Edit: The FM Ogre is covered up because bugs. But trust me it's there.

Hello all. I've revised my build since I last sought advice, and I have several questions. Also I'm looking for any general advice from experienced builders. This is my first (goal) build and plan to actually start it soon. Please guide or correct my thoughts and provide any input that might help! My background is that I am someone who has played with this stuff in software for at least 10 years, but got my first taste with the Neutron about a year ago (and have a list of limitations I frequently run into when using it, especially lack of attenuverters, triggerable EGs, lack of end of cycle triggers on EGs, and needing a third oscillator). I also own a handful of volcas but using a bunch of separate sequencers at once is unmanageable and my musical ideas tend to come in phrases of at least 2 bars in 16th notes or 16th note triplets. FWIW all the spontaneity and live recording functions of the Volca line are sucked dry as soon as you want a sequence longer than 16 steps. I also have played the cello and bass guitar from a young age so I'm not intimidated by music theory and like to have chord progression stuff possible in my workflow. In addition I'm very curious about plugging my bass and cello w/pickup into a modular synth. I studied physics and love learning about how instruments work and how to imitate them. At the same time purely electronic and imaginary sounds are amazing and I want to explore both at the same time. So here goes:

Goal: Primarily a flexible sound design lab with 3 operator FM, TZFM, and sequencing/sampling capabilities (keystep is a nightmare to program) so that I can minimize interaction with a computer. When I finish that amazing flute patch I worked so hard on I want to sample the instrument for later use. However I also have fun strong-arming the Neutron into generative patches and have gotten all kinds of contemplative sounding atmospheres, accidentally discovered the concept of filter pinging, and somehow gotten 5 part percussive rhythms plus a bassline out of just its 2 oscillators, an lfo, and a single filter. That type of exploration is extremely rewarding, so having exploratory generative tools (with a steering wheel, something like a very, very small Colin Bendersesque setup) is also important. I like the basic building blocks approach (rings is a compromise, more on that later, and damn would I die for a Trident). I like to be able to create the effects I want out of small parts so that I understand them before deciding on what compromises I'm willing to make when offloading them to another person's subroutine/module.

External hardware:
Behringer Neutron
Keystep
Behringer Xenyx 1202FX
Volcas/Bass Guitar/Cello
Mio2 MIDI router/filter/mapper

So that's the goal. On to specific questions:

Oh, but first, modules covered by the neutron so there is no confusion: 2 oscillators with soft sync (only to eachother, but there's a hardsync mod), pwm, and log FM, 1 resonant self oscillating filter with 2 outputs (Lpf+hpf, hpf+bpf, bpf+lpf), 1 clock syncable audio rate lfo, noise, 2 vcas, 1 attenuator, 2 linear ADSRs (no cv inputs except for gate, non loopable), overdrive w/tilt filter, grungy delay (input and rate cv controls, not clock syncable), S&H + hardwired slew (clock syncable, audio rate, used mostly as a bitcrusher), 1 slew (either linear or log not sure), 1 slew hardwired to pitch, 1 audio in that does no boosting of line level inputs but allows for self-feedback patches.

Ok. Questions.
1. Filter recommendations:
Neutron allows for notch behavior by combining its two filter outs. A second notch filter would be very nice, working towards the 3 notch/bp filter requirement for recognizable vowel formants which I could then control with CV. Also more notch filters = better phaser effects. The two in the build here are just placeholders, though I do like the way ripples sounds from the youtube videos I've watched. Smoother than Neutron and, like Neutron, no bass drop on Q increase.

1.5. Matrix mixer thoughts:
The doepfer 4x4 was my first choice because another thing I very much lack with the neutron is offsets and attenuvertors and it does this. But the main reason is for routing an inverted envelope follower from a kick drum to several sources at varying levels of attenuation. I know there's also a lot of other magic that can be done with matrix mixers and that is something I plan to learn. But the doepfer is big, the intellijel 1u quadratt or a shades gives offsets and attenuversion, so I think the smaller non-offsetting non-inverting 4x3 mixer is fine for now. Another alternative is that there are a couple of 4x2 mixers that can be chained into eachother later on, but I think i'd rather have the 3 available right away and then just trade up once I know more about the different things I want it for.

Case thoughts:
I originally wanted a Behringer Go because it's cheap, big, and has lots of power. However I will need to travel with my rig, so it needs to fit US carry-on specs. That turned me to the rackbrute 6u, but I'd still need to drop $100+ on a hard case for it, and it doesn't have enough power. Another power module and I would have to give up a module. Alternatively I could buy 2 rackbrute 6us and a soft case, and plug into the second case's power supply, but i'd still need another power module in the future. That plus the travel case costs make me think jumping straight into an intelligel 7u is the better way to go. It is ready to travel out of the gate and has more than enough power. Also I was able to recoup over $100 of the price difference by replacing modules with 1u versions. Accounting for the above issues, cost per HP is as follows:
2 rackbrute 6u (171hp2) + soft case = (600 + 30)/(171x2) = $1.84/hp (not accounting for a future second power module so probably not much better than the 1 rackbrute solution below)
1 rackbrute 6u + hard case = (300 + 100)/(171) (IF IT CAN HANDLE 1500mA) = $2.34/hp
1 7u intellijel (assuming the top row = ~5 modules of saved space) + no extra case = 650/~229 = ~$2.84/hp
1 6u Mantis case: I hate the way it looks. I think it looks like a Fischer Price toy. I'm as utilitarian as they come, but this is an instrument and if looking at it kills your inspiration then in this case (no pun intended) style is equivalent to utility.

Option 2 probably wont even work and gives me zero space to expand, option 1 gives the most space to expand but introduces so much uncertainty (will it fit in the case I buy? What sort of foam will I fill the case with?), option 3 isnt THAT much more inefficient than the other two and is easiest to travel with, while still leaving me room for a few more modules. It is also linkable to another case in the future.

Seq + Sampler thoughts:
In addition to what I said in my intro about needing seq/sample functionality, I love euclidean rhythms. The neutron can actually do pseudo euclidean rhythms with both oscs cranked way down as pulse waves and synced. One gives the main clock and the other gives evenly spaced but not quite perfect divisions of that clock and then resets each main cycle. Add in pwm and delay for more rhythmic coolness. Problem is now I have no more oscillators, so a dedicated module for clock play is very much desired. Hence Hermod and Marbles. Hermod is kind of a bang for buck beast here, as it handles midi connectivity with my Keystep and DJTechTools 4x4 continuous rotating potentiometer thing. And it can play with my Volcas if for some reason I decide not to sell them to help fund this. I love using Seq24/32/48/64/66 on the computer, and to a more limited extent Caustic 3 on android. Pattern based is what I jell with (pun somewhat intended). The 1010 musicbox stuff has that workflow as does the Hermod (though with a different interface).
Current choice: Hermod + Bitbox micro
Alternatives: Blackbox + Midi interface (poly 2): ~$1000. blackbox not triggerable by cv, and no guarantee auto multisampling will be implemented. Speaking of midi solutions. An 8 channel midi interface runs about $400 in EU. Add $100 to that and you end up with a Hermod, which does sequencing, midi, and prevents me from needing a dedicated euclidean module. I'd need to hear a very compelling argument to decide to NOT just get the Hermod and call it done.
Hermod + Bitbox Mk2: 16 samples would be nice, but it didn't have velocity multisampling last I checked and that is a non-starter for me. It's also bigger and unless i prove to myself i NEED more than 8 pads it seems like a waste to purchase it. Micro has more outs too, and I don't care about synthbox/fxbox firmwares.
Pyramid + Midi + Bitbox micro: Just no. $$$$$$$$$$
Still researching: Squid Salmpl, Vector, Eloquencer.

Getting late will finish tomorrow. In the meantime thankyou for any tips or thoughts you have!