From a design perspective the straight sections were much simpler than the bell.  Each section was divided into two identical halves so I just needed a single set of toolpaths for each section.  The biggest challenge with machining them on the CNC is that they are 28" to 32" long and my CNC has only about 24x24" machinable real estate.  This meant that I had to break each toolpath into 2 parts.  Vcarve make this pretty straight forward with tiling.   In my case, I can slide the work blank through the CNC in the Y axis and Vcarve will break the tool paths into pieces for each tile.  Accurate positioning is necessary so I had the CNC drill positioning holes into the spoilboard and the blanks that allow me to put the piece in the correct position as I moved it forward and back for different machining operations. 

Here you can see the beginning of the roughing toolpath on the first tile of the largest section.  In the foreground you can see the alignment holes.  There are 2 more 18" down the length and there are identical holes on the other side which make sure that it is positioned in the same place when I flip it over to machine the outside face.   I later added additional holes to make the positioning more rigid as I moved the piece to machine the other end. 

After finishing tile 1 on the inside, The workpiece is moved to machine the second tile.

Because of the thickness of the workpiece,  my clamps are not tall enough to hold it down so work-holding was a little sketchy. Fortunately the dowels in the spoilboard help keep it in place laterally but I had to pause the machining from time to time to make sure that the screw clamps and hold downs were still tight.    For subsequent sections, I used a combination plane to cut a groove into both  long edges  which made it easier and more secure to hold the work piece down. 

After the inside, flip it over and do it again (this is actually the other half of the first section after I added clamping wings) . 

I did not make a profile pass on this one to cut it out.  The finishing toolpath left a paper thin layer that I could simply push a flush cut saw through to cut off the tabs. 

Just holding the pieces together after freeing from the blank to make sure that they actually match.  You can see I had some tearout on the end that I fixed before gluing the halves together.   

Then it was just a matter of repeating the process 7 more times.  Each half section probably took about 2.5 hours of setup and machining, not counting the prepping the blanks.  Blank prep entailed ripping the 2x12s to width to match the thickness of the blank, planing, glue up, final planing to thickness and then cutting the grooves for the clamping wings 

Before glue up I did some basic sanding to remove some of the tooling marks, mostly on the sides where the finishing pass is mostly cutting with the side of the bit.  You can see some of that on the left side in the above image.  On the 3 smaller sections, I decided to apply a light finish of Tried and true varnish oil to add a little spit resistance.  I will have to figure out how to finish the inside of the largest section. 

Clamping was again done with giant rubber bands but in later sections, I also added some clamps for a little extra hold.  I used liquid hide glue to join the halves. 

Once everything was glued up, I used a block plane to clean up the tabs and glue joints as well as some machining marks that were going to take too long to sand away.  The shavings are from the 3 previous segments.

I then mounted each one on the lathe for sanding.  The lathe was just a convenient platform for sanding.  I did a little sanding with the lathe turning but I mostly just sanded lengthwise by hand  I also squared up the rounded corners on the ends where coupling sleeves will go. I was a little afraid I would destroy them in the last step, though I may still have some later steps that could destroy a section.  This is the smallest section just as the one above.  I designed them so that they would just barely fit on my lathe (actually planned ahead for once). I was pleasantly surprised that they were mostly round when when I turned on the lathe.

All four segments and the bell smoothed and sanded.

BTW, when I was buying the last two 2x12s at Lowe's, I realized that this is actually Douglas Fir, not SYP.  They have one stack of Douglas Fir right next too the SYP and because it looks so good is why I bought it thinking it was SYP.   Other than 4x4s it has been years since I have seen Doug Fir down here.   I did use one piece of SYP before I realized that the rest of the wood was DF, which stands out like a soar thumb on the bell.  

Coming up next is figuring out how to actually join the segments in a way that they can be disassembled for transport and still strong enough to not break apart.  I probably should have planned that out before now but I am still playing around with ideas.