I was looking at this document who's name I can only speculate. Its title and designation includes:
810-005, Rev. E, 201, Rev. B, DSN Telecommunications Link, Design Handbook, 201, Rev. B, Frequency and Channel Assignments, Released December 15, 2009.
The following section caught my eye:
2.2 Spacecraft Transponder Turnaround Ratios
To measure two-way or three-way Doppler shift, the spacecraft must transmit a downlink signal that is phase coherent with the uplink signal. Table 1 provides the recommended spacecraft transponder turnaround ratios for various uplink and downlink frequency bands. The tracking equipment at the DSN 34-m and 70-m stations can accommodate other turnaround ratios but this support must be negotiated through the JPL Frequency Manager, who is resident in the Plans and Commitments Program Office.
Doppler shift is an essential technique for measuring the spacecraft trajectories. Along with absolute range from delay measurements, the rate of change of range provides a precise one-dimensional projection of the spacecraft's velocity. Combined with sophisticated orbital mechanics calculations, these can be used to locate a spacecraft in 3D within the solar system.
For two-way doppler shift measurements, A signal of known frequency is (usually) transmitted from Earth to a spacecraft. When it is received, the uplink frequency is shifted by a stable internally generated frequency or by a specific PLL-controlled divide-by ratio to shift it to a new, downlink frequency and transmitted back toward Earth. The spacecraft often (always?) uses two different bands so that it's sensitive receiver is not saturated by the simultaneous high-power broadcast of the shifted frequency.
While range-rate can be done on passive objects (e.g. NEO asteroids or very large planets) using standard radar techniques and picking up the weak, passively reflected signal, spacecraft at deep-space distances are too small to reflect a useful amount of power, so a boost in power by the spacecraft's sensitive receiver and powered retransmission is necessary.
Question: How does a three-way Doppler shift measurement work in practice? I'm looking for some level of technical specificity here, not just hypothesizing, generalizations, or hand-waving.