Yearly Archives: 2011

From Square Sticks to Cubes

I recently wrote about the first of my puzzle making jigs to create square sticks as the first stage in creating the building blocks of many puzzles. With that jig successfully completed, and working pretty well, I had to move to the next stage and create some cubes. I said in the Square stick post that I’d tell you about it soon. Well, soon is now, and it’s time to make a crosscut sled!

The starting point for a crosscut sled

The starting point for a crosscut sled

For this jig, I needed a larger platform than the square stick jig, and as such I was going to be using both miter slots on the table saw. I cut myself a slab of MDF and marked it up for adding runners. I don’t have enough of the fancy metal miter bar that I used for the last jig, so I made my own. Starting with a strip of wood rough cut to the correct size for my miter slot, I sneaked up on the correct width by taking thousands of an inch off at a time until I had a snug fit. It didn’t take as long as I thought it might, and I think I now have an even better runner than the metal versions I used on the first jig.

My intention was to drill and counter sink holes in the top of the sled and screw straight into the runners below. Sadly that didn’t quite work out as planned as I didn’t have screws which would fit. What I had was either too short or too long so I had to go to Plan B. I decided to screw through from the runner side into the base. As it turns out, it wasn’t too much of a change and in the end achieved the same end result.

Runners mounted and a slot cut

Runners mounted and a slot cut

With the runners mounted, I flipped the jig over and tested the fit in the miter slots with the blade below the table. I had a couple of spots which were binding slightly, so I lightly sanded the offending areas until I had a tight but smooth fit with no wobble in the jig. For anyone wondering, the way I found out where the runners were binding was to run a sharpie along the length of the runner sides, then move the jig in the slots. When you take the jig back out, where the sharpie has been rubbed away is where you need to remove a small amount of material. Simple yet effective!

With the sled moving smoothly, I raised the blade up, and cut myself a slot part way through the sled. That slot will help aligning the crosscut fence as I need it to be perfectly square to the blade. Without having the blade through the sled that would be almost impossible.

Some reinforcement across the back of the sled.

Some reinforcement across the back of the sled.

Given the slot I’ve just created, I now have a potential weak point in the sled where it could flex, and degrade the cut. Despite how it may look in the photograph, I wasn’t trying to create a wing to add down-force to the sled.

Each of the three spacers on either side is screwed to the piece below with three screws, and gives enough clearance that with the blade raised to the point that the blade stiffeners are just below the base of the jig, it doesn’t cut into the cross bar. The cross bar keeps either side of the sled stable and prevents any twisting or warping of the sled. Yet again simple but effective. (I’m liking the fact that I can keep things simple. Less chance of things going wrong!)

A beautiful Dragonfly that stopped to see what I was doing

A beautiful Dragonfly that stopped to see what I was doing

Given that it was getting on for around 5 hours that I’d been working on the saw and building the jigs by this point, I took a break, stretched out my legs and back and relaxed for ten minutes. While I was doing that, I spotted a dragonfly floating around the garden. Standing watching him for a few minutes, he landed on the radio antenna for our car. Given that I still had the camera in my pocket from taking pictures of what I was doing, I grabbed it, as he posed for me! Looking at the pictures later, there really came out well, so I thought I’d share…

Anyway, back to the jig …

Adding the fence and a stop

Adding the fence and a stop

With the back reinforced, I had to add the fence to the sled. The important thing here is that it is at exactly 90 degrees to the saw blade. If it’s off, then the cut will not be square, which isn’t going to make for a good cube. Using the best square I have, I sat one edge against the blade, and the other against the fence board I’d cut. Keeping both edges in firm contact, I drilled one screw hole, counter sank and then screwed the fence to the sled at one corner. With one corner in place, I double and triple checked the fence was square, on both sides of the blade (using two squares). That may be a little overkill, since the blade shouldn’t be different on either side, but I figured it couldn’t hurt!

Everything checked, I pre-drilled and screwed the other side of the fence to the sled, and checked again for squareness. With everything looking good, it was time to add the stop. Now this was something that I’d been puzzling over for a good few days at this point, and hadn’t really figured out how I was going to perfectly measure the offset so that I ended up with good low tolerance cubes. As is ever the case, the answer came to me when I least expected it, in the shower in the morning before heading to work.

The answer. Make sure that the stop is far enough from the blade to make the biggest cut you’ll need, plus a bit. The reason … Well if you properly size a piece you want to cut (and I’ll get to that) then you can use that piece against the stop, then cut a ‘spacer’ by placing it against the fence, and the piece you want to cut, then cut that piece. You get a perfect spacer, and no complicated measuring required. (Other than the piece you want to end up with). Yes, I know that’s all very confusing, but I’ll annotate the pics below and it will make more sense!

Clamp removed now that the stop is screwed in place

Clamp removed now that the stop is screwed in place

Having screwed the stop in place I could remove the clamp. The stop is placed 6″ from the blade. I’m unlikely to ever cut a stick that needs to be 6″, unless perhaps I’m making 18 piece burrs, so this is lots of space to create whatever sized stick I need.


Creating the spacer for a single cube

Creating the spacer for a single cube

If you click on the image on the left, you’ll see I’ve annotated it to make my previous explanation simpler to understand. The single cube in walnut was created by shaving a few thousands at a time from the edge of the block and measuring after each pass until it matched the dimension of the square stick. In some regards, that is probably the most time consuming part of the process, as if you take off too much, then it’s a case of starting again.

With the first cube created by hand, it can be placed against the stop, then a long stick placed against it to create the spacer. In the diagram, you’ll see I created mine from some of the redwood sticks I’d cut when I was testing the square stick jig.

Batching out some cubes

Batching out some cubes

Now it’s a simple task of swapping the order of the pieces so that the spacer is against the stop, and batching out some cubes. I’ve used a clamp to keep the spacer in place, both against the fence, and hard up against the stop to make sure that it doesn’t move between cuts. It’s also important after each cut to clear out any dust that gets between the block and the fence as this will affect the accuracy of the cut. It is possible to adjust the spacer by adding feeler gauges between the stop and the spacer and before each run of cubes, I’ll need to check the sizes to make sure everything is within an acceptable tolerance.

Running through a few cubes, each cube came out almost perfectly. The worst cube had a tolerance of 0.001″ from the size of the square sticks. I can’t really complain when the cuts are that close. It’s going to be pretty good for any initial puzzles I make.

Making double size cubes

Making double size cubes

Now with a few cubes cut, I measured them, and selecting two that were exactly cubic, I placed them together and verified that the length was exactly double that of a single cube. Repeating the process I had used for the first cube, I cut a ‘two cube spacer’. This spacer will let me make double sized cubes, perfect for making any of Stewart Coffin’s Convolution/Involution/Involute puzzles.

Let the fun commence!

A Humbling Experience

Many of you know me on a few of the forums around and about the puzzling community, and a fairly well known Puzzle Box maker, let’s go with Allard’s name for him and call him ‘Stick guy’ posted asking what I was up to. It’s no secret I’ve bought a bunch of tools, and even started to use them to create the building blocks of puzzles, but I’ve never really mentioned what I was planning.

Well I answered Stick Guy’s “challenge”, and put up a brief summary of what I had been doing and what I was doing. You’ll know if you’re a regular reader that I designed a puzzle which I call Lock Cube some time back. I even prototyped it in Lego, then had it printed at Shapeways. Well at some point I’ll be making it out of wood too. (At least that’s the plan)

So here’s where things get interesting, and when I get to the point of the title of the post. Seems like a few people out there are interested in owning a copy of my Lock Cube, when I make it.

Now at this point, many things go through my head, including a few that I can’t print…

“Are you serious?”
“You really want one?”
“People want to own a puzzle I designed?”
“Is my puzzle good enough?”
“What will people think of it?”

The bottom line is that I was truly humbled by the response from quite a few people asking if I’d make a copy for them. I never expected to make more than just the one for myself, so this was a shock for me, and really left me not quite sure what to say. Quite impressive really since I’ve written an entire post about it!

So to everyone that has already shouted ‘Me please’ for a copy of a puzzle that I’ve not yet made from wood – Thank you.

If you want a copy, let me know. I’m not promising anything at this point, but I’ll keep it in mind as I make those early copies.

Zauberflote – Magic Flute

Eric Fuller recently offered a few new puzzles through Cubic Dissection and I picked up “Zauberflote” designed by Gregory Bendetti as well as “Stand Py Me” which I reviewed recently. Both puzzles sold out very quickly.

Zauberflote translates as “Magic Flute” and is an opera in two acts composed in 1791 by Wolfgang Amadeus Mozart. Gregory wanted to make a series of puzzles which had a link to the opera which he enjoys.

Zauberflote - Königin der Nacht pieces

Zauberflote - Königin der Nacht pieces

In a change from my usual style, I’m not showing the completed puzzle at the top of the post, but rather the pieces. I’ll get to the reason why soon enough. Eric has created this 4 piece version of Zauberflote from acrylic and yellowheart, and describes it as a pocket puzzle, given that its full length is just 2.25″. Gregory gave the four piece version the full name “Zauberflote – Königin der Nacht”, and each of the puzzle in the series with a different number of pieces in the flute has a different sub-name. I really like the use of the acrylic here, as even when the puzzle is solved (as you’ll see below) you can still see the internal burr of the wooden pieces, which is a nice touch. Eric made 45 copies of this puzzle, and they are all signed with Eric’s usual squiggle.

I spent about 30 minutes working on this puzzle, and after a few false starts I found a way to get all the pieces in place and the flute shape (or possibly more of a set of pan pipes) is easy to see. When I was solving it, I started by putting the smallest piece in first, and I required a couple of rotations to get the pieces into their final location.

My solution requiring rotations

My solution requiring rotations

Feeling quite happy with myself I put the puzzle aside for a few days. When I came back to it, I opened the trusty Burr Tools and created a model of the puzzle there. Now I fully expected burr tools to be able to put the pieces in place, but I didn’t expect it to be able to give me an assembly given that rotations were needed (when I solved it). To my surprise, Burr tools came back with 72 solutions and one assembly!

Solution found by Burr Tools

Solution found by Burr Tools

Burr tools notes a 14.4.2 assembly and shows that it is possible to solve the puzzle without using rotations as I had. If you look very closely at the two images, you’ll see that the internal burrs are in different locations showing that clearly it’s a different solution. Also Burr Tools puts the largest piece in first, although I believe it is possible to insert the pieces in any order.


So having used burr tools, I think there are more solutions than it shows, even without the rotations. I did talk with Gregory as to whether rotations were intended, and he admitted that he hadn’t checked for rotations, but it wasn’t cheating, since I still solved the puzzle without forcing the pieces, and had found a solution that he hadn’t. The solution with rotations is much shorter at 7.1.1.2 (if my counting is correct).

Overall, this is a fun puzzle, which isn’t too hard and is very nicely made by Eric.

Cutting Square Sticks

The start to many a good puzzle is a square stick. But not just any square stick. One which is accurately sized along its length and each side square to its neighbor. Not an easy task.

A while back, Scott Peterson sent me a tutorial for building a table saw jig to create square sticks. Over the Labor Day weekend which is a holiday for us in the US, I set about making the jig (mostly) following Scott’s tutorial so that I’m firmly on the road to making my own puzzles. Scott kindly game me permission to make the tutorial available to anyone that wants it, so feel free to download a copy.

The following is my attempt at creating square sticks …

Starting off with a sheet of 3/4″ MDF I cut myself a strip that was wide enough to go between the miter slot and just past the blade of the saw. I left a good sized area on the other side of the miter slot so I had room to attach a handle to help push the jig through the cut, and back after the cut. I marked the position for the miter bar and after pre-drilling the holes, screwed it in place.

Preparing to attach the miter bar.

Preparing to attach the miter bar.

Miter bar attached to the base of the jig

Miter bar attached to the base of the jig

The miter bar that I bought has plastic screws that protrude from one side to ensure that the bar is a perfect for for the slot. Taking some time I adjusted each of the screws to make sure that there was no wobble in the slot, as this would compromise the cut.

Taking Scott’s advice, I also added a UHMW plastic base to the cut side of the jig to back up the cut and help prevent tearout. Using blue tape I secured the plastic to the base, and marked out the positions for the screws, predrilled, countersunk then screwed the plate down. The Blue tape came in really handy as it’s almost impossible to mark the UHMW plastic with a pencil.

UHMW plastic set in place ready for drilling and screwing

UHMW plastic set in place ready for drilling and screwing

The UHMW screwed down.  Note the screws are all below the surface of the plastic

The UHMW screwed down. Note the screws are all below the surface of the plastic

With that done, It was time to make the first cut, and trim the extra material that I had left off the edge of the jig, ensuring that the blade ran along the edge of the jig which is needed for an accurate cut. Making sure that the blade was at 90 degrees to the table top, I made the first cut. And immediately noticed that there was a gap at the front of the jig which wasn’t there at the back. It was pretty small, but enough that I could see light between the blade and the jig, so I knew the jig wasn’t good enough. It seems that I didn’t do a good enough job of adjusting the miter bar, so I readjusted, and remade the cut. Fortunately, I’d left enough space between the edge of the jig and the screws holding the UHMW plastic down to be able to do this. This time things were much better and I got the accurate cut I was looking for.

Ready to screw the fence to the jig

Ready to screw the fence to the jig

The magnets to hold the steel bar in place

The magnets to hold the steel bar in place

Fence screwed gown and ready to go.

Fence screwed gown and ready to go.

Next up I need to install the fence on the sled which the stick will be clamped against when making the square sticks. As Scott notes in the tutorial, I have a steel bar as a spacer, and use some brass feeler gauges to get an accurate cut. I also embedded some Neodymium magnets in the fence to hold the steel spacer firmly against the fence. The fence is clamped securely against the saw’s rip fence, then predrilled and screwed down.

Those eagle eyed readers out there may notice a couple of extra drilled holes in the fence. Well yet again, I managed to get things a bit wrong, and screwed the fence in a little squint. Now that’s not a lot of use when you’re trying to get accurate cuts, so I re-drilled and re-attached the fence. So there’s a couple of extra holes there now. It won’t affect the jig, but a squint fence would have!


Adding some toggle clamps to the fence and I have something where my fingers are well out of the way, and the sticks will be firmly held against the fence.

Cutting a stick square

Cutting a stick square

Ten minutes later; five identical sticks

Ten minutes later; five identical sticks

With everything set, It was time to put some sticks in there and see if I could make some square sticks and whether they would be accurate or not. At this point I was looking to test the jig, so I used some redwood I had to spare, and ripped it into roughly sized sticks. Now you’ll notice the knots in the wood so even I knew things weren’t going to be perfect. Still it was good enough for a test.

Measuring the sticks using my digital calipers, I am getting sticks with a 0.005″ tolerance between either end of the 12″ sticks. That’s not too shabby. Given the knots in the wood, which will most likely make it twist as the knot is cut, I’m fairly happy with the results. At this stage in my puzzle making career, it should be more than accurate enough for making cube based puzzles, and burrs.

But that’s not the end of the story ….

I came back to the jig in the morning, used some walnut that I had bought a few weeks back for making puzzles, and used that to get some square sticks. The walnut is much nicer wood, no knots, and a beautiful grain to it …. The results. I cut four 9″ long sticks with a tolerence of 0.002″ across all four sticks. Now that’s a result I’m proud of!

Next up, I need to create a crosscut sled so that I can cut these sticks into accurate cubes for use in making puzzles. That post is coming soon, so stay tuned.

Stand Py Me

Stand Py Me is a new puzzle from Gregory Benedetti. It’s familiar shape is similar to Stewart Coffin’s Three Piece Puzzle, however it comes as four pieces (plus the stand) not three, and every bit as challenging to solve.

Stand Py Me by Gregory Benedetti

Stand Py Me by Gregory Benedetti

This version was made by Eric Fuller, and stands at 2.75″ with a Wenge frame and Zebrawood and Maple blocks to make the pyramid shape. Gregory comments that the name is a play on words; the Py, not coming from Pi (3.14159265…) but Pyramid. It’s as though a Pyramid is saying “Put me on a stand”. Well regardless of how it was named, it’s a fun name, and really fun puzzle.

The blocks which make up the puzzle are all joined on a half face or quarter face making for some interestingly shaped pieces. The puzzle is made significantly harder by the addition of the frame. It’s fairly easy to create a pyramid of the blocks outside the frame, but doing it so that the pieces are captured inside the frame is a lot tougher.

Four pieces, plus the frame.  Signed by Eric

Four pieces, plus the frame. Signed by Eric

To get all the pieces in place, you need to move them around much like a standard Burr puzzle, which leads to a 5.1.2 solution. In case you’re not familiar that means 5 moves to insert the first piece, one for the second and 2 for the third. So it’s a tricky puzzle, and took me around 40 minutes to solve it the first time. Now knowing how the pieces fit it takes just a few seconds, although I keep putting the pieces in ‘upside down’ so I don’t end up with the Zebrawood pieces in the corners the way it was made to be.

Having solved it, I modeled the puzzle in Burr Tools, and it confirms the 5.1.2 difficulty, and also points out that there are 2 solutions (not including rotations) but only one assembly.

Eric made 30 copies of Gregory’s design and has signed and dated them. I picked mine up at a recent Puzzle Paradise auction, although they all went very quickly. This is a really well made puzzle, and the choice of woods really shows it off well.

Interestingly, the puzzle doesn’t hold itself inside the stand overly well, and wants to fall out of the gaps, meaning that you really need to turn it upside down and hang it to get it to keep its shape. Not a problem, but it is something that even Gregory admits himself.

Coffin’s Rosebud and Cluster Plus

Before my visit to see Scott Peterson’s workshop a few weeks back, I’d been talking with Scott for a while via email. I’d introduced myself, and asked if there was any chance he could make a couple of puzzles for me, including a Rosebud, and a Peanut puzzle. I’d also mentioned that I was starting to make my own puzzles, and Scott was only too willing to help out with tips and advice, answering all my dumb questions with the sort of enthusiasm and excitement that I remember from my high school science teachers.

As it happens, Scott hasn’t made a Peanut puzzle for a very long time, and he tends to make a run of puzzles then move to the next design, however he had just finished a run of Rosebud puzzles. At the time he didn’t have any available, but he thought he may be able to make me one at some point. With an offer like that I duly said thank you and never thought about it again. In the time between that original email, and my visit, we’d exchanged a number of emails, mainly about woodworking, and Scott’s tips to help me get started. Not long after we arranged to meet up, Scott got back to me saying he had a Rosebud available if I still wanted it, and he had a jig too. Like a bobble head on a dirt road, I instantly said ‘yes please’.

Rosebud in its closed state

Rosebud in its closed state

This particular copy is made from East Indian Rosewood, and Flamewood. The strong red from the Flamewood, and its detailed grain pattern really make this an object of beauty. I must confess that the pictures don’t really do it justice. The finish is glassy smooth, and the fit of the pieces is incredible. This is a tight fitting puzzle, with no visible gaps, and yet the pieces slide past one another with ease.

The Rosebud is Stewart Coffin design #39 in the AP-ART catalogue, and is one of the few three-dimensional puzzles which assembles into more than one solution shape. The image above shows the rosebud in the shape which gives it its name, and is a coordinate motion solution where all six pieces are required to move along their three-fold axis at the same time to both assemble and disassemble the puzzle. Not only is this difficult, but the manner in which they are assembled is not at all obvious and requires considerable dexterity.

Stewart himself says “A few of these puzzles were produced some years ago and sold unassembled. After sufficient time had elapsed and almost none had been solved, the customers were given the opportunity to purchase (for an outrageous price!) an assembly jig and directions. With these, it is easy. Without the jig, it can be done with patience, using tape and rubber bands. Without such aids, it has been done but borders on the impossible.”

The reward for being insane enough to try this without a jig is a wonderful movement where the puzzle appears to open like the petals of a rose. It really is something which has to be seen to truly appreciate. My video review below will show you just that and give you a chance to see just how stunning the movement (and Scott’s work) truly is.

As you’ll see, both in the video review, and also in the photograph below, Scott’s work really is outstanding. The pieces of the puzzle will expand until just the very tips are touching, and it still remains perfectly stable. Only with very precise construction can you even get close to this, and it’s a testament to the years of puzzle making that Scott’s copy is as good.

Rosebud opened.

Rosebud opened.

As a puzzle, without Scott’s jig, I have assembled the rosebud. I didn’t go down Stewart’s suggested route of tape and rubber bands (as frankly I couldn’t see how a rubber band would help me much) but all I’ll say is it took over an hour, involved a lot of swearing, and wasn’t all that fun an experience. The jig on the other hand makes things much easier, and removes the worry about expanding the puzzle just a little too far. If it does collapse then you can easily re-assemble it. I’m glad I accepted Scott’s offer for the jig!

Cluster Plus

If you’ve watched the video review, then you’ll know this already, but while I was visiting Scott, he handed me a rather special puzzle.

Cluster Plus - Shop Copy

Cluster Plus - Shop Copy

The Cluster Plus puzzle is #114 in the AP-Art catalog, and has a striking resemblance to the Cluster Buster puzzle #49. The final shape however is where the similarities end. Originally only 11 of these were made, and almost half of those were given out as tokens at IPP-17. There don’t seem to be to many people have made these designs since, so that makes this copy even more special. This is one of the prototype copies Scott made with collaboration with Mark McCallum. As such, it’s not a finished quality and has the pieces numbered in the centre to help with assembly. The six dissimilar pieces are joined together in sets of three pieces which require a coordinate motion, and then the two halves of the puzzle are slid together. Looking at the two halves, it’s almost impossible to see how Stewart Coffin came up with this design. Even knowing how to put the pieces together, and the order I still found this a challenge.

Cluster Plus Halves

Cluster Plus Halves

When I first attempted to fully disassemble the puzzle, I split the two halves, then took only one half apart fully, leaving the other still assembled. Now, remember I have the pieces numbered as you can see above, and I’ve only taken one half apart. I still took over 20 minutes to get the puzzle back together.

The coordinate motion really makes it difficult to see how the pieces should come together, and add to that numerous ways that the pieces can fit together, and this is one tough puzzle. As an early copy, the co-ordinate motion on the first three pieces isn’t perfect, and this was something Scott was working on. His initial idea was to leave more of the centre block – one of the standard building blocks in Coffin puzzles, and the number part in the photo in case you’re not sure, with more of the central section un-sanded in the hope that this made things easier to assemble. The trick here was that the centre blocks normally have a pyramid peak. Stewart Coffin removed part of this to allow the coordinate motion which otherwise would not have been possible. In the end Scott removed more of these centre blocks as it turned out that this initial idea wasn’t great.

While Scott and Mark were working on this, Mark believed that there was more than one configuration of pieces which would make up the same shape. I took up that challenge and have been trying to find an alternate configuration. So far I’ve not been successful, however I’ve not given up and I’ll be sure to post if I find it.

At this point, I have a bit of a confession to make. When I was playing with the puzzle, I may have used a little too much force. I should point out that the fit here is not perfect, again since this is a shop copy, and given the nature of the puzzle, any inaccuracy is multiplied out to the end of the ‘arms’ which is where the fit needs to be perfect. Well when putting it back together at the end of a puzzling session, I used a little too much force, and with an impressive snap, one of the arms flew across the table, and the two halves slammed rather abruptly together. Sorry Scott!

A little bit of glue and an hour later, things were back to the way the should be and no real damage done. The glue joint had given out, but didn’t damage the pieces of the puzzle fortunately so it was a simple fix.

Thanks again to Scott for both of these puzzles. There’s great additions to my collection and I promise to try not to break them again!