On 3/22/21 3:38 PM, Robert McGwier wrote:
It is equivalent to QPSK with a good implementation.
I disagree with your statement of equivalence, but only slightly, for a couple of reasons: QPSK encodes 2 bits per symbol, so you should have stated BPSK (1 bit per symbol).
Additionally, GMSK's performance (probability of error) is close, but not equal, to BPSK for a given Eb/N0. The difference is up to several dB of required signal power for the same bit error rate. These numbers assume no channel coding to correct for errors on either system.
There's always a trade-off between bandwidth and power. GMSK trades off power for better-constrained bandwidth, while BPSK does the opposite. Not having to have a well-linear amplification system helps out with system efficiency, but noise performance offsets this gain.
An interesting paper by a pair of JPL employees explores this very subject:
"GMSK Modulation for Deep Space Applications" https://trs.jpl.nasa.gov/bitstream/handle/2014/44994/12-0053_A1b.pdf?sequenc...
--- Zach N0ZGO