To Whom it May Concern: The Parkervision PA Deception

To Whom it May Concern:  The Parkervision PA Deception.

 

 

We do not often find the need to use the word “disgraceful” when reviewing technical material, however the recent deluge of  bogus patents and outrageously flawed claims by Parkervision on their alleged revolutionary RF PA technology tempts us to use even stronger language. There is an irony. The technical core of their claims is a hardware configuration which may well not have been patented or used commercially before, but for a very good reason. It will not deliver any significant improvement in RF PA efficiency and/or linearity, in comparison to currently available commercial solutions.  

 

Our task is to try to draw the attention of potential investors and customers to our own views on their claims, (i.e.) that PV have a “disruptive” revolutionary approach to the implementation of power amplifiers in RF communications systems. Clearly, our view is that they don’t, and their latest attempts to demonstrate hardware are nothing more than a continuing display of technical ineptitude and naiveté in a well-mapped technical territory.

 

We are not patent lawyers, nor can we even claim to have much experience in writing patents, but reading the voluminous patents that have emerged from PV in recent months makes us quite angry. We perceive a deliberate and pre-meditated effort to “buck the system”, rather than a genuine effort to protect useful inventions. Their highly repetitious and duplicative style, and their technique of citing hundreds of references which are then not mentioned in the body of the patent, raises my suspicions about how genuine the applicants really are. we would like to think that such tactics are well known by the Patent Office inspectors, but we have to believe that the named inspectors have not had the time to review each any every cited reference. It seems to me that the basic core concepts in these patents could have been  described with a couple of figures and a few paragraphs of text, and the rest of the many hundreds of pages are nothing more than the embellishment of the core concepts with techniques that are already well known and widely practiced in the industry, such as bias control and digital predistortion.  We will here attempt to describe, in non-mathematical terms, why this core concept is flawed inasmuch as it will not enable any significant improvement in the current industry capability to implement highly efficient and linear RF power amplifiers.  

 

Some background first. A typical radio communications signal consists of a radio-frequency “carrier” which is “modulated” by varying its instantaneous amplitude and timing (or “phasing”). The modulation is what contains the information, or “intelligence” being transmitted. Such signals can be generated, along with all of the necessary coding and system management content, very easily using modern digital integrated circuits. Unfortunately, the power level at which these signals can be readily generated is too low for a typical mobile communications system, where the mobile transmitter (e.g. a cellular mobile phone) has to generate a signal that is strong enough to be detectable by a “base station” several miles (or even 20-30 miles in rural areas) away.  So it is necessary to take the signal and amplify it (usually to a level of 1 Watt or thereabouts for a mobile) using a power amplifier (“PA”). This process consumes a large amount of power. Even if the PA were 100% efficient, there would still be a drain of 1 Watt from the battery, which would run it down in a matter of an hour or two. There is another problem, which is that the PA will to some extent distort the signal, which can cause bad reception at the other end of the link, and also breach on-air regulations. Unfortunately, it is a matter of well established fact that there is a trade-off between efficiency and distortion in PA design. PAs which do not meet the regulatory specifications can have quite high efficiency, and in assessing claims for improved PA design techniques and technology it is vital that efficiency numbers are quoted at comparable (and acceptable) distortion levels. It is not difficult to make a PA with much higher efficiency than that typically quoted in current RF PA data sheets, if the distortion specs are ignored.

 

This linearity/efficiency trade-off is such a major issue in radio communications that sometimes system designers bypass it by using “constant envelope” modulation. This eliminates the amplitude variation as a means of conveying information, which is then carried entirely in the phase of the transmitted carrier. (The European GSM system is an example of this). The problem with amplitude modulation (AM) is that at times the level of the signal is very low, so that at these times the PA is consuming much more power than is really necessary. An analogy would be the ownership of a double-decker bus as one’s sole means of personal transportation. This would be just great for transporting 100 wedding guests on the very infrequent occasion that one of your relatives gets married, but highly wasteful of gasoline and parking space when using it every day to drive to work. A better solution would be to have another smaller vehicle to use for short personal trips. In the case of an RFPA, it is not so easy to “switch vehicles” during the time the signal transitions from high to low levels, due to the fact that the “switch” has to be enacted in a matter of microseconds, which even in electronic terms is quite fast. But over a the last seventy years or so, many schemes have been proposed for RF PAs which in effect do just this; they switch between a high power device and a lower power device at appropriate times in the signal modulation cycle.  The Doherty PA [1] is an example of this, dating from the 1930’s and currently being widely used in cellular basestation transmitters.

 

The core of the PV patents and claims uses another well-known and widely researched technique for bypassing the linearity-efficiency tradeoff in systems which use AM. The basic idea is to note that constant-amplitude, or phase modulated (PM) signals are much less distorted by an RF PA and this enables the PA to run at much higher efficiency (again, this is demonstrated in the GSM system). This technique has been the subject of many papers over several decades, and is generically termed “Outphasing”. In a classical and widely cited paper by Chireix [2], dating again from the 1930’s, two PAs are run with constant amplitude, phase modulated signals, and their outputs combined. The two PAs thus run very efficiently, and the AM signal can be generated at the summed output by suitable differential phasing of the two input signals. Even in the 1930’s, Chireix’s technique was recognised as a major breakthrough. The idea that AM can be created by combining two constant amplitude, phase modulated sinusoidal carriers is about as old as sinusoidal function theory, which takes us back a few centuries. The key breakthrough in the Chireix system was the use of a special kind of power combiner.

 

For several decades, it has been widely recognised in the RF PA industry that simple addition of  two amplified constant amplitude signals will not result in improved efficiency. The way in which PV flout ignorance of this well-known fact is the most striking aspect of all their technical output. 

 

 

PV appear to promote, and claim as revolutionary, what in mainstream PA circles has been termed the “bonehead” approach, where a simple power combiner, or parallel output connection, of the two amplifying devices is used to sum the two output currents. The problem with this simplistic approach is that the two devices “talk” to each other, and the output of one seriously modulates the output (technically speaking, the output load impedance ) of the other device. The result of this, which can be shown in very simple mathematical language, is that

 

the output voltage at each device now shows the very amplitude variations that the whole system was devised to remove.

 

So such a system, in the absence of a Chireix-style combiner, has much the same low efficiency as a conventional PA which amplifies the signal in its original, amplitude modulated form.

 

We should note that the Chireix style PA has been the subject of much attention in the PA community, with many papers and new patents appearing in the last few years. Even using the Chireix-style combiner, which in principle removes the amplitude variation at each PA output, many problems remain with this approach and it has seen little commercial deployment. In particular, the load modulation on which it depends creates a new source of non-linear effects which are absent in a conventional PA. The Chireix combiner also appears to be very limited in the RF bandwidth over which it can effectively operate.

 

The second core claim in the PV patent litany is the “MISO”, or “Multiple Input Single Output” configuration. This purports to extend the outphasing concept to the use of more than two input signals. This is certainly not new either (e.g. [3,4]). Implementors of the Chireix PA have reported that increasing the number of outphasing signals can have some benefits, such as reducing the range of phase deviation required in each signal input in order to obtaine a high dynamic AM range.

 

Beyond these core claims, PV indulge in lengthy and massively over-detailed embellishments, which in my experience are all standard established techniques in the industry, which are already widely used to improve the efficiency and linearity of RF PAs. Such techniques as bias adaption and digital predistortion are widely used and specific implementations have themselves been the subject of many patents. Indeed, we cannot help concluding that the later PV Patent,  xxxxxxx, which introduces such techniques, in effect contradicts the original claims in Patent xxxxxxxx, in effect admitting that the embellishments have become necessary to compete with existing techniques.

 

In conclusion, we think it is possible that PV can build a PA system which is based on their core concepts, which will have comparable, but not better, efficiency than that being routinely obtained and deployed using existing industry-standard PA technology.  We believe however, that the linearity will be significantly worse at comparable efficiency levels. This is based on our own experience developing Chireix-style PAs for WiFi applications. This degradation may respond favourably to linearization, such as digital predistortion (“DPD”. This however will not be an acceptable “fix” in a mobile transmitter application. DPD is now extensively used in high power multi-carrier basestation PAs, and the capabilities and requirements of the digital hardware are well established. In particular, a typical DPD system consumes several amps of current due to the need for very high speed sampling and processing in such a “real-time” DSP application.  Such a drain of power would obviously be unacceptable in a mobile transmitter, especially when current mobile PA products have no need for it.