Advanced Search >>

« March 2009 | Main | May 2009 »

1 posts from April 2009

04/14/2009

Future Of RF Large-signal Measurement Systems; Should We Continue To Expect So Little?

April 2009

Johannes Benedikt is CTO at Mesuro Ltd, leading the commercial introduction of new measurement solutions that enable systematic waveform engineering at RF and microwave frequencies. He took up the position, which evolved from his work at Cardiff University, UK, in October 2008. Benedikt received a Degree in Electrical Engineering from the University of Ulm, Germany in 1997. While studying for a PhD Degree at Cardiff University, which he received in 2002, he held an additional position as a senior research associate supervising a research programme with Nokia on RF power amplifiers. In December 2003 he was appointed a lecturer at Cardiff University with responsibility for expanding research on large-signal measurement systems and their application in RF PA design. He has been involved in the development of a number of time-domain characterisation and active harmonic load pull systems.

To comment or ask Johannes Benedikt a question, use the comment link at the bottom of the entry. The first 5 people to comment will receive a copy of the Electrical Engineering Handbook (please include your e-mail and mailing address).

At university we are taught how semiconductor devices and amplifiers respond to applied current and voltage waveforms. We learn how the input voltage generates a current at the output, which in turn varies with an increasing output voltage. The resulting IV output plots - constrained by the device IV plane - can then be directly used to identify an optimum load that will bring maximum power out of the device. Yet, as soon as we emerge into the practical world of RF and high-frequency engineering these fundamentals are quickly jettisoned in favour of other concepts such as small-signal theory and measurements.

Admittedly, small signal parameters contain lots of information enabling amplifier designers to maximise and eventually trade-off key performance figures such as output power, gain, and bandwidth for a given device. The same design principles are applicable even when the device is in a large-signal state and is generating a significant harmonic content. Here, I mean the harmonic content that is directly produced by the current generator as device parasitics and packaging often form an excellent filter for harmonic frequencies.

So, what is the problem? It is the fact that although when performing small-signal measurements a number of key parameters such as output power, efficiency or reliability can be determined, we are still left in the dark when we try to explain the results. This is where PA designers become frustrated because they need to identify the underlying problems and be able to take corrective measures.

Without this understanding we are forced to tweak design parameters, such as fundamental load and bias, and observe their impact. If we do it often enough with a particular device and amplifier design the resulting experience will tell us straight away what should be changed. This design methodology works well until we are asked to use a new device technology or develop a new type of amplifier; maybe a Doherty or Envelope Tracking PA. In this case, a large portion of the existing experience becomes irrelevant, making the development of new designs time consuming.

The introduction of new device technologies and amplifier architectures often means that we need to start taking into account harmonic load and in some cases even harmonic source impedances. This introduces even more parameters that have to be tweaked during the PA design phase, leaving us with a multi-dimensional parameter space with an almost infinite number of combinations.

Measured RF I & V waveforms can offer additional insight and understanding that goes far beyond the small-signal world. Not only can we derive all the small-signal parameters, but often a single glance at the I & V waveforms allows us to explain the observed performance. An early clipping of the current waveform at low or high voltage values will pinpoint clearly any knee-walk-out or soft pinch-off issues within a GaN technology. In cases when the current waveform starts rising at voltage values that approach the voltage breakdown region we get a clear indication that this mode of operation will impact on the reliability of the device.

So how did we end up in a position so distant from the fundamental concept of current and voltage waveforms?

At first glance, this question appears to relate directly to available microwave receiver technology, as the measurement of waveforms requires the capture and calibration of broadband multi-octave signals. In fact, to capture a genuine waveform we are required to detect all relevant spectral components ranging from DC, baseband and fundamental to the highest harmonic frequency of interest. Appropriate receivers such as nonlinear VNAs and calibration techniques have recently been introduced, which indicate awareness within the established providers of RF measurement systems of the need to go beyond small-signal parameters.

However, the underlying measurement technology needs to go further than just the measurement of I & V waveforms. If we place a device into a measurement system and measure the existing waveforms, we will obtain new information that will allow us to explain in detail why the device has low output power or efficiency but it will give us very limited options as to how we can improve the performance. The main reason for this is the 50 O characteristic impedance of measurement systems. It works perfectly for small-signal measurements but unfortunately it is a rare event when 50 O constitutes an optimum loading for a large-signal device.

So let’s face it, if we say ‘RF I & V waveforms’ we don’t mean only their measurement but the creation of a very specific set of waveforms that will result in a mode of operation for a device that will give us an improved performance with examples including class-B, C, E, F, F-1, or even Doherty PAs. This is actually the Achilles heel of all present measurement solutions, as specific current and voltage waveforms necessitate an accurate impedance control at all spectral components that are present at the device; to be more precise we need to control the impedances inside the device at its output current generator.

Consequently, to make I & V waveforms a viable proposition to PA designers, measurement systems have to evolve beyond the provision of calibrated waveform data. What we actually need is an integrated solution that is capable of capturing waveforms and controlling the impedance environment to engineer the specific waveforms that are required by the various PA modes.

The required impedances need to be generated inside the device, thus any impedance control will have to overcome the signal losses that are introduced by the device parasitics and packaging. Also, the impedance control has to be established for all spectral components at the fundamental and harmonic frequencies of interest. For modulated signals the impedance control will need to include the specific tones that are contained within the signal.

So what are the potential benefits of such new systems? At Cardiff University we’ve been developing such integrated solutions in the last ten years [1-3]. As these developments matured over the years our research effort increasingly focused on their application for PA design. The results so far are very encouraging as new and relatively inexperienced students were repeatedly producing PA prototypes with very attractive performances that were achieved within their first design iteration [4], which in a way also confirms the soundness of this new approach.

If some of you were wondering whether such new systems will make it easy to design PAs, I would like to categorically say that this is not the case, as the engineering of waveforms is not a simple matter of introducing linear changes to the harmonic loads. An example of this might be the I & V waveforms of a class-F amplifier that assumes an open load at odd harmonics and thus does not allow for any current flow.

But what happens if the device tries to produce a signal at the third harmonic? Is the energy from the third harmonic transferred to the fundamental RF signal or will it go a completely different way? Another discussion point is how we can remove the effects of device parasitics and packaging from the measured waveforms to obtain more meaningful waveforms? This is actually quite important as all PA modes of operation are defined by waveforms at the intrinsic current generator at the device output. And eventually we need to consider how accurately we need to do this de-embedding to obtain useful waveforms? The bottom line is that the capability provided to measure and engineer waveforms is only as good as its users, as it is really up to them how the new information is utilised during the design. This is actually very much the same as with other instruments.

However, what the new integrated systems offer is a bridge between the fundamental nonlinear circuit theory - that is based on current and voltage waveforms - and practical PA circuit design. The new systems give us the opportunity to systematically identify underlying technology and design issues that prevent us from reaching the theoretically possible performance and tackle them without getting lost in the endless tweaking opportunities within a multi-dimensional parameter space and thus shortening the PA development cycle while ensuring a theoretically optimum design.

COMMERCIAL AVAILABILITY
As a direct response to the existing lack of integrated measurement solutions Mesuro Ltd has been formed with the aim to develop systems that make ‘I & V waveform engineering’ commercially accessible at low and high power levels. The planned market introduction of these new measurement systems is June 2009.

These systems are based on the longstanding developments at Cardiff University and offer a number of unique advantages. To measure genuine waveforms the utilised receiver technology is based on well-known microwave sampling oscilloscopes as they enable the accurate measurement of all spectral components ranging from DC to several tens of GHz. The required control and implementation of suitable sampling algorithms for the measurement of waveforms has been developed within the research group including necessary calibration techniques.

However, the development of suitable load and source-pull technology had to be based on a new approach, requiring the deployment of active load pull technology - maybe also some sort of combination of active and passive systems. More specifically, we need an open-loop load pull architecture, as any active systems that are based on closed feedback loops are narrowband and have the potential to oscillate.

Combining such an open-loop system with newly available arbitrary waveform generators creates a particularly elegant solution that realises simultaneous input drive for the device under test, active harmonic source-pull and load-pull at any frequency that can be produced by the generator. Figure 1 is a simplified block diagram of the new integrated measurement system.

Fig1Apr

Figure 1. Active Harmonic Source-pull and Load-pull Test System

The integration of both the microwave receiver technology and the active load system is more involved than simple ‘plug and play’ and needs to be carefully considered from a hardware and software point of view, necessitating a careful co-development of all system components. A disadvantage is that this approach is rather difficult to realise with multiple vendors providing separately developed building blocks for such a system.

An example of this is the integration of a measurement system with a load pull system, as any excessive perturbations from that 50 O characteristic impedance have to be avoided over the entire frequency range as they potentially trigger unstable device behaviour. However, such impedance perturbations are easily introduced when using conventional multiplexers to establish separation between the harmonic frequencies for either their independent control or for amplification within active load pull systems.

As for the software, it should empower the user to quickly start with waveform engineering without the need to become a master in the use and set-up of the utilised instrumentation such as sampling scopes or signal sources.

CONCLUSION
The strong demand for power efficient wireless systems and the continued challenges that PA designers are facing in realising them has created unique market openness for new solutions which in turn creates an opportunity for the establishment of new RF test and measurement solutions that are more appropriate for RF PA design.

References
[1] Z. Aboush, J. Lees, J. Benedikt and P.J. Tasker, "Active Harmonic Load-Pull System for Characterization of Highly Mismatched High Power Transistors,” in 2005 IEEE MTT-S International Microwave Symposium Digest, Long Beach, CA, USA, June 12-17, 2005, pp. 1311 - 1314.
[2] A. Alghanim , J. Lees, T. Williams, J. Benedikt, and P. J. Tasker, "Reduction of Electrical Baseband Memory Effect in High-Power LDMOS Devices using Optimum Termination for IMD3 and IMD5 using Active Load-Pull,” in 2008 IEEE MTT-S International Microwave Symposium Digest, Atlanta, Georgia, USA, June 15-20, 2008.
[3] C. Roff, J. Benedikt, P.J. Tasker, D.J. Wallis, K.P. Hilton, J.O. Maclean, D.G. Hayes, M.J. Uren & T. Martin, "Utilization of Waveform Measurements for Degradation Analysis of AlGaN/GaN HFETs,” 70th ARFTG Microwave Measurement Symposium, Tempe, Arizona, USA, Nov 2007.
[4] P. Wright, A. Sheikh, Ch. Roff, P. J. Tasker and J. Benedikt, "Highly Efficient Operation Modes in GaN Power Transistors Delivering Upwards of 81% Efficiency and 12W Output Power,” in 2008 IEEE MTT-S International Microwave Symposium Digest, Atlanta, Georgia, USA, June 15-20, 2008.

Main
My Photo

Categories

Blog powered by TypePad
Member since 11/2008

Other Horizon House Sites:

Microwave Journal Online: Home | Current Issue | News | Buyer's Guide | Events | Resources | Archives | Subscriptions | Privacy Policy

Advertiser Information:
2009 Media Planner

Find out why more companies advertise in Microwave Journal than any other publication in the industry.

Read More >>

Microwave Journal
Editorial Information

Editorial Planning Guide and Information for Authors

Read More >>


©2009 Microwave Journal & Horizon House Publications ® All rights reserved.