Science: Begin at the Beginning for Transmitter Design

Earl McCune PhD is a successful serial Silicon Valley Entrepreneur, founding both the Digital RF Solutions Corporation in 1986 and Tropian Inc. in 1996. He is a member of several IEEE conference committees, and is a worldwide invited speaker on modulation issues and polar-coordinate signal processing.

Any wireless system exists in service to an application. Consider these example cases:

  • television or radio broadcast in service to the entertainment and marketing (advertising) businesses,
  • cellular phones in service to the transformation of the telephone network from having phone numbers relate to a place to where they represent a person, or
  • Bluetooth in service to the desire to separate the cellular phone transmitter from being next to the head of the user.

In all of these cases it is the application that sets the economic value of the wireless technology to be used, and not the other way around. System implementation costs must be low enough so that providing the needed wireless products is a profitable business within the application. Researchers, standards committees, product managers, and development engineers must all be aware of these underlying costs to successfully provide useful standards and products that are profitable.

Fundamentally, any wireless system must accurately transfer information from one place to another. This transfer makes use of modulated electromagnetic signals: radio waves. These modulated signals need a certain amount of bandwidth – spectral space – in order to successfully communicate. A primary metric is bandwidth efficiency, meaning how much bandwidth is needed to transfer a particular data rate. Physics unfortunately sets a limit to achievable bandwidth efficiency, so we are not free to select arbitrarily high application design goals. And this limit includes the effects of fancy signal processing so elaborate algorithms, including adaptive modulation, allow us to just approach the limit. Only the use of higher transmitter power can increase this physical limit.

Transmitter power is expensive, so we do not want more than we really need. In fact, any specification has a cost impact. The selection of tight specifications by standards committees effectively “bakes in” a higher cost foundation to any product implementation. This is acceptable if the value of the high cost features is important to the application, and the application users are willing to pay this cost. Otherwise a high baseline cost can kill the desired business opportunities. It is my habit to follow the “Keep it Simple” process, always starting from simple and low cost implementations. Higher cost features are only accepted after proving the value of the added feature exceeds its cost.

Energy efficiency in wireless technology is critical. This is true whether to achieve a long battery life in a mobile device, or to draw less electric power when plugged in to result in a (comparatively) low carbon impact. Energy efficiency is maximized when dissipated power is minimum – the device must not get hot. This leaves us a problem though, because circuits which generate high quality signals necessarily get very hot.

We usually have to accept increased distortion for improved energy efficiency. Adaptive correction is increasingly used to improve energy efficiency with no distortion increase. But as always this comes at a price, here being much higher implementation complexity and its associated additional manufacturing cost. The low cost of additional complexity using digital integrated circuit technology unfortunately applies only in a limited way to the radio transmitter application.

Proper selection of signal type can help – or hurt – the need to meet economic goals in wireless applications. The GMSK (Gaussian Minimum-Shift Keying) signal used for GSM mobile phones is a very low cost signal to implement, which definitely helped the rapid growth and adoption of mobile phones. The more recent signal OFDM (Orthogonal Frequency Division Modulation) is more controversial: the bandwidth efficiency is no better than a conventional QAM (quadrature amplitude modulation) signal, but the OFDM transmitter must be designed for about 8 times higher power. The digital signal processing is easier in an outdoor environment, at the price of much higher product cost. The value evaluation of this price is still pending.

Newer system types as currently proposed do not use new signals, but rather use existing signals in different ways. Two proposals, Smart Antennas and MIMO, both try to improve bandwidth efficiency. But each in its own way encounters the physical limit discussed above and is forced to acknowledge the need for higher power to actually get the desired bandwidth efficiency. Physics cannot be cheated. We need to understand this and work smartly within these immovable bounds to make money – the true ultimate objective.

Dr. McCune’s latest book “Practical Digital Wireless Signals” presents principles and economics of digital wireless signals emphasizing a physical approach, complementing the mathematical approach of textbooks.


Leave a Comment

We don't require your email address, or your name, for anyone to leave a comment. If you do add an email address, you may be notified if there are replies to your comment - we won't use it for any other purpose. Please make respectful comments, which add value, and avoid personal attacks on others. Links are not allowed in comments - 99% of spam comments, attempt to post links. Please describe where people may find additional information - for example "visit the UN website" or "search Google for..." rather than posting a link. Comments failing to adhere to these guidelines will not be published.