Spread Spectrum is the art of secure digital communications that is now being exploited for
commercial and industrial purposes. Hardly anyone can escape being
involved, in some way, with spread spectrum communications these days. Applications for commercial
spread spectrum range from wireless LAN's (computer to computer local area networks), to
integrated bar code scanner/palmtop computer/radio modem devices for warehousing, to digital
dispatch, to digital cellular telephone communications, to "information society"
city/area/state orcountry wide networks for passing faxes, computer data, email, or multimedia data.
On this page and on our Tutorial Page, we endeavor to provide you with some
basic knowledge about this fascinating and useful technology. Some of our tutorials are aimed at interested
laypeople, with easy to understand explanations and little math, while others are aimed at working
engineers. Please explore some of our introductory topics below:
Virginia Polytechnic Institute's Spread Spectrum Introduction
The term spread spectrum describes a modulation technique
which makes the sacrifice of bandwidth in order to gain signal-to-noise
performance. Basically, the SS system is a system in which the
transmitted signal is spread over a frequency much wider than
the minimum bandwidth required to send the signal. The fundamental
premise is that, in channels with narrowband noise, increasing
the transmitted signal bandwidth results in an increased probability
that the received information will be correct. If total signal
power is interpreted as the area under the spectral density curve,
then signals with equivalent total power may have either a large
signal power concentrated in a small bandwidth or a small signal
power spread over a large bandwidth.
From a system viewpoint, the performance increase for very
wideband systems is referred to as "process gain". This
term is used to describe the received signal fidelity gained at
the cost of bandwidth. The numerical advantage is obtained from
Claude Shannon's equation describing channel capacity:
C=W log2 (1+ S/N)
where C = Channel capacity in bits, W = Bandwidth in Hertz,
S = Signal Power, and N = Noise Power
From this equation, the result of increasing the bandwidth
becomes apparent. By increasing W in the equation, the S/N may
be decreased without decreased BER performance. The process gain
(GP) is what actually provides increased system performance without
requiring a high S/N. This is described mathematically as:
GP = BWRF/RINFO
where BWRF = RF Bandwidth in Herz and RINFO = Information
rate in bits/second.
The baseband signal is spread out to BWRF over the channel
(see Fig. 1). Then at the receiving end, the signal is de-spread
by the same amount by a correlation with a desired signal generated
by the spreading technique (more on the different spreading techniques
later). When the received signal is matched to the desired signal
the baseband/information signal is retrieved.
Fig. 1 Bandwidth Spreading
Signal Spreading works quite well in situations with strong
narrowband interference signals, since the SS signal has a unique
form of frequency diversity. The actual signal spreading may be
achieved with one of three basic techniques. These include:
frequency hopped and pulsed FM or hybrid forms.
Direct Sequence Spread Spectrum (DSSS)
This is probably the most widely recognized form of spread spectrum. The
DSSS process is performed by effectively multiplying an RF carrier and a
pseudo-noise (PN) digital signal. First, the PN code is modulated onto the
information signal using one of several modulation techniques (e.g. BPSK,
QPSK, etc ). Then, a doubly balanced mixer is used to multiply the RF carrier
and PN modulated information signal. This process causes the RF signal to
be replaced with a very wide bandwidth signal with the spectral equivalent
of a noise signal. The demodulation process (for the BPSK case) is then simply
the mixing/multiplying of the same PN modulated carrier with the incoming
RF signal. The output is a signal that is a maximum when the two signals
exactly equal one another or are "correlated." The correlated signal is then
filtered and sent to a BPSK demodulator.
The signals generated with this technique appear as noise in the frequency
domain. The wide bandwidth provided by the PN code allows the signal power
to drop below the noise threshold without loss of information. The spectral
content of an SS signal is shown in Fig. 2. Note that this is just the spectrum
of a BPSK signal with a (sin x / x)2 form.
Fig. 2 BPSK DSSS Spectrum
The bandwidth in DSSS systems is often taken as the null-to-null bandwidth
of the main lobe of the power spectral density plot (indicated as 2Rc in
Fig. 2). The half power bandwidth of this lobe is .88 Rc, where Rc is the
chip rate. Therefore, the bandwidth of a DSSS system is a direct function
of the chip rate; specifically 2Rc/RINFO. This is just an extension of the
previous equation for process gain. It should be noted that the power contained
in the main lobe comprises 90 percent of the total power. This allows a narrower
RF bandwidth to accommodate the received signal with the effect of rounding
the received pulses in the time domain.
One feature of DSSS is that QPSK may be used to increase the data rate. This
increase of a factor of two bits per symbol of transmitted information over
BPSK causes an equivalent reduction in the available process gain. The process
gain is reduced because for a given chip rate, the bandwidth (which sets
the process gain) is halved due to the two-fold increase in information transfer.
The result is that systems in a spectrally quiet environment benefit from
the possible increase in data transfer rate.
Frequency Hopped Spread Spectrum (FHSS)
Frequency hopping relies on frequency diversity to combat interference. This
is accomplished by multiple frequency, code-selected FSK. Basically, the
incoming digital stream is shifted in frequency by an amount determined by
a code that spreads the signal power over a wide bandwidth. In comparison
to binary FSK, which has only two possible frequencies, FHSS may have 2*10^20
The FHSS transmitter is a pseudo-noise PN code controlled frequency synthesizer.
The instantaneous frequency output of the transmitter jumps from one value
to another based on the pseudo-random input from the code generator (see
Fig. 3). Varying the instantaneous frequency results in an output spectrum
that is effectively spread over the range of frequencies generated.
Fig.3 FHSS Spectrum
In this system, the number of discrete frequencies determines the bandwidth
of the system. Hence, the process gain is directly dependent on the number
of available frequency choices for a given information rate.
Another important factor in FHSS systems is the rate at which the hops occur.
The minimum time required to change frequencies is dependent on the information
bit rate, the amount of redundancy used, and the distance to the nearest
This article is re-printed from "Spread Spectrum Scene" magazine
(paper version), Volume 3, Number 3, page 8, and was written by Randy Roberts,
RF/SS Consulting (Retired)
This frequently asked question is really rather hard to answer. There is no real unbiased way
to compare today's crop of commercial direct sequence radios to the frequency hoppers that are
available. Sure, claims and counter claims abound, but the truth is hard to find.
Why? A little history helps explain what has evolved in the commercial SS world. First of
all, the FCC's first Part 15 rules (published in 1989), did not require any SS radio to have processing
gain - nor did these initial rules differentiate between fast and slow hopping. Thus the earliest SS
radios produced, could use almost anything as long as they met the then defined Part 15 rules.
Some of these early radios used post detection correlation and thus, were not "TRUE
DIRECT SEQUENCE" radios, at all. Only when correlation is done before detection, can all of the
anti-jam and anti-interference benefits of direct sequence be seen. Some of the early hoppers
changed frequency so slowly that they transmitted tens of thousands of bits on a single frequency
dwell (and made no provision for error detection - let alone correction).
It's no wonder then, that some of these early radios (of either variety) were very short of the
long hyped interference immunity that they were supposed to have. In fact, in Europe and the United
Kingdom, Direct Sequence has gotten such a bad name from early trials with overly simple Direct
Sequence radios, that frequency hoppers have almost become a standard.
The FCC tried to rectify this situation in 1992, with new Part 15 technical rules that require
a minimum processing gain and better definitions of hopping speed and numbers of hopping
channels required. But, out of intense lobbying efforts, came "grandfather provisions" that allowed
existing approved designs to be sold for 5 years beyond 1989. The most recent actions of the FCC,
however, have granted "dispensations" to those "grandfathered" manufacturers who yelled the
loudest. The "deal" that was struck allows slow hoppers and post detection correlation (Non-TRUE
DIRECT SEQUENCE) radios to continue to be sold if they keep their power output below 100
So if a manufacturer cannot furnish a radio with significantly more power than 100 mW,
they are probably peddling an old, inferior design - Caveat Emptor!
So the answer to the which is better is still unclear -- neither is any good, if it's an old
design! Fast hoppers (no more than a few bits per frequency dwell) can have almost identical
performance to Direct Sequence.
Real (or TRUE) DS and FH radios can each be vulnerable to certain kinds of interference.
No one modulation is best against any and ail interferers! However, the best that can be done with
SS is to use a hybrid, or combination of DS and FH, that adapts to channel conditions in real time.
The BEST SS modulation is thus seen to be not either DS or FH -- but both, when used optimally
against adverse interference, multipath and channel conditions.