To be honest, in order to leverage technology as it existed at the time, I'd think about abandoning the AEGIS-feeling framework entirely. I might instead try to fit, say, six to eight X-band PESAs from MiG-31 technology, on trainable mounts. As many as I can, anyway.
(snip)
EDIT: Best of all, if you have one of your radar units crap out on you, as Soviet electronics tech was wont to do, you can cover the blind arc with a different PESA tracker/illuminator. And if you have to put to sea with some radar units missing entirely, as again their ships often did, the combat system is degraded rather than dysfunctional.
EDIT2: Doctrinally, this is also a good fit for the Soviet/Russian use case. Ticos had to worry about SSGNs and Backfires and whatnot enveloping the carrier group from all directions. Kirovs and Slavas, the pinnacle of their era, were apparently relatively content with sector defense, with the reasonable expectation that all severe attacks would come from a carrier on a single vector. So I think this really meshes well with the fit and feeling of procurement officers who would have come up treating SA-N-6 as gospel.
Well, yes and no. The Kirov/Slava paradigm might be limited by 1) a typically 80s environment with a low density of targets, particularly AShMs, 2) an operational concept implying, as you said, a mostly uni-directional threat, and 3) fairly bulky AA systems (chiefly the Top Dome and Tomb Stone missile direction radars of the S-300). You can see the limits of that concept in the fact that both classes had omni-directional close-range air defenses (which kinda negates the uni-directional threat argument), and in the numerous Russian voices expressing disappointment in the Slava class for its lack of long-range SAM coverage on more than one azimuth. The weakness was known for a long time, but was considered a trade-off for a lighter and cheaper ship.
In any case, the Sovremenniy doesn't have that problem, seeing how it bristles with illuminators in the first place. With 2 missiles rails, a reload time of 12s on each and a range of 25km on the initial version (about 30s flight time), they still saw it necessary to fit 6 Front Dome missile illuminators (plus 4 optical backup). Redundant sector coverage is ensured, so in your analysis, it makes sense to replace the Front Domes with slewable scanning arrays one-for-one.
The type of slewable targeting array is mostly what I had in mind for the later SAM generation replacing the Uragan (SA-N-7). Considering the Soviet (justified) obsession with backwards compatibility, the phased-array directors for that new missile should be able to handle legacy Uragans, so that they can be retrofitted on Sovremenniy hulls in MLU and the new ships can sail out with previous-generation missiles until enough of the new ones come down the pike.
Come to think of it, this kind of slewable targeting array would pretty much be an enlarged version of the 5P10 Puma (the oddly proportioned box above the bridge, which you also see doing gun FC on the Talwar and Steregushchiy classes), which, I was surprised to learn, is a 64x64 AESA. This IRL version probably requires even more smarts to handle multiple gun- and rocket-type weapon than our one-trick pony of an illuminator. Come to
really think of it, who says this Zaslon-based array of yours can't fit inside the Front Dome cover, thereby rendering the new iluminator visually identical to the original, and making the entire point irrelevant to Shipbucket?
Now, as much as I like the idea above, I think it's more appropriate as an upgrade for legacy Sovremenniy hulls, possibly alongside a new SAM system.
I don't know if I mentioned it already, but the PAR-equipped Pr.956.2 I've been doing above is as much a block upgrade as an evaluation/development platform for the new PESA fixed array that will enter service on the second batch of Pr.1244 frigates and the future 10000ton destroyer.
Besides, though I don't have sources confirming it, the abundance of redundant mechanical scan arrays even on 90s projects leads me to consider that the goal of the initial phased arrays was thought less in terms of target designation than raised scan rate. I mean, why else do you put
twice the same 3D radar on the same ship? And don't talk about reliability when that same radar is built into all manner of ships for 20 years, at a rate of
one per ship. At that stage, if you don't have a half-way civilized MTBF, just scrap the damn thing and start over.
So, yeah, the target (ca. 2015) configuration for such a ship would probably be a fairly smart fixed-plate AESA, if weight allows in more than 1 frequency band, backed up by an assortment of dumber PESA slewable arrays for sector scan and illumination. A fall-back solution is to switch to a dumb PESA scanning array and smarter AESA "heads", which is the current IRL version.
(I note, FYI, that the OE-120 was in USN service in the early 1980s. 2D round phased arrays are apparently drop-dead simple.)
Not to be overly contarian, but I'm wondering if this isn't a case of the bottleneck being data treatment (computers) rather than antenna (radar) as such. Again, Ive seen similar things in the Russian industry nowadays, so why not. I'll have to figure out how to introduce this properly, or turn it into its own thing.