1

In a point-to-point mode, that is, one transmitter and one receiver, I understand how a spread spectrum receiver can read the transmitted signal from a spread spectrum transmitter by knowing the code. But I cannot understand how different codes can be sufficient to separate signals from multiple transmitters on the same bandwidth.

When there are multiple transmitters transmitting at the same frequency and at the same time, even if they have different codes, don't their signals still interfere with each other? How come codes are not affected by interference?

Noob_Guy
  • 395
  • 2
  • 11

2 Answers2

1

There are (at least) two types of spread-spectrum.

Frequency hopping and direct sequence spread spectrum (DSSS)

In DSSS all transmitters share the same frequency but with a different spreading code.

Multiple transmitters do interfere but after doing a cross-correlation using the specific code the unwanted signals get reduced by the cross-correlation factor. Ideally there will be zero cross-correlation but in practice, the codes used are not perfect and other effects such as doppler and range attenuation will make things worse.

Kevin White
  • 32,097
  • 1
  • 47
  • 74
  • So the magic is in the 'cross-correlation'? Meaning even if the signal is garbled due to interference we can still recover it using cross-correlation? – Noob_Guy May 11 '20 at 06:43
0

CDMA (code division multiple access)

There are a multitude of frequency slots and any one transmitter occupies several of that multitude. The precise slots occupied are based on the "code word" and so another transmitter coming along is very unlikely to occupy any significant portion of the slots already "owned". The relevant receiver understands the code word and can demodulate the signal based on the knowledge of what slots are intended to be used by its partner radio.

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • Only true for FHSS. DSSS uses the same frequency. – user110971 May 10 '20 at 21:48
  • No it uses multiple simultaneous frequencies hence, the spectrum is spread. – Andy aka May 10 '20 at 22:29
  • what I mean is multiple channels occupy the same bandwidth. So, when you use DSSS, the same frequencies are used at the same time by all the channels. – user110971 May 10 '20 at 22:31
  • Does it mean FHSS relies on the low probability that multiple transmitters will occupy the same frequency? – Noob_Guy May 11 '20 at 06:52
  • @user110971 CDMA spreads the spectrum (hence the name) and it spreads the spectrum in a way defined by the codeword hence, two users may overall occupy the same approximate area of the spectrum but their respective key spectra do not (or should not) clash. To say they use *the same frequencies are used at the same time* is missing the point. – Andy aka May 11 '20 at 07:05
  • @Andyaka In DSSS all channels occupy the exact same frequencies at the same time. Consider a BPSK modulated DSSS system, e.g. GPS. During a single chip you either transmit a 0 phaseshifted signal or a 180 phaseshifted signal. This is rect times sin or rect times cos in the time domain. In the frequency domain it becomes two sinc functions with the same magnitude centered at the carrier frequency \$\pm f_c\$. As you can see, all channels occupy the exact same frequency range at the same time. – user110971 May 11 '20 at 07:59
  • @user110971 maybe you should make an answer to this question: *But I cannot understand how different codes can be sufficient to separate signals from multiple transmitters on the same bandwidth.* asked by the OP? – Andy aka May 11 '20 at 08:14
  • @Andyaka I was going to, but Kevin’s answer is a good one. Basically, the despreading operation is the same as the spreading operation. So, when you despread your target signal all other signals get spread again. Of course, as Kevin said, all the codes are not completely uncorrelated. Hence you get some interference due to the cross-correlation. From a theoretical perspective, it’s not that surprising. Remember, the Shannon—Hartley theorem yields \$C = B \log_2(1 + S/N)\$. Thus even if multiple channels occupy the same frequency range, i.e. SNR < 1, you can increase the bandwidth accordingly. – user110971 May 11 '20 at 08:27