Stowed configuration. The Nebo SVU occupies three vehicles. A semitrailer carries the antenna/radar system, a 6x6 truck the diesel generators, and a 4x4 truck the operator cabin. The system can be deployed in 20 minutes, which is less than half of the time required for other Russian VHF radars (NNIIRT).

Commercial VHF band MOSFET transistors rated at 500W are now available in the global market at unit prices of around US$250.00, so building a VHF band TR module rated at 2 kiloWatts with four ganged MOSFETs presents no great difficulty, the only issue being effective cooling. With the low packaging density for a VHF AESA, it is clear that this did not present any  obstacle for Krylov and his designers. Western designers have been building kiloWatt class L-band TR modules for well over a decade.

Low noise solid state receivers for the VHF band are also a non issue, and the low packaging density requirement for such an array would give the designers considerable freedom in layout.


The turntable mounted antenna mast doubles up as a structural mounting for three major RF modules. These include modules containing phase shifters, controllable attenuators, summing networks, exciter stages and power feeds (NNIIRT).

What remains to be disclosed is how NNIIRT designed the phase components for beamsteering control of the antenna, as the claimed module providing this function is quite compact. At such a low frequency a Digital RF Memory (DRFM) style solution might have been adopted, rather than a classical analogue delay line or phase shifter solution - if we assume a carrier frequency of the order of 150 to 220 MHz typically seen in Russian VHF radars, a fourfold sampling frequency of 600 - 880 MHz would be more than adequate to ensure high signal purity. This is again well within the reach of Russia's industry base. If the radar is limited to small beamsteering angles, primarily for angle measurement, the range of phase/delay increments required per element will not be as challenging as for a design having large off boresight beamsteering angle requirements.

The radiating antenna element design is a three element hybrid - a vertically polarised two wire 3/8 l folded dipole [Kraus 11-39, 11-61] with a single parasitic director, using additional support frame mounted reflector elements. The well documented dimensions of the Ural 4320 truck and good close up imagery allows a fairly accurate estimation of the wavelength at ~2 metres with a symmetrical ~1 metre array element spacing, ie a regular square lattice. The choice of a 3/8 l folded dipole was clearly driven by its compact size allowing tighter element spacing in the array. Gain is of the order of 3-4.5 dBi per element, but is likely to be reduced by array coupling effects.

The choice of vertical polarisation is unusual for a VHF design intended to track aerial targets, and is best explained by the dual role use of the radar for ballistic missile defence purposes, as the shape of ballistic missile targets presents a higher RCS in the vertical polarisation. The 1L119 array design with a regular element spacing has the capacity for growth to a selectable polarisation, with embedded mechanical drives to rotate each antenna element through 90° to select optimal polarisation for a given target detection regime. The principal penalty in the hardware is additional complexity per element, and the need for different processing optimisations for either polarisation. With an electrical motor drive in each element, the rotation and polarisation change could be effected in seconds.

The problem of element spacing versus grating lobe performance is interesting in this design. If we assume that the electronic beamsteering capability is used primarily for precision angle tracking of targets near the antenna boresight,  grating lobes do not impose quite the burden they do in fixed X-band AESAs, and there is some flexibility in operating frequencies.

If the electronic beamsteering capability is used for sector searches, with significant deflection angles off boresight, then grating lobes become a potential problem in the design, and the <1/2 l element spacing rule limits the upper frequency of the design to around 150 MHz, with degraded gain in the 3/8 l folded dipole imposing the lower limit on frequency agility. The range of measurement error in array geometry indicates that the design was sized for larger deflection angles, so ±45° to ±60° off boresight is achievable, subject to aperture foreshortening, sidelobe performance limits, and the shaping of the hybrid two wire 3/8 l folded dipole element mainlobe. Were the design limited to small off boresight steering angles, the element spacing would be greater [1].

With only 84 elements, the 1L119 uses a sparse array, in AESA terms, so highly accurate calibration of module phase/delay and gain are absolutely critical to achieving the intended sidelobe control and beamsteering accuracy for repeatable target angle measurements.

As the design is an AESA, digital control of angle/delay and amplitude per element is a given. This also presents considerable freedom of choice in taper (illumination) function across the array, for control of sidelobes. The absence of any auxiliary antennas as used in the 1L13 for sidelobe cancelling can be accepted as proof that the 1L119 uses amplitude control in its antenna channels. Not surprisingly, NNIIRT have not commented on the choice of taper function, only that the radar has 'adaptive sidelobe suppression'.


Polar plot for estimated azimuthal plane sidelobe / mainlobe performance using a Chebyshev taper, for a peak sidelobe level of -24 dB (Author).

The design may include active jammer nulling by notching the mainlobe, with at least one Russian translation appearing to claim this, but given the quality of so many technical Russian to English translations this could however be a misinterpretation. As an active array this capability could be integrated in the design and simply not disclosed, so from an analytical perspective the safer assumption is that this capability already exists or will exist in a future variant of the design.


NNIIRT chart showing detection range performance for a 'MiG-21 with RCS=2.5m²' target, in the absence and presence of a jamming signal with a power density of 100 Watts/MHz. Russian data on range performance is consistent, but cited RCS values for identical ranges vary between 1.0 and 2.5 m².



 

There is some inconsistency in cited error bounds for target tracking. This chart shows worst case performance (range error = 200 m; azimuth error = 0.5° and elevation error = 1.5°), with the best case range error at 100 metres and best case azimuthal error at 0.3°. This performance is of the same order as the S-band 64N6E family of PESAs used as SAM battery acquisition radars (Author).

[Chart - Under Construction - Mainlobe Squint for Angle Measurement]

The radar's cited angle measurement accuracies are 1.5° in elevation, 0.5° in azimuth, and range accuracy is 200 metres. This  performance is almost identical to the S-band 64N6E Big Bird PESA used as a target acquisition component in the S-300PMU-1/2 Favorit / SA-20 Gargoyle and S-400 Triumf / SA-21 Growler SAM/ABM systems.


Russian literature covering the 1L119 describes it as capable of detecting and tracking aircraft and ballistic missile class targets. Tthe antenna can be tilted at least 17° in elevation, the latter cited specifically for  'ballistic missile  acquisition'. Ballistic missile target detection is likely to have imposed the choice of vertical polarisation, less than favoured otherwise due to poor ground clutter rejection performance.

The antenna can also be rotated at 3, 6 or 12 RPM for aerial target acquisition, or pointed in a fixed direction to cover a specific threat sector. The cited altitude limit for search modes is 100 km, for sector tracking modes it is 180 km. Using a circular sweep pattern the antenna is claimed to be limited to an elevation angle of 25°, but in its fixed azimuth/sector target tracking mode the highest beam elevation angle can be as high as 45° to 50°. If we assume the design is mechanically limited to a tilt angle of 17° this suggests an electronic beam deflection angle in elevation of ±28° to 33°. It follows that a similar bound would apply to horizontal deflection angles, through commonality in delay/phase shifter hardware.

Clearly the design will retain all of the DMTI processing features introduced in the 1L13, and Russian literature clearly states that the radar has features for adaptive rejection of ground clutter, precipitation and chaff, including Doppler compensation. Clutter rejection is cited at 45 dB, chaff at 30 dB, and noise jammers at 24 dB, figures which likely understate actual performance. Given that the earlier 1L13 uses Barker codes, it is likely the 1L119 will use the same, or other similar low cross correlation codes.

A curious statement in a number of Russian documents is that the radar employs "Complete space-time digital signal processing". This may well be an attempt to explain that the radar is fully digital throughout, or it may be a poor translation of Space Time Adaptive Processing (STAP), only recently adopted in Western radar designs. Is STAP a feasible proposition for a Russian radar designed very recently?

Given this radar is a VHF design with a modest sampling rate, and Commercial Off The Shelf (COTS) computing power is not a problem for a design carried by ten tonne Ural 4320 trucks, then the only issue with STAP is the ability of the Russian designers to implement the algorithms required, and whether it is useful enough to justify the effort.

Most Western research on STAP is focussed on airborne radars using STAP to adaptively reject ground clutter. For a VHF DMTI the issue is rejection of ground clutter, but also other unwanted effects such as Doppler shifted chaff and weather. At this stage the issue of STAP capability in the 1L119 remains unresolved, but it is a likely capability in this family of radars longer term. There are no fundamental problems with dividing this array into multiple receive path phase centres, since cables are already routed from the TR modules to the central phase control, amplitude control and summing modules.

From an analytical perspective the safest assumption is that this capability already exists or will exist in a future variant of the design. Engineering a STAP capability for a fixed ground based DMTI will be easier than doing it for an airborne X-band radar.


The 1L119 Nebo SVU makes full use of modern COTS technology, with 17 inch LCD displays for operators (above), and flexible digital display presentation (below). Display formats can include geographical data such as borders, air defence zones, and range boundaries (NNIIRT).


The display system software for the operator consoles and interfacing to the array management processor (array control) was developed initially in the 2000 to 2002 timeframe, using COTS software and hardware, specifically Intel architecture, Linux and С/С++ high level languages, and Xlib, Xt, Xaw, Qt libraries/toolkits. This is the same basic technology used in state of the art US military equipment for this purpose. This also supports NNIIRT claims that the 1L119 is a fully digital system.


There are still many uncertainties remaining in understanding the full capabilities of the early production 1L119 Nebo SVU. For instance, the phase/delay control capability and thus maximum off boresight mainlobe angles in azimuth and elevation have not been disclosed. An optimistic estimate is that the design is limited by phase/delay control to modest off boresight angles, used for precision target angle measurement only. A more pessimistic estimate is that the design can achieve an off boresight elevation and azimuth deflection angle of ±45° to ±60° similar to existing US S-band and L-band AESA/PESAs, thus allowing it to perform agile fixed sector searches. While the more optimistic estimate still makes it a highly effective combat capability, the pessimistic estimate would make it exceptionally capable. The basic antenna array design, given sufficiently good engineering,  supports both regimes of operation. The safer assumption from an analytical perspective  is that the capability for larger electronic beamsteering angles already exists or will exist in a future variant of the 1L119 design.

Table 3 Best Case versus Worst Case Capabilities

Parameter	Best Case Capability	Worst Case (Design Potential / Bound)	Notes
Electronic Beamsteering Angles	±~30° azimuth / elevation	±45°- ±69° azimuth /  elevation
Electronic Beamsteering Interval	3.3 msec	0.4 msec
Jammer Suppression	as per 1L13	CRPA (adaptive mainlobe notching)
Clutter Processing Technique	DMTI only	Space Time Adaptive Processing	STAP
Angle Tracking	Slow Sequential Lobing	Fast Sequential Lobing
Deploy / Stow Time	20 min	5 min
Integration	Digital Track Output	48N6/9M96E midcourse guidance via 30N6E

Table 3 maps out the design potential of the 1L119 and the range of capabilities which exist in the current configuration of the design, and would be certain to appear over a 30 year design and production lifecycle.

The cited error box of the Nebo SVU is small enough to enable a SAM or AAM with an active or infrared seeker to be flown near enough to the target to acquire it and initiate terminal homing.


Engaging a VLO/LO target (above) using the 1L119 Nebo SVU as a VHF band acquisition and tracking radar, and the 30N6E Flap Lid as a missile uplink channel (Author).


S-300PMU-2 Favorit / SA-20 Gargoyle 5P85TE TEL in deployed configuration, with the telescoping network antenna retracted. The PLA remains the largest export customer for S-300PMU variants (Xinhua).

Whether the Nebo SVU is used as a cheaper substitute for an SA-10/20/21 S-band 64N6E radar or paired with a 64N6E,  the radar has the required performance cue an X-band 30N6E series engagement radar. If these systems are all networked following current Russian practice, the battery's 5N63 / 54K6E series command post can launch the missiles remotely and datalink them to the aimpoint through most of the flight trajectory. When near enough, the missile switches to its own terminal homing seeker to complete the engagement.

What the Russians have not disclosed, but is clearly obvious, is that pairing the Nebo SVU and 64N6E allows the operators to discriminate between a low observable and conventional radar target and adjust tactics accordingly. If the target is invisible on the decimetric band 64N6E but visible on the VHF band Nebo SVU, then clearly it is low observable, and a missile trajectory flown under datalink control using updates generated by the VHF radar is needed, rather than a conventional engagement sequence where the 30N6E locks up the target and completes the engagement autonomously. Missile range performance permitting, this opens up other options such as flying a 'dogleg' curved missile trajectory to effect a beam attack terminal phase, so the missile's seeker is illuminating the less stealthy beam aspect of the aircraft rather than its most stealthy front aspect.