Very minimum-shift keying

VMSK, for Very Minimum Shift Keying modulation, is one of several digital modulation methods claimed to send high speed digital data through very low bandwidth (or narrowband) channels. A typical claim is a data rate of 6 Mbit/s in a bandwidth of 1 kHz or less using the same (or even less) transmitter power than conventional schemes.

Analysis

The claims of VMSK proponents are to communications what perpetual motion is to energy production, as they run afoul of firmly established mathematical principles such as the Shannon-Hartley theorem and the Nyquist-Shannon sampling theorem. Proponents claim to have built working hardware. However, a careful analysis shows that all these schemes actually produce various forms of ultra wideband modulation to which a strong, narrowband, information-free component is added as a distraction. Only the strong narrowband component is easily seen on a spectrum analyzer, so it is incorrectly alleged that it must convey the information being transmitted. This claim is readily dismissed with the Nyquist-Shannon sampling theorem which states that for a receiver to be able to distinguish among data-dependent changes made to a signal at a given rate, the channel bandwidth must be at least one half that signaling rate.

The information in all these allegedly ultra-narrow-band signals is actually conveyed in the very weak, very wide band spectral components that generally fall well below the noise floor of a spectrum analyzer. The conspicuous narrow band component conveys no information; it merely wastes most of the transmitter power.

That VMSK is extremely wasteful of transmitter power as well as bandwidth can also be seen from the time-domain waveform.The VMSK waveform for a digital '1' is nearly identical to that for a digital '0'; they differ only in a very small region near the centerof each bit. The energy in the rest of the bit does not help the receiver distinguish between a '1' and a '0', so it is simply wasted.This is in stark contrast to conventional, power-efficient schemes such as BPSK, with an antipodal signal set, and FSK,with orthogonal signals. The waveforms for 0's and 1's must be as different as possible to help the receiverdistinguish between them in noise.

According to proponents, the basic idea behind those schemes is "to find ways to slightly mark the carrier wave with the modulation so that the least distortion of the carrier wave is achieved." [http://www.vmsk.org/Comments.pdf] . However, it is impossible to do so without generating spectral components whose bandwidths increase proportionately with the modulation rate. The amplitude of the generated wide band components can be kept low if the amplitude of the modulation is kept small, but this simply reflects the inefficientuse of transmitter power.

While proponents of VMSK concede that filtering out those ultra-wideband components (sometimes called "useless grass" in most VMSK literature) makes the data unrecoverable, they counter that a "zero-group delay filter" is required to preserve the data and make the (high frequency) phase modulation information recoverable by the low-frequency carrier alone. This amounts to a concession that VMSK is a wideband signal, as "any" filter with a small group delay must necessarily have a large bandwidth according to Fourier transform theory.

In other words, these schemes represent viable -- though extremely inefficient and useless -- means of sending digital data. Only the bandwidth claims made for them are impossible, along with the implication (never actually demonstrated) that these schemes could increase the capacity of the radio spectrum beyond that achievable with conventional modulation methods.

Part of the misconception rests on the fact that there is no single, universally applicable definition of the word "bandwidth". Different definitions exist, and using the wrong one can lead to meaningless or absurd results. VMSK proponents are especially fond of citing the "noise bandwidths" of their signals, but the concentrated power in the information-free spectral line gives pathologically small results. The correct definition of the minimum bandwidth of any digital signal is given by the Nyquist sampling theorem: one half the symbol rate.

Personalities

Today, most VMSK or related modulation schemes seem to be regularly studied, developed and documented by H. R. Walker, owner of the [http://www.vmsk.org/ "Welcome to the Ultra Narrowband Club"] website, and author of several papers and publications on the arguments. Some of Mr. Walker's works have even gained the attention of telecommunications sector publications like [http://www.edn.com/article/CA47127.html EDN] , [http://www.mwrf.com/Articles/Print.cfm?Ad=1&ArticleID=6992 Microwaves & RF] and [http://www.commsdesign.com/news/OEG20020227S0026 CommsDesign] , as well as several cooperations and paper co-authorships with academic environments [http://rcl.physics.auth.gr/En/Profile/Com.htm] .

Interestingly, Mr. Walker often refers to the works of "Professor Howe" in many of his VMSK related papers, who allegedly had proposed and proven that a VMSK-like modulation scheme could work since 1939 ("Professor Howe, as published in K.R. Sturley, Frequency Modulation, Chemical Publishing Co., Brooklyn, NY (from "Wireless Engineer," November 1939, p. 547."). However, other than in Mr. Walker's works, references to "Professor Howe", his academic career and publications--even his actual existence--are obscure at best.

xG

Early, sketchy descriptions of xG technology made it appear to be another "ultra narrowband" scheme similar to VMSK. However, it has been more recently described as a hybrid of anarrow band, low speed channel used to coordinate the actual data transmission on an ultra wideband channel. This would remove it from the "ultra narrowband" category.

External links

* [http://www.vmsk.org/ Welcome to the Ultra Narrowband Club] VMSK developer's website
* [http://www.xgtechnology.com/ xG Technology]
* [http://www.ka9q.net/vmsk/ The VMSK Delusion] Phil Karn's analysis of ultra narrowband claims