Ah.... I realllly feel like you did this before in like 2010? But it's still good. Much has changed in the interim on drawing small objects I've noticed.
I know I drew a series of sort of freehand SCB-227-inspired designs back in the long, long ago, with SPG-59. I don't remember if we ever worked through MPQ-53 refits.
MPQ-53 as originally built could track 50 targets, control 9 missiles at a time against (sources vary) up to nine targets, and control 3 of those missiles in terminal CWI homing mode FYI. Operator limits didn't matter, it was capable of fully automatic modes and based on the video of the monochromatic displays I've seen you were pretty much hitting airliner destroying territory long before you had 50 tracks on screen. It didn't help that the actual firing sector was a V-shape, so whatever form factor your display has, it's just gonna be that Vee that actually matters for data with a 1970s TV screen!
And we still use this!
The big limitation here is just time sharing on the radar, the separate TVM uplink antenna helps with that, but the duty cycle on this radar was like 10% as built, and your dividing that over and over again.
My thinking was that it seems like TVM in that era ought to be at least somewhat computationally limited, not just array limited. At least in midcourse (maybe not in terminal), you need to be running launcher-side calculations to tell the missile where to go, but your track accuracy might be sufficient that the radar isn't any more loaded in midcourse (per target) than it is in tracking.
Shipboard you could probably track more targets at some kind of precision, but actually engaging more targets at once would be very hard to brute force. Being a C-band radar you aren't scanning huge areas of sky with each beam you emit. Making the antenna bigger doesn't help with that much at all, scanning for new targets still involves a whole lot of duty cycle and so does tracking existing ones. A more powerful radar can generate more total power and have more effective range, but that doesn't mean out of hand that it can actually engage more targets with PESA. But perhaps a warship could have had two traveling wave tubes from the get go? I mean I'm sure on paper it could!
Well, right off the top, we're giving up on MPQ-53 handling long-range search, although it would need to perform some sort of target localization based on the coarse data fed from SPS-49. It seems like this ought to help quite a bit. In a casual, hand-waving sort of sense, C-band sucks for search anyway, and so I'd figure we'd recover more track/engage capability by giving up search than (say) doing the opposite. Whatever time-sharing of the array we're using for long-range search, it's not very efficient use of that time.
Shifting from SPS-49 to SPS-48 ought to give further improvements by this metric, although I'm reluctant to give up L-band on a long-range platform like this.
An alternative path, which I'd definitely considered, is retaining at least one trainable director (the bow, presumably) and duplicate real-life-scale MPQ-53 in that location for enhanced sector capability, i.e., eighteen missiles in the air over something like a 300 degree field of fire.
The maximum approach might be to retain all three trainable directors (replaced with MPQ-53), keep SPS-49 aft, fit SPS-48 forward, and fit only the three fixed MPQ-53 arrays on the main deckhouse. With all arrays operating, this would give 360deg coverage with at least two arrays, and three-array coverage across broad arcs abeam. The result certainly would not be aesthetically graceful, but that's not why we build warships.