As I mentioned in part one, I initially started with just 4 slices of the bell.  I decided to use some nice straight grained SPF 2x8 I found at  Lowes (pretty sure it was white fir) for my first attempt.  Since the blank needed to be more than 1.5" thick, I ripped strips whose  width were the target  thickness and laminated them together for the first prototype.  Since the depth of cut was over 2" deep, this required an extra long end mill and ball nose bits. Unfortunately, the fir turned out to be too brittle for my machining strategy and the tear out was pretty bad.  In hindsight, I probably could have adjusted my machining strategy to reduce the tear out but after switching to SYP things worked better.    I can also find much better quality SYP with a more consistent grain stategy down south than SPF. 

 

 

You can see that I decided to embellish the sides of the bell with some 3D clipart.  

After switching to SYP, there was still a little tear out to deal with but overall the results were much better.  I also switched from 4 to 6 slices at that point which also made things work better.  Here are some actions shots of the top and middle slices being machined.  

 
Middle slice

Top Slice.  Since this is a two-sided process, I used dowels in holes cut symmetrically on both side of the blank and  into the spoilboard to precisely position the piece when flipping to the other side.   

Still a little tear out but much more manageable. 

The two slides stack on top of each other

 

It was at this point that I decided to upgrade my CNC.   What I learned from this first attempt is that my fixed speed Porter Cable router on the CNC was not quite up to the task.  I needed more control of the RPM and I started having issues that sounded like the bearings were going out.  I've pushed this PC router pretty hard since I got it about 6 (?) years ago and I decided that a variable speed spindle was the way to go.  That caused me to go down a rabbit hole replacing the electronics on the CNC so that I would have machine control of the spindle speed.  Between that rabbit hole and some medical issues, that caused about a 2 month break in the action. I've blogged about some of my issues with my CNC and I may eventually write about  my experience upgrading my machine but we can skip ahead to stay on topic.

 

The 3rd slice took a little head scratching but I eventually just decided to take a similar approach to the top two slices.  I thought about orienting it on its side but that would have required a much thicker blank and wasted much more wood.    As it is, this approach, wastes wood on par with bowl turning.  Not only are you shaping hollows and rounds but because of the curved shapes end to end you wind up with cutoffs too small for much of anything.  Here are some shots of the making the nearly vertical slices.  

 

 

Then it was a matter of creating mirrored versions of each slice and repeating the process for the other side's slices. I realized early on that glue up was going to be challenging.  I decided to use hot hide glue because it allows you use a rub joint for the initial stick and light clamping to get a good bond.  Here are the two halves before gluing on the 3rd slices and breaking on of them in half. 

When gluing on the 3rd layer slices, I tried to use some cauls to apply uniform pressure but I got too much and split the bell in half down the middle.  Fortunately it was a clean break and I used HHG and a rub joint to successfully mend it.  

 

I sanded the inside of the bell to remove machining marks and slight gluing misalignment.  I should have looked at the bottom during glue up.  The misalignment caused some gaps that I filled with sanding dust mixed with liquid hide glue.  

 

I was finally ready to assemble the two sides of the bell together.  Again, I used HHG to get everything to stick together while I clamped it using giant rubber bands.  I had to sand the edges flat to get a tight joint.   Here is the dry fit.  

 

This was before filling the misalignment gaps. At the bottom you can see the repair with a rub joint from where I broke the piece in half by over clamping. 

 

I used many more bands during the actual glue up.  The nice thing about the bands is that they sort of automatically push everything into alignment so very little tweaking was required once the first bands were in place.  

 This was a little stressful and difficult with only two hands but I knew that the hide glue would allow corrections later if needed.  

 

As you can see below, I had an alignment issue when I glued on the 3rd layer, was well as some tear out.  This is where using hide glue pays off.  I whipped out my steam generator (wallpaper remover) to soften the glue and then applied clamps to push things into closer alignment and clamp it back together.  

 

After the clamps were removed.  Not perfect but I have more work to do fixing tear out and adding the sleeve that will cover most of this anyway.  I didn't want to loosen the entire joint so this is the best I could do without risking breaking something. 

 

This is when I realized that a decision I made to not machine the cutout for the sleeve to join the segments together was going to bite me.  Because  you cannot put a square corner with a round bit, especially using a ball nose bit for the finishing pass, I removed the recess for the sleeve when I exported the STL files from Onshape. 

 

 

I decided that I would cut or file this by hand later but I should have let the CNC rough it out so that I could just square it up and fine tune it.  I will add them back for the straight sections.  I also discovered that this end was not round.  I think this is a combination of the glue up challenges and having to over sand the joints between the two halves to get a tight joint.  I will save fixing this to a later step as it  very possible that I could destroy the bell in the process.

Next up is making the straight sections.    

Thanks for following along.