to the circuit post-simulation for both small and large signal analysis in a non-invasive manner. Multiple examples illustrate how to use the probe on real-life circuits. After reading this Application Note, you’ll look at stability in an entirely different way and the circuits you design will reflect that.

Table of Contents

  • Stability in modern wireless design: the perfect storm
  • The Trouble with “K-Factor”  
  • Feedback Theory
  • Techniques to compute loop gain  
  • Rigorous Stability Analysis Techniques 
  • Modern Implementations
  • Active-Passive Circuit Bifurcation
  • Summary of Techniques
  • A Modern Approach to Stability Analysis
  • Unifying Stability Analysis Techniques: The WS-Probe 
  • The WS-Probe in action
  • Using WS-Probes to Compute a “Virtual Load Pull” 
  • Stability under Large Signal Drive
  • Extending WS-Probe Analysis to Large Signal
  • Practical Application Examples
  • Example 1: Using the Probes on a Transistor 
  • Example 2: Using WS-Probes in a Multistage MMIC Amplifier
  • Example 3: Using EM Simulation to Visualize Instability 
  • Summary 
  • References

Stability in modern wireless design: the perfect storm

Wireless technology is advancing faster than ever before. Whether you’re working in Aerospace / Defense / Satcom, Commercial Wireless, or Automotive, the trends that are driving these businesses forward are twofold. First, frequencies are increasing. Think about your car: a few years back, the highest frequency hardware might have been the FM radio. Now, most cars have some form of millimeter-wave radar or sensors built-in. Consider commercial cellular; frequencies increased very modestly between each new standard upgrade, about 1.4X over 17 years. Now, going from 4G to 5G, frequency increases by a factor of 38 – in just 12 years!

Second, wireless communication systems are becoming very complex. For example, Aerospace/ Defense hardware designers cannot just build high-performance, single functioning systems anymore. The systems also need to be flexible and reconfigurable, which leads to dense heterogeneous integration. These heterogeneous designs are orders of magnitude more complicated than their traditional chip and board predecessors. At the same time, 5G cellular and automotive communications increasingly require MIMO (Multi-In-Multi-Out) configurations, which imply the need for beam steering phased arrays. These advanced systems used to be developed exclusively by defense companies – now they are being produced commercially at high volumes. 

Both trends lead to many hardware design challenges; new circuit topologies are needed, antennas must be co-designed, digital and wireless blocks must co-exist in smaller spaces – but one unexpected nemesis could throw a wrench into all of this: amplifier stability is becoming a major problem with high frequency, tightly integrated circuits, and systems. At first glance, this may seem a little surprising – stability has always been a circuit design challenge, but one that was manageable with best practices, basic simulation, and a little handiwork in the lab. Why is stability poised to become such an issue now?