What are your crossover points?
I’m assuming you’re about 1.2k to 8k with the S2 and the Raal above? If so then you probably want the more rapid horn expansion for the 600Hz horn, which will reduce beaming higher up.
So I’m guessing your lower mid range is 250 or 300 to the 1.2k? No small solution there!
Are you sure you don’t just want to get a whole MEH thing to shove with the Labs? Like one of those 80s mini systems!
Well the second one is finally completed. My speakers now look like this:
Now listening, they sound great (tbh they did before, but hey). I’ll measure them when I have recovered from 2 months without music!
Now you just need to do the bass horn.. 
They will look good once you’ve got around to painting the square horn white.
I know….all in good time……they’ve only been like that for 16 years 
I might print a round-to-square transition for the throat of the mid bass, so they’re not finished yet.
Also I know that I won’t ever bother to do this, as it won’t make a scrap of difference.
Therefore I will never paint them.
what xover are you using Adam? Nadja?
I’ve been playing with this recently; I stopped when I decided to print new horns as I will need to remeasure.
Basically I’m using rePhase to generate the XO, and JRiver to run the convolver. There’s a post here by the guy who wrote rePhase:
I use the “reject low” crossover. It’s an asymmetric XO that sums flat. The idea is that my horn-loaded won’t like being played low, but should be relatively well behaved higher up. So this seemed kinda perfect for a system like mine.
So I now have a new 3d printer - a Bambu Labs P1S. Worth the £479?
Bambu is very different in philosophy from Ender (and the likes of Prusa) in how they approach printers. It’s very much a complete integrated system; it’s Apple versus a hobbyist, tinkering system. After saving the model in Fusion it’s about 3 clicks to getting a completed print. Will take some getting used to.
Now accepting new design projects 
It’s a cool project for sure, although the pleasingly rounded shapes by JMLC are more my kind of thing
There are a few multicell models available on the on-shape directory I have be eyeing them but I think there 21 different cells in this assembly so lots of very well organised printing
With last few bits to my first printed set of horns now complete I thought it timely to add something to this thread. I read carefully as Adam made his horns, at the start I was wondering whether this would be too big a project for me to tackle? - There were lots of unknowns. One important lesson is learning how best to slice large items.
Example
My first 2 part (split axially) throat adaptor had lots of overhang meaning that lots of supports were needed (left view) = 22 hours printing.
The 3 piece version has a split mounting flange but a non split flare =11 hours printing - And yes its plenty strong.
Another thing I have been doing is counterboring(?) and pressing in nuts when needed - looks nicer and saves time
They look great.
I was told over and over that the best plan was to use super glue rather than bolts, so eventually I gave in and glued. It was a bad idea. Modelling in bolt points, as you have done, will give more accuracy and better joins.
Yes - It’s all a bit experimental- Not many people have made horns this size on a home sized printer and then made an advice video avail. The lug and bolts method does add lot of additional hours effort at the modelling stage but likely saves time in the end.
I did end up with some minor gaping - you can see it at the split in the image above
When I do my next one I will model in two additional lugs here to pull that gap in with bolt pressure. I will glue them eventually but will do some FR testing first.
The other issue with using dowels is the overall weight and mass of PLA used - what did yours weigh?
I don’t think that the dowels and bolts add much weight. Mine are about 2-3kg, I think - awkward to manoeuvre but not heavy. Doubling that wouldn’t really make much difference to how they handle.
I think that a better understanding of how to manage the slicer would be good for me. I’d like to have the main wall extra thick but the infill and opposite wall are just there for strength; standard slicer settings treat all external walls the same. I might experiment more if I choose to redo them, to get the strength and thickness where it’s needed.
No my weight comment was due to the extra wall thickness needed to fit dowels isn’t yours 10 or 12mm wall? - I ended up with a 8mm wall thickness for my throat adaptor and 5mm wall on all of the other the flares. That gave a 1.515 kg finished weight including the split flange. The centre of mass being a bit closer to the flange also reduces stress on the mount
That’s not how it works, it’s mostly infill which is only 15% PLA and the rest is air!
Scroll up and down in your slicer software and you’ll see how it’s made up. At no point is there anything like 10mm thick PLA.
