Building the Woodgears Box Joint Jig

I have been fascinated by the Woodgears.ca website for years. Matthias Wandel builds fascinating gizmos, jigs, tools as well as interesting wood projects, and posts new items regularly.

In particular I have been thinking about building a version of his Screw Advance Box Joint Jig for a few years now. Not that long ago I decided that I would try and get started on it over this Christmas holidays as a present to myself. And of course if I'm doing that, a photo essay here was almost mandatory.

If you want detailed build instructions, please go to his website and read his web page, and/or watch his instructional videos, and/or buy his plans -- they're quite inexpensive. My intent here is to just post some photos of my own build, with a focus on what I did different, or areas where I had particular challenges, or anywhere that I thought his plans needed a little elaboration. So I will NOT be posting detailed instructions or plans here, and if you are not familiar with his jig you will likely not understand what I'm doing!

Note that I wrote this document while I was building the jig, so this is mostly written in the present tense, and with no knowledge of what is coming up next.

Building The Jig:

So, onward.

I actually got started a few months ago when I bought the Freud Box Joint Cutter set at my local tool store. The blade has a good reputation in various online forums, and was also recommended by the guy at my store (whom I trust.) Now, curiously enough, Matthias actually uses all kinds of different blades with his jig -- none of which, to my recollection, are actually purpose-designed for box joints. So I hope this will work out, we'll see.

The first thing that I advise is to do is go out and buy yourself one of these. Look closely, it is not just an ordinary tape measure! Matthias works in metric; his plans and instructions are almost entirely in metric, so you will want to have a metric tape measure at minimum. I also have a combination square which reads in metric, which is proving to be extremely useful.

I am not used to working off plans other than my own. I've been designing my own projects pretty much since I began woodworking. So working off of plans is already a stretch.

One thing unique about Matthias' plans is the inclusion of these 1:1 plans. I've watched several of his videos where he demonstrates how he uses 1:1 plans in a lot of his projects. It intrigues me, so with this project I went ahead and printed out the 1:1 plans and taped them together. (This took two tries, two sets of plans, to get right. It is not quite as simple as he makes it look in his videos.) And so I am trying out his method of taking some dimensions and cutting instructions directly off of the plans.

You'll need to go shopping for some hardware. Some of the requirements you might already have in your shop. Check through your bolts and jig parts. The threaded rod, 2x10, and other random t-nuts and so on, were easy to find at the local Big Box lumber/hardware store. I also had some jig knobs/bits from Lee Valley which I'd picked up years ago for various projects. The metal bar for the miter slot is also fairly easy to find at the local metal supermarket, though I haven't yet picked it up.

The bearings are another matter. Home Depot did not carry them. Princess Auto has some, but not the right size. I happen to have some broken down Rollerblades which also has pretty decent bearings in the wheels -- I plan to try to use one of those for the Idler assembly, since I don't believe that the bearing size is critical there. However, for the bearing that supports the threaded rod, size is critical. You'll need to look up a bearing supplier in your hometown. It was only about $3 each, so I grabbed two to have a spare.

Here is the stock holding box glued and clamped.

Typically I build with hardwood. Typically, Matthias uses a lot of softwood in his projects. If you watch his videos you will quickly realize that he is extremely frugal, and uses a lot of found or scrounged wood, as well as often looking for more inexpensive ways to accomplish something rather than simply buying the latest whatsit.

I do have extra hardwood on hand, and briefly considered just building it all out of hardwood. But as this is my first time using one of his plans, I decided to just go with his plans as much as possible and use the same materials he does. As well, this is going to be a fairly big jig, so I am also conscious of how much it is going to weigh!

I will make one comment on this step. The plans note that a single 8ft 2x8 piece of construction lumber will provide enough wood for the needed pieces. I would suggest instead buying the nicest 2x10 that you can find. Around here, the 2x10 lumber in the local big box store is FAR nicer and clearer than the 2x8 lumber. Check the rings on the end of the wood. With 2x8 lumber they can use smaller trees than with the 2x10s, so by buying a 2x10 and cutting out your pieces from the outside edges, you will get closer to quarter sawn, which will be more stable, and farther from the pith of the tree.

Another thing: Do buy your 2x stock a few weeks ahead of time to let it acclimatize to your shop. Construction lumber is not always the most dry and stable. Even so I would also build the box the same day that you cut and plane the pieces.

Here I am cutting out the slots for the corner splines in the stock holding box. In the instructions/video he just uses a jury rig with a cross cut sled. I can understand that, as it is only a few cuts. However I found that procedure a bit unsteady. As well, building a jig to hold this stable was only the work of about 10 minutes. You can see that it is just 4 pieces of scrap plywood assembled with nails/screw/glue.

NOTE: If you build a little jig like this, think in advance of where you are going to cut. For the sides, I kept the nails in the upper half of the boards. For the cradle portion, I only put in two screws, and kept both of them at the extreme edges. This leaves lots of room in middle for the saw blade. Note that for this cut you are raising up the saw blade quite high, so you need a lot of clearance around it.
Keep those nails away!

Early in the process, I realized that I needed the plans close at hand. Fortunately I happened to have an old clunker laptop that I could bring into the shop, which I did not care too much if it died due to the dust or other shop hazards.

This way I now had the plan package close at hand, and I could flip through the files, photos, instructions, as needed. Even more importantly, I could play snippets of the instructional videos. I found I would play roughly 1-3 minutes of video at a time, which would give me pointers as to what would be the next step in the process.

For the thin piece of hardwood to glue to the front of the stock box, I used some scrap maple hardwood flooring which I had obtained from a neighbour some time ago. The baked on finish on this stuff is REALLY hard. I learned the hard way to NOT try planing it off. Instead I rip about a 16th inch off of the top with my bandsaw. You can see the scrap on the left -- the finish is really strong as you can see. Beside it is the glue-up of two pieces of flooring

This is what happens when you cut a dado on the WRONG SIDE OF THE LINE. ARGH. The tolerances on this project are very precise, so I did not think I could just adjust the result, so I glued in a carefully ripped pieced of maple and then proceeded. Thank goodness for modern glues and long-grain joints! And thankfully I caught this before I'd make the next cut, as that would have really set me back!

Here is the next cut, which was fairly complicated to set up. This is the half-dovetail which is cut along the front of the stock-holding box, for where the box rides along the front rail.

Got the t-nut mounting piece fabricated, as well as the dado in the stock box where the threaded rod fits.

Gluing up some maple for the gear mount block. Working on this is way out of sequence, but I discovered that I did not have a piece of 12mm Baltic birch plywood that was large enough for the base. So, I'm kind of stuck until I can get to the lumberyard to buy some more.

This gear block is a pretty beefy piece of wood. So you likely will need to glue up a few chunks of hardwood, unless you happen to have some particularly thick stock on hand. The left-hand frame piece, where you mount the bearing, is another such piece. Three cm thick hardwood (1-1/4") is also not the most common. I wish I'd taken note of these pieces at the beginning, so I could get them all prepared at once, rather than having to come to a stop right now as I wait for the glue to dry.

Here is my first quibble with the plans -- and it's a minor one. (These plans are awesome.) The 1:1 plans show the screw locations for fastening the frame-front-rail to the base of the jig. The instructions suggest that you just lay the plans over the base and tap a scribe on each mark to transfer the location to the wood and then drill the holes.

My quibble is that the screws for the front rail are all centered on the front rail, which makes perfect sense. However at the left side of the front rail, we have earlier cut the rail skinnier, so it is only 3cm thick there. So those two screw locations are NOT positioned over the center of the rail. On my build, my front rail split when I screwed into it from the base. It was a small split and I was able to force in glue and clamp it closed, and then reposition the screw. I'm sure part of that is just the way the grain moved through my piece of wood, but I do think that positioning the screws so they were properly centered there on the left would have helped.

The front rail is just spruce -- very soft wood -- from a 2x10 from the local building center. So in the original how-to video he did not pre-drill these holes, and as I followed along neither did I. I probably should have, as that may have helped also. I definitely did pre-drill when I redid the hole in a new location!

I have written to Mattias about this, and it will be changed the next time he updates his plans.

Oops. The instructions call for drilling a stepped hole in the left side of the frame, where the bearing is inserted. However, I managed to somehow drill the step on the WRONG side of the board. I considered milling up a new board, but decided to just reverse this board, as it was already milled to a perfect fit, and drill the CORRECTLY POSITIONED stepped hole on the other end.

Also, I just happened to have a 7/8" plug cutter. (seriously, it's the only one I have, from a project I built years ago.) So I drilled out the stepped hole so it was 7/8" diameter the whole way through, and then used the plug cutter to make myself a matching oak plug, which I glued into the offending hole. A slim piece of oak was also glued into the slot. I don't think a whole lot of strength is really the issue there, but I liked the fact that I could (mostly) close up the mistake!

I then worked on the idler pulley assembly. Here I made use of a bearing from a junked Rollerblade wheel. Since the inside/outside diameters really are not crucial, this is a good place to substitute for something free.

Unfortunately, on my first try, it is positioned too low. Note from the photo how the bearing is riding against the softwood section of the front rail. Not good. The bearing is already digging in.

Fortunately, it was a simple matter to reposition the hinge a few mm higher on the side of the box.

Later on, since I had an extra rollerblade wheel bearing, I changed the idler pulley assembly to ride on TWO bearings. This spreads the load out over the wood also. (You can see this in a few photos way near the end of this document.)

A view of the jig partially assembled. Next I'll be working on the whole screw advance part of the jig.

Drilling out the axle hole from one of the two gears. This is a mistake, by the way. After I drilled out the holes and then cut out the gears I moved on in the instructions and realized that I was supposed to mount a T-nut in the larger 16-tooth gear. This of course requires a larger hole. The smaller 12-tooth gear requires the 3/8" hole just to accommodate the threaded rod. (In my defense, I'll point out that the drawing from the plans does not have this noted. But really, it's my own fault. I was just following along with the youtube video instructions, and this is noted in the written instructions which I was NOT reading. I am now...)

The plans call for 39cm (15") of 3/8 threaded rod, and of course I could not find that exact length in the store. It's easy enough to cut with a hacksaw. After cutting, roll it along a grinding wheel at a 45 degree angle, and you'll take off the burr and rough edges caused by hacksawing, and you should have no trouble getting nut on or off. (This is wonderfully illustrated by Ron Walters's in one of his video shop tips. If that link goes dead, simply google "Ron Walters cut threaded rod".)

What is going on?!? This is NOT going to fit in that space. There is no way to tip in the threaded rod with the bearing and nuts into place. I was perplexed. On the video, Matthias said it was tight, but this is way more than tight! I double checked against the 1:1 plans and the dimensions of the gap are correct.

The problem is the same as on the gear noted above. Actually I realized both problems about the same time. In the written instructions, it is explained that he was using 9mm Baltic birch plywood, which requires a 2mm wooden washer between the two gears. I had been watching the video, as mentioned, and NOT reading the written instructions, and I simply stuck in the washer. I was using 11mm Baltic birch, and that extra thickness is enough. The parts list notes that either thickness of plywood is fine. I think that this is probably the only spot where you might need to make allowances for the different thicknesses of plywood.

The video instructions are so good, that I had mostly been building this by watching the Four step-by-step videos a little bit at a time throughout the build. But These two situations drove home that I also need to check the written instructions for more details that are omitted or glossed over in the video.

Fortunately this was easily remedied by removing that wooden spacer.

And another view from the other side. (Just a note for clarity -- in this picture I am just positioning the rod + gears beside their mount point to verify that they fit. It is not yet mounted into place.)

I'm about to insert the screws that clamp the bearing into place in the bearing holder frame. What I often do when fastening screws into hardwood is to dip the tip of the screw into some paste wax. A bit of wax on the first couple rows of threads just makes the screw go in quick and easy. Of course I had also pre-drilled the holes (in the foreground). But, this oak was quite hard, and I really did not want any accidental splits. Cheap insurance.

Annnnnd another oops. This is supposed to be the gear mount block. I'm not sure how I got the orientation mixed up on this, but as you can see I ended up with the boards that make up this block turned sideways. That is not terrible, but the other issue was the holes that the bolts go into. You are supposed to drill these holes a touch oversized (the gear mounts use a 5/16" bolt.) because later you cut out the notch in the block and then it clamps down on the bolt when you tighten the lock down bolt. I don't remember what size drill bit I used but there was no way it would properly clamp down on the bolt.

The truly silly thing here is that the blank that I cut the gear mount block out of was large enough to provide two. But when I cut it out I was not really thinking about saving the rest, and so I cut the block out of the center of the blank, partly to avoid a small knot, but mostly just because I didn't think I needed to save the rest of it. Oh well, time to reach for the re-cycled maple hardwood flooring and quickly glue up another blank. Just for fun I slipped a piece of scrap cherry in the middle. In time that will darken up (you can't really tell right now) and give me a racing stripe in my block.

Drilling the mortise for the carriage bolt that clamps the block closed.

And that is much better. Time to make another big gear!

I printed out the 48-tooth gear plan, stuck it to some Baltic Birch, and started work. I forget why, but I got the idea that this was a good size gear to start with.

Later on, I found a section of the build video where he explained a bit of the gear ratios... The threaded rod is 16 tpi (threads per inch), so each turn of the rod is a 16th of an inch in travel for the stock holding box. Using a 48-tooth gear with the larger 16 tooth gear on the shaft gives a 3-to-1 ratio, and with the 12 tooth gear a 4-to-1 ratio. (48 divided by 12 gives 4.)

so if you use a 48 tooth gear on the lower setting, so it engages the 12 tooth gear on the shaft, then with each turn of the 48 tooth gear, the 12 tooth gear spins 4 times, which yields four times 1/16th of an inch, or one quarter inch of movement of the stock box. 1/4" fingers in my box joints is what I'm aiming at to start, so that works well.

But back to the cutting of the gear. I first tried drilling a few holes for the back of the gears, as you can see in the photo, but I didn't like how it was working -- I mostly have twist drillbits. This would work a lot better with a brad point bit! So I cut it all out on the bandsaw. I tilted the table to the right for cutting one side of the teeth, and then I wanted to tilt to the left for the other side. My table does tilt a bit to the left, but I don't like to do that, since it involves resetting the 90-degree stop that is on the saw. So I built this little ramp for simulating the left-tilt. Works quite well.

In hindsight, I wish I had first taken the time to review his videos on making gears. (Link Here) I was worried that drilling the holes was affecting the teeth. In that tutorial + video, he uses a drill for drilling out the insides of all his gears. I now realize that the depth of the notch between each tooth does not really matter much. What matters is the spacing between each tooth. I had to do a fair bit of filing and fiddling to get my gears to mesh nicely. I'm still not 100% satisfied. If you build it, go check out that tutorial first!

On to the next problem...

I had the gears meshing well, I thought. But the mechanism was still not working well. The gears did not seem to fit properly, and they would jump out of the shaft gear's teeth all the time. I was very puzzled for a time... Eventually I realized that this was yet another problem with plywood thickness. Again, I used 11mm plywood for my parts, and I think Matthias was using 9mm. The plans say that either will work, but I don't think that is quite true.

Take a close look at the photo and you can see that the gears are not fully mated together. The piece of plywood on the gear block is 11mm thick, which I think is too thick. But I don't HAVE any 9mm plywood. I tried to rip it skinnier (see the chunk out of the corner?) My bandsaw blade just popped over and would not rip it thinner. I was scratching my head for a solution. I want this to be stable, so it had to be good quality BB plywood.

Fortunately, while I did not have any 9mm plywood, I did have some 19 or 21mm (3/4" equivalent) Baltic Birch. So I had a go at resawing that down to 9mm in thickness, which actually worked perfectly.

After resawing, and then some sanding and drilling the proper mounting holes, I now had a thinner mount under the gear, and the gears would now full engage each other. Whew!

I was now nearly complete. The next step in the plans is to use some 3/4" x 3/8" bar stock to fit into the miter slot of the tablesaw and fasten to the base of the jig. I don't have a tap-and-die set, and could not find one at a price I was willing to pay. So I waffled on this for a while, and eventually decided I would first give it a try using just hardwood for the runners. I've had hardwood runners on my Table saw cross-cut sleds for 10+ years with no worries so I thought this was at least worth the attempt. I also discovered some well seasoned tight grained white oak in my leftover pile. This is very hard, and very dense, and I was able to cut out some strips that were practically quarter-sawn with their grain arrangement. I'm hoping that these will be stable enough to serve. I did go ahead and put two of them on, one for each slot. The plans call for just one steel guide bar.

The final step was to cut out a handle and fasten it to the front, which brought me to the end of the build. Here are a few more photos showing it from all sides.

Testing It:

Now the jig is built, the moment of truth is the actual use of the jig to cut some finger joints. But first, let us digress into a bit of math.

On a woodworking forum I was writing about this jig and someone asked about errors. Essentially, since I am just turning a crank (the big gear), and there is nothing in the jig that makes the crank alwasy stop in the same spot, can't errors creep in? If I don't always stop the gear in exactly the same place, won't that affect the spacing?

The short answer is, of course, NO!

The more detailed explanation takes some basic math. Just a bit further up in this web page, there is a section where I write about cutting out a 48 tooth gear. As part of that I learned a bit of the math for how this thing works, and wrote about that there. Here then, is the second half of that discussion, which should explain why there really is not any worry about error.

It goes a bit like this:

I'm cutting 1/4" finger joints. I've got a Freud box joint blade set in my saw, set to cut a nice square cut exactly 1/4" wide.
I have a threaded rod that moves the stock box, which is 16tpi, (threads per inch). So each full turn of the rod moves the box 1/16th of an inch.
I have a 12 tooth gear (the smaller one) mounted on the rod, which I am using.
I have a 48 tooth gear mounted in the gear block, which is meshing with the 12 tooth gear.
This big gear has a handle on it, which I use for turning the gear. 48/12 = 4. So each full turn of the 48 tooth gear turns the 12 tooth gear 4 times.
Therefore, each full turn of the 48 tooth gear turns the threaded rod 4 times. 4 x 1/16th = 1/4.
Therefore, each full turn of the 48 tooth gear moves the stock one quarter inch.
Therefore, for making 1/4" finger joints, I turn the 48 tooth gear twice after each cut. That moves the stock 1/2". 1/4" for the bit just cut, and 1/4" for the next finger.

So, the question remains. I am just eyeballing this as I turn the gear. What if I'm off? What kind of error do I get?

First of all, the gear mounting block has a clamp in it. The 48-tooth gear is removable. (So you can put in different sized gears for different cuts.) So when I set up the jig to make my cut, I loosened the clamp, lifted up the gear, and then moved it and set it back down such that the handle on the gear was positioned right beside the secondary gear. This gives me a nice visual reference for stopping the gear each time I turn it.

But secondly, and more importantly, let's go back to math.

Suppose I do make a mistake. There are 48 teeth on this gear. Let's say I make a turn and am sloppy and am off by one or two teeth when I stop the gear, what happens? Remember, each full turn of the gear moves the stock box 1/4". We're dealing with small measurements here, so I think it is easier if we start talking about thousandths of an inch:

1/4" is is 0.25 inches, or 250 thousandths of an inch.
250 divided by 48 teeth is 5.2.
Therefore, each tooth in that 48 tooth gear, in this setup, is equivalent to moving the stock 5.2 thousandths of an inch.
So if I'm off by two teeth, that is 10.4 thousandths of an inch. One thousandths of an inch is .001"
10.4 thousandths is .0104" which is less than 1/64th of an inch (.0156") --- It's actually about two-thirds of a 64th... Ummm, that is a bit under 3/256th of an inch. (I'm the one doing the math, and I find 3/256th to be a silly fraction!)

So, in my example, if I am off by turning the gear two teeth, it is only going to make a difference of less than a 64th of an inch. I don't think that is really going to make much difference. (And I have a hard time imagining being off by more than two teeth when turning the gear.)

PLEASE REMEMBER: This little bit of figuring is based on my current setup. I am using a 48 tooth gear, meshed to the 12 tooth gear. A different setup is going to result in different amounts of movement.

Now that the jig is built, it is time to test it. I did not take pictures during my first cut, but I did during my second test cut. I took several shots, which I hope does a good job of explaning how to use the jig. As well, there are several videos on the woodgears website that show the jig in use.

To start with, I want 1/4" finger joints, so I put in a Freud Box Joint blade set. (That was the most expensive part of deciding to build this jig!) Blades like this are not mandatory, you can make box joints with a regular dado set, or with other blades. What you really want is a blade that leaves a flat-topped cut, and my dado does not.

I layed a piece of stock down, and set the blade height to be just a touch higher than the thickness of the stock. My goal is to have fingers that are just a bit proud of the joint, which can then be trimmed flush.

Next, stand up the two pieces for the test cut. If I was building a box, I would have all four piece in at once. Snug up the pieces in the corner of the stock box, and turn the crank until the pieces are just touching the side of the blade.

One optional step here is to loosen the knob that is holding your gear in the gear block, so that you can lift it up and turn it such that the knob (or some other visual reference mark) is lined up with the two drive gears on the threaded rod. This gives a nice visual reference for lining up the gear as you turn it.

I am making 1/4" finger joints. So I need to turn the 48-tooth gear one full turn to move the stock box 1/4". Then carefully slide over the front piece of your stock, while leaving the back piece untouched. And then clamp them securely in place.

Honestly, this was one of the most difficult bits to get right, as the two pieces wanted to slide together. I would think it would become easier with practise, and also if I was using heavier hardwood stock.

Turn on your saw and begin the cutting process. Push the jig with the stock through the blade and back. Then turn the gear the required number of turns and repeat until done. Again, I am making 1/4" finger joints, so I need to move the stock double that -- 1/2" -- each time. So I turn the large 48-tooth gear, which is meshed with the small 12-tooth drive gear, two full revolutions.

And here is the result.

As an aside, I do not like the thin finger on the right edge of these pieces. However, for utility shop projects, I probably would not care. As well, these were just random test pieces that I grabbed from the scrap pile. For ideal finger spacing, I would want to use stock that is sized to be a multiple of my dado width, which is 1/4".

The joint slipped together perfectly. It was like magic. No sanding, no chiseling, no fiddling. I just slipped it in, and tapped it together. Note that this photo is actually of the very first test joint that I cut with my jig.

And that is pretty much the end. I think I'm going to really enjoy using this jig, and I can forsee a lot of use of it in future projects!