Direct2Power (D2P) Analysis

Related Terms:

The latest version of the ParkerVision story is D2P (Direct2Power). It claims to be an RF power amplifer/transmitter technology that is significantly more power efficient than existing PAs. D2P was announced on January 20, 2005 (Breakthrough at ParkerVision Enables Common Silicon Semiconductors to Create Ultra-Efficient RF Power Amplifiers). In the conference call (Audio of D2P call, Transcript of D2P call), Jeff announced ParkerVision had working D2P silicon and that commercial shipments would begin by the end of 2005. To date (Dec. 2007) no commercial silicon has been demonstrated and no shipments have been made.


We now know that the "working silicon" did not, in fact, work. Parkervision appears to have implemented D2P as described in D2P Patent 7,184,723, which unfortunately doesn't actually result in high-efficiency (D2P patent analysis). Instead of announcing the problems with the first chip (apparently known problems by late 2005), ParkerVision has been busy trying to "patch" the problems of the original D2P idea. These patches are shown in the long list of provisional applications (List of D2P Applications).


The provisional applications use a highly nonlinear power amplifier (using dynamic bias adjustment and other techniques) combined with digital logic to implement predistortion to "undo" the nonlinearities. This digital logic can be implemented either in the baseband logic or in the PA itself. Sometime in late 2005 or 2006, Parkervision sharted showing a breadboard implementing the ideas in the provisional applications. The breadboard consists of some (nonlinear) RF circuitry combined with a FPGA (Field Programmable Gate Array) to implement the digital signal processing required to enable the predistorion. The FPGA is very power hungry and the breadboard is decidedly not high efficiency. ParkerVision claims that the functionality of the FPGA can be implemented in CMOS, and will result in very low power and hence high overall system efficiency.


We disagree. The significant signal processing requirements (as described in the provisionals) lead inherently to a very high power digital signal processing logic somewhere in the system - it doesn't matter whether the DSP is embedded in the baseband or in the PA, it will still consume significant power compared to the savings of the nonlinear amplifier. Especially at low output power (e.g. 16 dBm - where the efficiency claims of ParkerVision are the strongest), the constant and high power of the DSP will be too high.


We make three additional comments.


In the December 3, 2007 announcement (ParkerVision, Inc. Responds to Barron's Article), Jeff stated that D2P doesn't need a DSP. This is totally untrue, as shown by reading any of the provisional patent applications which clearly describe digital logic which implements the predistortion required. Also, the FPGA used in the ParkerVision breadboard clearly implements DSP functionality. Jeff is relying on the figleaf of not requiring a standalone DSP chip, but rather that the DSP logic is embedded. Since almost no consumer systems implemented today use standalone DSP chips (for example in a cell phone, there are many DSPs implemented in the baseband chip), this is a meaningless quibble. Clearly D2P needs DSP functionality.


Jeff Parker has repeatedly claimed that ParkerVision has "working silicon". This phrase (and the similar ("silicon verified IP")  has been used in most (if not all) of the conference calls of the last three years and was used in the Dec. 3, 2007 call. We find it very misleading, in that what ParkerVision has implemented and shown is a breadboard with an FPGA. To anyone in the industry, "working silicon" these days would be a single chips (CMOS or SiGe in ParkerVision's case) with all of the logic implemented and demonstrating not only functionality but low power and high-efficiency. ParkerVision has not done this.


Finally, the idea of using a nonlinear amplifier combined with analog or digital predistortion has been around for years. It is commonly used in cell phone base stations, where the PA power typically 20 (or so) watts. DSP logic that can save many watts of power in the amplifier at the cost of a few watts of power in the DSP is clearly worthwhile. The same technique doesn't work to save power in a 50 milliwatt amplifier (i.e. at 16 dBm output power), and, in any case, is not a new (or broadly patentable) idea.


 This page was updated in December 2007. For a link to our old page, see Old D2P Analysis.