What it is for
Testing a receiver in the field is not very repeatable: the environment changes, interference varies and it is impossible to reproduce exactly the same condition twice. A channel emulator takes the signal from a transmitter or generator, modifies it with a mathematical model of the channel and delivers the resulting signal to the receiver. This makes it possible to compare design versions, locate defects and verify performance under difficult conditions, with reproducible results.
Which effects it can reproduce
- Multipath: several copies of the signal with different delays, amplitudes and phases.
- Fading: variation of each path over time according to a chosen statistical distribution.
- Doppler effect: frequency shift caused by relative motion.
- Additive noise: added at a controlled level to set the signal-to-noise ratio.
- Interference: other signals added to the wanted one.
- Multi-antenna environments: correlation between the channels of systems with multiple antennas.
How it is used
A propagation model is selected, often one defined in a specification document, and its parameters are set. The receiver is connected by cable, with suitable attenuation, so that the only thing disturbing it is what the emulator introduces. Performance, for example error rate or data rate, is then measured while the severity of the channel is varied.
What you learn
You find out at what signal-to-noise ratio the receiver starts to fail, how much delay between paths its equalizer tolerates, how it reacts to speed of movement and how well it estimates the channel. A receiver that works well with noise but poorly with fading has a problem in estimation or equalization, not in sensitivity.
Limits
A model is a simplification: it reproduces statistics, not a specific building. That is why emulation complements field tests but does not replace them. In addition, accuracy depends on the calibration of the levels and on the model being appropriate for the environment to be represented.