Everything in this set depends on two things: cutting the beads and cutting the grooves. And those two things both rely on using the right router bits.
Watch the video below for the short and sweet summary. Read the text below for more details.
The Right Router Bits
There are two bits used for this project—a ball-end bit and a bead profile bit. In this case, both bits have a main diameter of 3/8”. Through some testing and some math, I determined that 3/8” grooves and beads gave me enough strength in the narrow parts of the pieces, but also kept the posts and panels to a reasonable size. I tried 1/2” diameter bits, but they required 2 1/4” wide posts, which felt like baseball bats and would have weighed a ton all together. 3/8” was a better way to go.
The ball-end bit was easy enough to find. I used a Whiteside bit, purchased online for around $30. But I couldn’t finalize a ball-end purchase until I knew I had a perfectly matching profile bit to match. And that profile bit was very hard to find. In fact, it was impossible.
About 18 months ago, I started my online and catalog search for a bit that would cut half of a “keyhole” profile. It seemed straightforward enough, and a lot of bits looked like they were close. But they all had one problem or another: one would leave a “neck” on the bead that would be far too narrow, another’s would be too thick. Some would cut such a deep neck that the bead would be “floating” about 3/8” or more from the rest of the panel, making it quite likely that the entire bead could just be snapped off. I wrote to Amana, Infinity, and other manufacturers and the answer was always the same: We don’t have anything like that, and we’ve never seen another set like it.
Originally, Amana seemed interested. I spoke with Lonnie Bird, Amana’s spokesman and a contributor to Fine Woodworking, and sent them detailed specs in a PDF. He said they were interested in producing the set, but after a few months I had heard nothing and started my search again.
The second time around, I also looked at getting a complete custom bit made. I was quoted $120 to $150 for a one-off solid carbide bit. That was a bit more than I wanted to spend, but it was an option. Then I got a great suggestion from the product information expert at Infinity. He said to speak with a sharpening service to see if they could modify an existing bit to match my measurements.
I did a quick search online and found a few firms in the general area that seemed reputable, including one in Canton, Ohio called Cammel Saw Company. I sent an email to the Contact Us address on their site and quickly got a return email from Dennis Cammel. I sent him my PDF specification and had a quick phone call with him, and he said they could do it. They sell high quality Whiteside bits. Dennis made a recommendation on which source bit to start with and said the additional custom sharpening would cost a mere $10.
About a week later, a package arrived in the mail from Cammel Saw with my first set of bits—the 1/2” diameter ball-end and profile bits I mentioned above. After my first attempt I knew I was on the right track. The bits cut perfectly and left a mirror finish. With a little adjustment to my router table fence, to make the bead just a hair narrower than perfectly round, I got two pieces that slid together easily, but with fairly little slop. But the 1/2” set was just too large.
I called Dennis back and asked if he could do the exact same thing, but with 3/8” diameter bits. Again, 10 days later a new set arrived, as perfect as the first. I would highly recommend working with Dennis Cammel and his team at Cammel Saw if you ever have similar needs, and the odds are high that I will also be using him for any regular sharpening I need for bits or blades.
Step By Step
1. Joint or plane the planks to get the desired thickness. I started with 4/4 planks, and planed them down to 5/8”—thin enough to be easy to handle by little hands, but thick enough to stand up to their abuse. And I used 1 3/4” posts, cut from 8/4 boards. I used all ash wood, because ash is hard and durable, and has a high elasticity. It will bend before breaking. There’s a reason it used to be the go-to wood for baseball bats. There’s a lot of maple used in bats these days, but they always shatter when the break. Not good for a kid’s toy.
2. Cut the planks and posts to the right width. The posts are all 1 3/4” on a side, so that’s easy at the table saw. The planks were a bit trickier because I had to calculate the right widths so that two narrow planks and a post would be equal to a wide plank. You can make you planks whatever width you like, but it’s important to keep that ratio of two narrow planks and one post equals one wide plank so that the structures you create can interchangeably use wide or narrow panels and still fit together.
3. Cut the planks and posts to length. The length of each post and panel is again a matter of choice. I chose 9” lengths for posts and regular panels. That was a good scale for the various dolls and figures we already owned. Much taller than that would use a lot of wood, be getting pretty heavy, and would physically be a bit set to manage. Any smaller wouldn’t feel like a castle wall, when placed next to a 6” figure. So 9” worked for me. The small posts and panels are simply half that height so that, again, two small posts are the same as a large post.
4. Cut a test groove in a scrap post. The 1/4” dado worked well with a 3/8” ball-end bit. That would leave a decently wide throat for the neck of the bead to slide through, but still be narrow enough for the ball-end bit to cut a clearly wider groove. It took a little trial and error to get the dado blade depth and the router table ball-end bit depth to work together. The router table had to cut slightly deeper to that the square corners of the dado cuts met the edge of the ball-end bit. Once I had this working, I carefully measured and wrote down the fence positions and bit/blade heights on the test piece. This would prove to be super handy down the road.
5. Cut a test bead on a scrap panel. Once I had a test post, I needed to figure out the right fence and bit height settings to cut a matching bead. My first step was to place the test post on the router table, and raise the profile bit until it seemed to match the shape of the groove. I removed the test piece, moved the fence in to guide the panel and did a test cut. Each time I did a test, I ran the panel across the bit, then flipped it end-for-end and ran it through again. This was to mimic the actual process I would use for the final pieces, and would ensure that as I moved the fence closer and closer, I would always have a symmetrical and centered bead. Eventually, it got close. I would take a pass on each side and then test it in the test post. I’d take a hair more of each side and test again. The micro adjustment knob on my Kreg router table was very useful for this process of sneaking up on the right bead width. When it was right, I once again measured and recorded all the settings on the panel itself.
6. Mass post production. With a matching test post and panel, I could get to work cutting the rest of the pieces. With a stack of 9” posts on a table, I began cutting dozens of dados, checking every once in a while that the fence was still secure and the cuts were all of consistent depth, and were centered on the faces of the posts.
When the dados were all cut, I moved to the router table. Using the test post as a guide, I inserted the ball-end bit again and raised it until I could just slide the test post over the bit. I then slide the fence over until it met the edge of the test post. I removed the test piece, turned on the router, and ran the test piece over the bit again to ensure that no additional wood was being removed. Assured of its setup, I began running the dados over the ball-end bit, one by one, 8 passes per piece (4 in each direction).
With each finished post, I would pick up the test panel and quickly slide it through each groove. Normal human motions will occasionally mean I applied more less pressure against the fence during a given cut. Even a few thousandths of an inch can be felt by the sliding pieces. Every once in a while I got one that faintly stuck in a certain place, so I ran that post over the bit again to clean it up.
7. Mass panel production. With the posts complete, I turned to the panels. Again, I used my sample panel as a guide to quickly get the bit height set. I then set the fence about halfway to its final position and did light passes on both sides of all of the panels. Trying to cut the full profile in one pass would have led to lots of tear out or splintering of the wood and likely would have damaged the bit or the router. Much better to take more, lighter passes.
So after I completed a first pass on both faces, on both sides of all of the panels, I moved the fence to it’s final spot and re-ran one panel to ensure a good fit. When I saw that it fit perfectly with the posts, I began re-running all of the panels at their final depth.
8. Cut pieces to final length My source wood for the posts had started as roughly 3’ long planks. That worked well to produce four 9” posts per length. So at this point I cut the posts into their 9” and 4.5” final lengths.
The panels had started as 6’ to 8’ long planks. Since I would be running these planks on edge on the router, I cut them to final length before routing. I wasn’t comfortable trying to manage long panels of that width on edge. In hindsight, I should have constructed an add-on fence for the router table to better support panels that wide up on edge.
9. Sanding, sanding, sanding. I realized there were a LOT of edges in these pieces, and since I was building a toy for very young children, I needed to hit every edge to make sure nobody got splinters. I also expected to sand the faces of the planks, but when I began that process I realized the glossy surface left by my new Dewalt 735 planer, when set at high quality cut, was superior to what I would get with 180 grit sandpaper. So I just left the faces alone. Thank goodness for small miracles. I still ended up building myself a quick downdraft box to minimize the dust from all of this close-quarters hand sanding. I also built some specialized sanding tools (wrapping adhesive sandpaper around the pointed tip of a pencil) to get into the nooks at the end of each post.
10. Beeswax Finish. Again, because this was going to be used by little kids and even an infant, I had to make sure the finish was safe. I wasn’t too worried about dings and scratches—it’s a hardwood toy; it will get dinged, but it will be tough. But I did want to keep the wood from getting too rough, and I wanted a nice feel to the wood. The Salad Bowl finish was perfect. It’s a blend of mineral oil and beeswax. You can rub it on with your fingers and it leaves the wood just slightly darker. Two added benefits are that it smells great and natural, and it actually makes the pieces slide together even more easily. With the finish added, the pieces have a smooth, fluid, quiet action. And if I ever need to add more, I can reapply finish in the house and immediately hand the pieces back to my kids to keep playing. No smell, and no waiting necessary.
11. Building the Box. Once I had all of the pieces built, I was finally able to plan out and design a box to store and carry them all. At first I thought of building a typical toy box with a flip up lid, etc. Then it occurred to me to make the box part of the set. By using posts for the corners of the box, with regular grooves cut in the outside faces, the box can become the center of the castle and the walls can be attached and built out from the corners. That meant, once again, I needed to build it so that the height, width and length were all multiples of the narrow panels and post widths, so everything could still meet at a right angle. Fortunately, the math worked out so that I could build a box about 2 wide panels x 1.5 wide panels x 1.5 posts tall. The last feature of the box was to cut permanent “doorways” into the sides. These doorways allow the castle residents to get into and out of the main room. Also, they conveniently double as carrying handles on all four sides of the storage box. I decided to leave the top of the box open to make it easy to see the pieces inside, reach the people or furniture you put into the castle, and eventually to put the pieces back when playtime is over.
I built the box out of 1/2” baltic birch plywood, with each edge reinforced with a strip of solid ash. The corners are joined by simple dados and Titebond woodglue holding the plywood in place. The bottom is another piece of 1/2” plywood resting on 1/2” ledger strips that are glued and screwed into the bottom edges of the sides.
The box is also finished entirely in the Salad Bowl finish.
At the end of the day, once I had the router bits I needed, the whole project probably consumed about 60 hours of cutting, routing, sanding and finishing. If I had to do it again, I could probably shave a few hours off now that I have good samples to start from. But it’s still a manual labor-intensive undertaking. If you had two table saws, and two router tables, and some friends to help with sanding and finishing, you could knock out a whole set in a day. Maybe.
Thanks for reading. I hope you enjoyed the blog.
-- Pete in Wilmette, IL