A VOR receiver receives two sine signals repeating 30 times per second: reference and variable, and determines the bearing by measuring the delay (phase) between them.
The reference information sent in all directions is identical for all receivers, disregarding the strength which depends on distance. It is timed to be equal to, or more accurately to have the same phase than the variable signal when the receiver is located north of the VOR. For the other bearings the phase difference increases by one degree for each degree of bearing counted clockwise.
In the CVOR, the variable signal is the space modulation product of two AM modulated carriers sent from four antennas (or four slots on a cylinder) spaced by less than a meter. Space modulation creates a field pattern in space which depends on the relative characteristics of the two original HF signals. The parameters are selected to produce a directional cardioid rotating 30 times per second. When sensing the electric field strength of this pattern, the receiver sees an apparent 30 Hz AM. All receivers sees the same strength variation cycle, but the timing relative to the reference cycle depends on the receiver bearing.
Similarly the variable signal is produced in the DVOR by the apparent rotation of two opposite antennas at Mach 4, introducing a Doppler shift of ±480 Hz on two otherwise information-less carriers they transmit. The rotation is done electronically by scanning an array of about 50 antennas on a circle of 14 m. The perceived Doppler shifts appear to the receiver as a 30 Hz FM. All receivers will have the same shift cycle, but again the timing depends on the receiver bearing. The advantage of the DVOR is the variable signal information doesn’t change with distance, the Doppler shift remains.
When the variable signal is sent in AM the reference is sent in FM, and vice-versa. This prevents, in the original CVOR, one signal to influence the other, the DVOR just reuses the modulations but swaps the information they represent.
The same receiver is used. It computes the phase difference between 30 Hz AM and FM signals, to obtain a value which for the same bearing has an opposite sign for the CVOR and the DVOR, because the information carried by AM and FM are swapped. To associate the correct angle with a given reference value, the DVOR rotates counter-clockwise and inverts the Doppler shift.
Both types of VOR (and their different subtypes) are required by ICAO to produce a bearing with a ±2° tolerance. The DVOR is better in this aspect because it’s easier to produce an accurate Doppler shift than to form a precise pattern with space modulation.
Once the signals are emitted, they are subject to reflection on terrain and obstacles. The final composite error at the receiver location cannot exceed ±6.5°, as ICAO limits the error according to a bent (slow variation) tolerance of ±3.5° relative to theoretical course and to a superimposed scalloping (rapid variation) of ±3°.
On this aspect, the DVOR is also superior, the variable signal is FM modulated and less subject to multipath errors. The smaller CVOR is difficult to run on airfields due to reflections on close obstacles, the large DVOR must be used for terminal VOR. The CVOR is rather used for enroute navigation and implanted on isolated sites.
Large counterpoises under the antennas limit the reflection problem, and when the site used to built the VOR is carefully selected, the operational accuracy can be far better than the ICAO requirements in a large part of the service volume.
Both types use the following principle:
All receivers receive the same reference signal created in the VOR. This signal follows a sine function of the time, its frequency is 30 Hz. Where the cycle starts doesn’t matter, the only requirement is the reference must be equal to the variable signal when the receiver must indicate a north bearing.
The VOR receiver also receives a variable signal, a signal which value is determined by its own position relative to the VOR station. This signal is also a 30 Hz sine curve. The receiver works by phase comparison between the reference and the variable signal. A phase is an angle. The phase difference is equal to the bearing of the receiver (no difference therefore indicates north of the VOR station).
The variable signal is produced by the position of the receiver relative to the station. The two types of VOR differ by how this signal is created.
A first method is to use space modulation, an esoteric name referring to how two signals originating from separate antennas combine in space (actually how their modulation indexes are altered).
Two HF carriers are individually AM modulated by 30 Hz sine waves derived from the reference, one represents the sine of the reference angle, the other represents the cosine. The two HF signals are sent at right angle, in horizontal polarization, creating two individual figure-of-eight radiation patterns, themselves equivalent to a larger composite figure-of-eight along the symmetry axis:
Figure of eight obtained in the CVOR
As the AM modulations are linked to the Cartesian coordinates of the reference angle, the final figure-of-eight rotates at 30 revs per second.
The reference information is FM embedded into a 9.96 kHz subcarrier (332x30, sometimes called the 10 k subcarrier), which AM modulates the VOR carrier.
The variable signal carrier and the reference carrier interfere in a way which reinforces the field on one side and weakens it on the other side. This creates a directional cardioid, rotating at the frequency of the figure of eight pattern, 30 Hz:
Cardoid pattern in the CVOR
The cardioid EM field which constantly increases and decreases in strength at a given point in space as the pattern rotates is sensed by the receiver as a signal modulated in amplitude at 30 Hz, this amplitude variation becomes the variable signal. It has the same frequency than the reference but is lagging the time required for the cardioid to be steered from north to the direction of the receiver. As the system is calibrated to have a null difference of phase at north, the phase difference is constantly representative of the bearing:
Bearing determination in the VOR
The method used to create the cardioid changed with time. Initially two perpendicular pairs of Alford loop antennas were used to send the AM individual signals, and the reference sent using the four antennas. A vertical radiation creates bearings errors when the aircraft is banked, the Alford loop antenna (also found in the DVOR) was prevented to radiate vertically by adding a system above the horizontal radiators, giving the radome its pointed shape. Today a 4-slot cylinder antenna is used, its tall vertical shape is also due to the polarization filter.
Whatever the technology used, the signals sent from the antennas have the same characteristics. There is a carrier and two mirrored sidebands containing the AM modulation and the FM subcarrier:
CVOR signal spectrum
Space modulation acts on the amplitude to create the cardioid (complex sum of the respective phasors), it cannot generate FM. The Doppler VOR was designed to move to FM and improve accuracy.
The DVOR is entirely compatible with the CVOR, there is no specific DVOR receiver. The Doppler effect creates the FM variable signal. Physically a DVOR is a large circular array of about 50 Alford loop antennas, but only two diametrically opposed ones are used at a given time, and the system switches to the next pair counter-clockwise, scanning the whole array 30 times per second.
These antennas transmit a constant information-less HF signal (no modulation) on two different frequencies close to the VOR frequency ($\small f - 9.96$ kHz and $\small f + 9.96$ kHz). Albeit the VOR doesn’t send any information, the receiver sees the constant frequencies from the "moving" antennas with the Doppler shifts added. All receivers on the same bearing will see the same Doppler shifts, and the shift timing will vary only with their bearing:
Principle of the Doppler VOR: The variable signal generation
To maintain compatibility with the CVOR FM signal the Doppler shift must be 480 Hz (16x30), repeating 30 times per second. For 110 MHz, the required relative velocity is 1,300 m/s, it can be obtained at 30 revs per second with an array diameter of about 14 m.
The receiver receives three signals from three antennas but treat them as being sent by a single antenna. As the composite spectrum is identical to a CVOR spectrum, the CVOR receiver is able to continue working.
DVOR frequency spectrum
However the reference and variable signal modulation are now swapped, the computed phase difference is actually opposed. By inverting the scan direction (counter-clockwise in the DVOR), the result of the phase comparison is associated with the opposed sign (complementary) angle, which is the correct bearing.
The reference is still sent in all directions by the antenna at the center of the array along with the carrier. As in the CVOR, the carrier is used in the receiver to compensate for the Doppler shift created by the aircraft own velocity, and only the Doppler shift resulting from the array scan is taken into account.
Answer to specific questions
Do they generate the same navigation signal?
There are two differences:
The CVOR actually sends two modulated signal to generates a cardioid interference pattern which when steered creates an apparent AM modulation. The DVOR sends only “moving” unmodulated sidebands which create by Doppler effect an apparent FM modulation.
To maintain compatibility with CVOR, the DVOR reference signal must be sent in AM, and the scan done in reverse direction of the cardioid rotation.
Can they be used by the same receiver, or do VOR receivers need to be specially adapted to DVORs?
The CVOR receiver is used for DVOR. It computes a phase difference which is wrong (opposed cosine), but because the array is scanned counter-clockwise, the correct bearing is infered.
Is one more accurate than the other?
Wikipedia states, but with no reference, and I’m assuming it’s for a DVOR: “The predicted accuracy of the VOR system is ±1.4°. However, test data indicates that 99.94% of the time a VOR system has less than ±0.35° of error”.
The same article states about ICAO Annex 10 volume 1: “This document sets the worst case bearing accuracy performance on a Conventional VOR (CVOR) to be ±4°. A Doppler VOR (DVOR) is required to be ±1°”. But I can’t find these figures in the version I have.
Annex 10 and Doc 8071 volume 1 (testing procedures) rather give three tolerances:
- A maximum VOR equipment signal error of ±2°
- A radial bend tolerance of ±3.5° from the perfect course.
- A radial scalloping error of ±3° around the bend.
The two last errors are due to the environment rather than to the instrument. VOR are located at places where these errors can be minimized.
So the acceptable total deviation of ±6.5°, as attested by Doc 8071 volume 1 §2.3.48, when the radial is bent by ±3.5° and a scalloping of ±3° is superimposed. However scalloping is a rapid oscillation around the course, and in most cases is ignored by the pilot who keeps working on a mean deviation value (but there are limitations of scalloping when the autopilot is used, because the A/P can react to oscillations).
ICAO VOR tolerances on bearing value
The DVOR was designed to minimize the multipath effect, the variable signal is frequency modulated for this reason. The size of the counterpoise was increased so that reflection occurs on a controlled surface, rather than randomly. This way VOR could be located on airfields, allowing better VOR approaches and easier maintenance.
However DVOR introduces a new error: The array antennas are close to each other, therefore the inactive antennas on each side of the active one interact on the signal. First they delay the wave, introducing an error in the phase before the wave is Doppler-shifted. Second they absorb a large portion of the energy, reducing the range. As explained, the 14 m diameter can't be changed to space further the antennas, and they cannot be spaced vertically without impacting the maximum elevation of the signals and increasing the cone of silence, which is already relatively large (100°).
Another error is the discretization of the bearing information due to a scan over only 46-50 antennas (the matter is complex), a fix is the so-called power blending: The two or four antennas around the active antenna are fed with a smaller power HF signal and the transition from one pair to the next is smoothed.
The last error, which has been partially corrected by the DSB DVOR (the one with two sideband antennas described here) is related to the portion of counterpoise between the antenna and the receiver which varies with the antenna location. If a reflection occurs on the ground, and the one from the reference antenna occurs on the counterpoise, then their path is not equal and the phase difference is altered (however the modulation wavelength being 10,000 km, a path length error of 1 km is not very significant).
It's difficult to quantify a mean error, it really depends on the conditions the VOR is used and at which elevation angle the receiver operates. There are tools for simulation like the Ohio University Navaid Performance Prediction Model (OUNPPM) you may already know. Here is a study of the CVOR/DVOR accuracy with this tool.
Is one more susceptible to interference than the other?
For interference from signals or from obstacles, the CVOR is more prone to errors because of the use of AM for the variable signal. In the DVOR the reference is AM modulated, but a receiver can maintain a local reference using a PLL oscillator synchronized with the VOR reference when the signal is good, and reject transient spurious values (it’s like maintaining a local time reference, synchronized from time to time with a master clock).
Here is a study of wind turbines interference on DVOR.
EUROCAE ED-52 provides a guidance for VOR siting.
For foreign signals, VOR are subject to a requirement from ICAO (Annnex 10, volume 1, §220.127.116.11): "The VOR receiving system shall provide adequate immunity to interference from two signal, third-order intermodulation products caused by VHF FM broadcast […]".