Hi Nicholas
some thoughts that might help!!!
keep percussion out of the rack - ie use a drum machine
for a 9u 84hp consider 3 or 4 voices as a decent target - and in that size case I find it hard to justify more than 1 type of sequencing - pick midi-cv, a dc coupled audio interface (cv direct from computer, eg expert sleepers es8 or 9), a traditional sequencer (metropolis or the like) or generative (something like bloom or marbles) - if you must have more than 1 then consider 1 of them being outside the rack - a beatstep pro for example, or an sq1, or cv.ocd
in 252hp you have enough space to ignore small modules as much as possible and go for ergonomics over trying to cram stuff in
a single voice may be made up of some vcos - if more than one (or more than 1 output on a single vco) then a sub mixer will probably also be needed, some audio modifiers (filters, delays, reverbs, wavefolders etc etc) and the modulation and utility modules that are needed to support those modules
NB you can also get some modules that are complete voices - for example plaits can be used as a complete voice as it has an internal lpg (combination of vca and filter) or you can choose to bypass those and use external ones
then there are the modules that are needed to support these voices together - more mixers and other utility modules - possibly more effects - end of chain reverb is nice as a minimum to bring everything together
vcas are probably one of the most important types of utility module - get more than you think you need - make sure you have both linear and exponential (possibly in the same module) - linear is good for cv, exponential is good for audio - and audio can easily eat up 2 vcas per voice - one to open and close it and one to increase gain over time, perhaps - modulation can eat up a lot of vcas too - say you decide to use 2 modulation sources - you may want to modulate both of these modulation sources so that they do not happen all the time - another 2 vcas are needed for that - want to patch an auto-panner so that a mono signal is panned across the stereo field over time - mult the audio signal to 2 vcas (one outputs to L, one outputs to R) and use an lfo and an inversion of that lfo to open and close them - vcas can also be used with other utility modules to patch compressors
matrix mixers are a great way of expanding modulation sources - take mults of modulation sources and patch to modulation mixer, you now have 4 more related modulation sources (that can be mixed to your taste) - I'm in the use fewer modulation sources and more utilities to get more, complex modulation camp, as opposed to the more modulation sources and less utilities for the same thing camp - modulation and utility should take up about 50% of the rack (if you want a balanced rack imo)
a good starter set is - a sound source (a vco), a sound modifier (a filter or delay), a modulation source (I really like maths as a starter modulation source - as it is a great learning tool - see illustrated manual for more details), a way to play (see above) and a way to listen (I would use a quad cascading vca for this - and whatever adapter cables you need to take to the output from that to whatever you want to listen with - to a mixer, headphones, audio interface etc etc basically whatever you have already!) - and a basic starter set of utilities - I like links, kinks and shades as a relatively inexpensive and fully functional start - although any combination of modules with their functions will do nicely
I'd recommend starting slowly and then growing organically from there - ie some of what you want/some of what you need and only getting new modules after learning what you already have - experiment with and reesearch the modules you have and are thinking of buying - maybe a second filter before a second vco - because some filters can self-oscillate and so be used as a vco
after that I would look to adding a second voice (you can integrate one piece at time easily using mults and mixers)
at this point I strongly recommend getting a tuner - either on your phone or a guitar pedal or in rack - so that the 2 voices can be in tune with each other - instead of discordant - unless that is what you want - you may want this earlier if for example you are playing with other instruments - or have perfect pitch and find out of tune notes jarring
and then an end of chain mixer (possibly out of the rack) and some more effects - before adding any more sound generation options
I would also consider buying the book Patch and Tweak - as it is a good starting resource for learning about modular (and modules)
"some of the best base-level info to remember can be found in Jim's sigfile" @Lugia
Utility modules are the dull polish that makes the shiny modules actually shine!!!
sound sources < sound modifiers < modulation sources < utilities