0
$\begingroup$

In this answer to Have deep-space spacecraft always used some form of spread-spectrum for data downlink? space scientist Mark Adler explains that while spread-spectrum proper (using specific codes in the mod/demod processing) is used for ranging and Doppler signals, for data and voice it's generally not used.

But that doesn't mean they use the narrowest possible spectrum. Because of the directionality of the deep space antennas and scheduling, they don't worry about the spectrum being crowded.

But this discussion got me wondering if there are modulation schemes that don't require fixed codes that can use substantially more bandwidth than the data requires. I don't just mean occupy, I'm asking about schemes that make use of a much larger bandwidth to do things like improve S/N or error correction.

The biggest one I can think of is AM radio, where each sideband contains the same information. But I am not sure if a standard AM radio receiver gets any improvement in S/N over a single-sideband receiver would (from a signal of the same average power).

So I've tentatively excluded simple AM radio in the title and asked for schemes that are designed to use a much wider spectrum, but without requiring codes.

Question: Are there modulation schemes that use more bandwith than they need to, but don't require a spread-spectrum code to decode? (besides AM)

$\endgroup$
3
  • 2
    $\begingroup$ Your question contains factual inaccuracies. Specifically, spread spectrum coding does not improve SNR or error correction. For a given transmit power and white noise, in theory at least spread spectrum performs neither better nor worse than its mother signal without spreading. Spread spectrum enhances performance when there are interfering sources that are not white noise; error correcting coding does spread the spectrum (or, at least, increase the number of raw bits transmitted), but it is not referred to as "spread spectrum". $\endgroup$
    – TimWescott
    Apr 16 at 3:18
  • $\begingroup$ @TimWescott The purpose of putting those in there is to give a sampling of my current understanding to assist answer-authors in setting me straight. I'm curious; if the N in S/N is only thermal noise coming from the temperature of the receiver front-end, then I think I understand what you mean. $\endgroup$
    – uhoh
    Apr 16 at 5:26
  • $\begingroup$ @TimWescott But is it always clear - especially when picking up signals from deep space with the cosmic microwave background and interplanetary plasma and the Earth's ionosphere - that spread spectrum never helps in these cases, and is only helping when substantially "non-white noise" (e.g. other signals with information) are present? That would certainly explain why NASA doesn't use it except in the case of ranging and Doppler measurements (where they use a coherent transponder to return the transmitted signal). $\endgroup$
    – uhoh
    Apr 16 at 5:26

1 Answer 1

5
$\begingroup$

Yes, analog wideband FM is another example where the occupied bandwidth is wider than the bandwidth of the modulation signal, and provides a similar processing gain to digital spread spectrum modulation.

Here is a reference detailing the SNR for wide and narrow band FM:

https://charleslee.yolasite.com/resources/elec321/lect_snr_afm.pdf

See equation 29.46 on page 29.9 in that reference showing how as $\beta$ increases (for FM $\beta$ is defined for a sinusoidal modulation signal as $\beta = f_{dev}/f_{mod}$ where $f_{dev}$ is the peak frequency deviation and $f_{mod}$ is the rate of the sinuosoid), the output SNR increases, up to a lower threshold of minimum input CNR as explained in the paper and demonstrated nicely in their figure 29.8 copied below:

SNR for FM demod

For digital modulations, Ultra-Wideband (UWB) is another case where very narrow pulses are transmitted, resulting in a much wider bandwidth than the modulation signal.

Ultimately, the use of bandwidth in waveforms comes down to the trade-space of power efficiency vs bandwidth efficiency. A power-efficient modulation like BPSK, compared to a bandwidth-efficient modulation like 4096-QAM, will use "substantially more bandwidth than the data requires": If we want to send 4 Mb/s, we can do this in about 350 KHz of BW with a 4096-QAM modulation (and of course a lot more power!). The BPSK system in comparison would require about 4.2 MHz of BW with the benefit of requiring much less power - which is a similar example of using more bandwidth to improve SNR performance (with more bandwidth we can transmit a given data rate with less SNR). All of this is well captured in the Shannon-Hartley theorem showing the trade of bandwidth and power in achieving a data capacity. In the simplest case of this we could send the same message multiple times to improve SNR performance at the expense of occupied BW compared to the eventual data rate we actually get. This opens us up to many alternative bandwidth expanding modulations including frequency chirps (where the frequency is swept over the occupied bandwidth in a ramp such as used in FMCW radar; this could equally be applied to communications as well where for example a ramp up could represent a "1" and ramp down represent a "0").

Note regarding the OP's opening paragraph; I read the link for what Mark Adler wrote and see he was talking about deep space specifically and only - not communications in general. There are MANY terrestrial applications that have used spread spectrum for both wired and wireless data communications: GPS uses spread spectrum signals both for ranging, but also for data transmission. Spread spectrum is used for voice and data communications in military systems given its resistance to fading, jamming, and eavesdropping. Wifi 802.11, 802.11b and 802.11g use Direct Sequence Spread Spectrum (DSSS). (802.11 also used Frequency Hopping Spread Spectrum). The wireless protocol Zigbee for low power low bandwidth datalinks uses DSSS. Cordless Phones use spread spectrum for voice communication.

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.