3
\$\begingroup\$

What is the difference between Pulse Amplitude Modulation (PAM) and Pulse Code Modulation (PCM)?

Both seem to be effectively sampling an Continuous Time (CT) analogue wave and converting to a Discreet Time (DT) wave.

Images & animations would be super helpful!

Related question: link.

\$\endgroup\$
3
\$\begingroup\$

PAM encoding uses the physical amplitude of the sample as the final modulation (as seen on this page); i.e. it is an analogue modulation technique (the amplitude used for modulation is the actual sampled value, not the nearest approximation as is used in PCM, although it can be bounded).

Put another way, it is not (as part of the modulation) converted to a digital data word.

This modulation method forms the basis of QAM

When used as an analogue method, there is no inherent quantisation noise.

PCM takes the same thing a step further by requiring a step size (which may not be linear across the range) and then encoding that as a digital value (which will have a certain number of bits per sample); this will always introduce quantisation noise.

PCM data is widely used in the telephone system in most countries.

\$\endgroup\$
0
\$\begingroup\$

To be accurate, PAM needs to settle to a final value. For 10 bits, you need 7Tau settling; at 2*pi tau per cycle, this bandwidth is just a small amount faster than the channel sampling rate.

For PCM, you need e.g. 10 bits of NRZ or of other modulation method, per each sample. For trustable data recovery, you need to settle to a healthy logic "1" or "0", probably 3 tau, to have wide-open data eye. Thus you need 10bits * 3, or 30 tau. At 2*pi tau per cycle, you need bandwidth to be 30/6.28 or ~ 5 times the channel sampling rate.

Thus PAM can be usefully slower than PCM, but PCM has much better noise tolerance.

\$\endgroup\$
  • \$\begingroup\$ What do you mean by "PAM can be usefully slower than PCM"? The Stratix 10 line of FPGAs have transceivers that use "56 Gbps PAM4" \$\endgroup\$ – Sam Feb 6 '17 at 0:07
  • \$\begingroup\$ Usefully slower? If you have a band-limited channel, with high SNR, then PAM may be the better approach. The reason the rocket/missile telemetry engineers converted from analog to digital was the extreme ranges needed and the resultant low SNR. \$\endgroup\$ – analogsystemsrf Feb 6 '17 at 1:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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