They are not based on Doppler- but rather radial pulse timing. Very clever use of technology back in the day.

The doppler signal encodes the station identifier, i(t), optional voice, a(t), navigation variable signal in c(t), and the isotropic (i.e. omnidirectional) component. The navigation variable signal is A3 modulated (greyscale). The navigation reference signal is delayed, t+, t?, by electrically revolving a pair of transmitters. The cyclic doppler blue shift, and corresponding doppler red shift, as a transmitter closes on and recedes from the receiver results in F3 modulation (colour). The pairing of transmitters offset equally high and low of the isotropic carrier frequency produce the upper and lower sidebands. Closing and receding equally on opposite sides of the same circle around the isotropic transmitter produce F3 subcarrier modulation, g(A,t).

t = t + ( A , t ) ? ( R / C ) sin ? ( 2 ? F n t + ( A , t ) + A ) t = t ? ( A , t ) + ( R / C ) sin ? ( 2 ? F n t ? ( A , t ) + A ) e ( A , t ) = cos ? ( 2 ? F c t ) ( 1 + c ( t ) ) + g ( A , t ) c ( t ) = M i cos ? ( 2 ? F i t ) i ( t ) + M a a ( t ) + M n cos ? ( 2 ? F n t ) g ( A , t ) = ( M d / 2 ) cos ? ( 2 ? ( F c + F s ) t + ( A , t ) ) + ( M d / 2 ) cos ? ( 2 ? ( F c ? F s ) t ? ( A , t ) ) {\displaystyle {\begin{array}{rcl}t&=&t_{+}(A,t)-(R/C)\sin(2\pi F_{n}t_{+}(A,t)+A)\\t&=&t_{-}(A,t)+(R/C)\sin(2\pi F_{n}t_{-}(A,t)+A)\\e(A,t)&=&\cos(2\pi F_{c}t)(1+c(t))\\&+&g(A,t)\\c(t)&=&M_{i}\cos(2\pi F_{i}t)~i(t)\\&+&M_{a}~a(t)\\&+&M_{n}\cos(2\pi F_{n}t)\\g(A,t)&=&(M_{d}/2)\cos(2\pi (F_{c}+F_{s})t_{+}(A,t))\\&+&(M_{d}/2)\cos(2\pi (F_{c}-F_{s})t_{-}(A,t))\\\end{array}}}

where the revolution radius R = Fd C / (2 ? Fn Fc ) is 6.76 ± 0.3 m .

The transmitter acceleration 4 ?2 Fn2 R (24,000 g) makes mechanical revolution impractical, and halves (gravitational redshift) the frequency change ratio compared to transmitters in free-fall.

The mathematics to describe the operation of a DVOR is far more complex than indicated above. The reference to "electronically rotated" is a vast simplification. The primary complication relates to a process that is called "blending".[citation needed]

Another complication is that the phase of the upper and lower sideband signals have to be locked to each other. The composite signal is detected by the receiver. The electronic operation of detection effectively shifts the carrier down to 0 Hz, folding the signals with frequencies below the Carrier, on top of the frequencies above the carrier. Thus the upper and lower sidebands are summed. If there is a phase shift between these two, then the combination will have a relative amplitude of (1 + cos ?). If ? was 180°, then the aircraft's receiver would not detect any sub-carrier (signal A3).

"Blending" describes the process by which a sideband signal is switched from one antenna to the next. The switching is not discontinuous. The amplitude of the next antenna rises as the amplitude of the current antenna falls. When one antenna reaches its peak amplitude, the next and previous antennas have zero amplitude.

By radiating from two antennas, the effective phase centre becomes a point between the two. Thus the phase reference is swept continuously around the ring – not stepped as would be the case with antenna to antenna discontinuous switching.

In the electromechanical antenna switching systems employed before solid state antenna switching systems were introduced, the blending was a by-product of the way the motorized switches worked. These switches brushed a coaxial cable past 50 (or 48) antenna feeds. As the cable moved between two antenna feeds, it would couple signal into both.

But blending accentuates another complication of a DVOR.