I think the end result is quite good and I managed to get the registration right first time.

The front panel has been fixed in place and the shaft extension rods fitted to the attenuation pots. In terms of pure construction and finishing, the case is now completed and I only have the circuit to finalise.

I’ve spent a few days swapping op amps in and out. The crossover still has no cover fitted and is located out into the room, so access/swapping is simple. I power down the unit and pull the op amps out of the sockets with a removal tool and clip the new ones straight in. I use a suitable grounding wrist band connected to the chassis, just in case I pick up any static from the carpet. I have all the things needed, including the packets of various op amps, laid out ready next to the unit, so swapping a pair takes under a minute.

Op Amps tried (comments in brackets are how they have been reviewed by others)

OPA 2134 (Full rich bass if not the deepest, soft and warm mid/treble)
LME 49720 (Transparent, good mids not bad all round)
LM 4562 (Transparent, good mids not bad all round)
AD 8066 (Good deep bass and clean detailed top end but not as good for mids)

Low Pass
I tried the AD8066 in the LP filter and while I can agree it gives a deep and transparent bottom end, it seems to affect only a narrow band and has no impact on upper bass/lower mids. To me the OPA2134 is almost as deep but is slightly fatter and richer and has a broader influence on upper bass/lower mids which I liked.
Over all I preferred the OPA2134 for bass compared to any of the others, drum solos seem more realistic and the richness that carries over into the mids works well on acoustic instruments.

Band Pass
(LP element <6090Hz passing) The LM4562 worked best, it is very transparent and clean sounding for the interface with the tweeter.
(HP element >501Hz passing) The LME49720 worked best here. I know some say the LME49720/LM4562 are the same chip renumbered, and the spec sheets are identical. I am sure they are fully interchangeable in all circuits, but I find as others do, that they sound different. The LM4562 is slightly cooler and sharper sounding, so for the interface with the rich bass the LME49720 and it’s slightly warmer presentation Dove tail better.

High Pass
The AD8066 does work very well in this filter, cymbals and triangle are very clean with a long decay and a nice shimmer on cymbals. I also found that the upper range in Piano, located more in the upper mid region, sounded more percussive and realistic using this op amp, compared to the others.

Input Buffer
This seems to have an overall effect across the range. The LM4562 was very transparent, the AD8066 even more so, but upper mids/ treble became too dominant and tonally un-realistic. The slightly warmer LME49720 worked very well here also and gave the best tonal balance when used with the choices made for the individual filters.

It’s worth pointing out that changing the fitted op amps, with similar dual channel, unity gain stable devices, does not affect the cross over points in this circuit, it only effect’s things like transparency, timing, soundstage and tonal colour. So similar to valve rolling, you can tune to your own preference, but also like with valve rolling, there is a madness that can set in when searching for that illusive ultimate combination.

The combination I have settled on for now, is very good and suits my preference for realism over musicality, but it is never the less, still very easy on the ear, so the unit can be fastened up and I will live with it for a while.