The pursuit of Excellence in Fine Woodworking

Make An Accurate Box Joint Jig, Simple & Fast

The box joint, also known as a finger joint, is probably the most popular and widely used joint for drawer construction and box making. Aesthetically it has a very nice look, and structurally it’s very strong. Here we’ll show you a method that is fast, simple, and accurate. And best of all it costs next to nothing to make.

Specific Tools Used

Mitutoyo Dial Calipers
Forrest Woodworker Flat Top Table Saw Blade

14 thoughts on “Make An Accurate Box Joint Jig, Simple & Fast”

  1. Are you using the Forest Flat bottom box joint set or are you using multiple Forest Woodworker II #1 grind 10 inch blades?

    1. I use the Forrest WW II #1 grind which is the same as the box joint blade. For making 3/8 box joint I use their 8″ dado set. Hope this helps.

  2. thanks for your training.
    Been away from wood working for 40 years..
    thanks for taking the time to teach the techniques that make all the difference

    Trying some finger joints tonight.

  3. Why do you use a 3/8″ or .375 dado? a 4″ drawer side divided by .375 = 10.666 But you say and it appears to be perfect across the 4″ rather than having a smaller 2/3 piece at the end.

    1. 3/8 dado is actually .365 and the size of the fingers I make smaller which is around .360 If you add those up It will only be off by .01. I try to keep the video short and to the point and sometimes I have to make a judgement on editing. I hope this helps. Appreciate the comment and for watching.

      1. I followed your video as close as I could and ended up with a lone skinny finger. Now I understand why. But I’m happy how the joints slid together with that ‘just right feel’. Glad I intended my 1st try as an exercise in planning and execution and didn’t immediately jump into a large project. I wrote down my ‘actual’ measurements (slot in fence, spacer & distance from spacer to blade) and I bet my results would agree with your explanation above.

        Now that I know about your website/videos, I’ll follow more closely. Too bad I live on the opposite side of the country, else I would have already registered for one of your classes.

        Thanks you for the well produced videos and your sense of humor.

  4. A further question arises due to Bill’s comment. By offsetting the cuts to achieve a 6/1000″ greater width between the fingers the way that is done in the video, the 6/1000 added width on each successive cut is then CUMULATIVE, I believe. Using 0.365″ dado and 0.371″ fingers per finger & notch couplet, you are right in saying that this results in a couplet width of 0.365″ + 0.359″ = 0.724″/couplet. I think this means that as successive cuts are done, however, the overall width displaced increases in such a series so that with couplet’s width totaled, the overall widths will accumulate according to, just to begin, the following series: 1=.724″; 2=1.448″; 3=2.172″; 4=2.896″; 5=3.620″; 6=4.344″; 7=5.068″; 8=5.792″; 9=6.516″; 10=7.240″, etc. The error increases 0.006″ times the number of couplets, and we may note that as each successive couplet is cut, the .006 error creates a greater overall departure from the ideal series we might want to have (in the above, 1=.725″” 2=1.500″; 3=2.175″; 4=2.900″; 5=3.625″; 6=4.400″; 7=5.075″; 8=5.800″; 9=6.525″; 10=7.250″, etc.) Now there are n fingers and n-1 notches on the sides, and there are n notches and n-1 fingers on the front and back always under the method of uniform finger & notch widths proposed. so one must consider that fact as well. So, for the sides, the series above must be adjusted by a factor of +0.360″ for the additional half-couplet that is a finger, and the front and back must be adjusted for -0.365 for the missing half couplet that is a notch. Thus “perfect drawer side heights using this system become, and I will start at four couplets in the above assuming that few if any want to build a drawer less than 3″ deep!): 4=3.256″; 5=3.980″; 6=4.704″; 7=5.068″; 8=6.152″; 9=6.876″; 10=7.600”. Yet, even in this finger no. adjusted ideal series, there is no drawer height of exactly 4.000″, the closest being five couplets plus a finger equaling 3.980″, negligibly close, of course, to the 4.000″ value in a four inch high drawer as Dr. Ng stated, and is rightly here illustrated (Even “perfect” is relative in the realm of man! There is nothing “straight”, nothing “flat”, nothing “round”, etc. [If you don’t believe this pour yourself a “flat” slab of concrete, let it dry, stand at its centroid and slowly pour out a bucket of water directly downward–now watch the water wind its way across the slab finding any slight imperfections in its “flatness”. Do it again, day after day, and note the water always follows the same exact route to the edge of the slab. Imperfect reality is why engineering drawings always indicate to machinist’s and other craftsmen “tolerances”. No experienced engineer will use the words flat or round or rectangular or straight without them!] HOWEVER, depending on the drawers actual height and in treating other drawer heights than the nominal 4″ height, care must be taken to understand the above!. Of course, the critical finger is the last finger cut at the bottom of the (The reference mark being the top of the drawer sides.), and the question then arises as too how much “snaggle tooth” can be tolerated in the last bottom finger, a matter of individual conjecture and preference, BUT, I believe a better approach would be to simply know that there is not going to be a perfect fit for an integer number of couplets (or fingers or notches) adjusted as above unless one plans ahead by adjusting the actual height of the drawer (or box or whatever) slightly to a size that matches suitable closely one of the values in the stepwise progression demanded by a given combination of 3/8″ nominal notch and finger widths that are integer values. But, maybe I’m all wet and have made some error here–and you thought Dr. Ng was picking a few nits? Us engineers have killed millions of ’em!

  5. Using the heights of the fingers and notches that Mr. Ng has given above in a spreadsheet gives the drawer/box side heights that are matched by n notches and n+1 fingers of the widths he uses, noting, of course, that one side has n fingers and n-1 notches, and that its mating side must have n notches and n-1 fingers. There are specific set of side heights associated with a correct combination of fingers and notches, and while a four inch side height means that an error of only 0.001″ exists with a 4″ height, that will not be true for other drawer heights. With a 4″ drawer height, (5x(0.365″+.0.360″))+0.365″=3.990″, but with a drawer height of, say, 9″ then you get (12x(.365″+.360″)+.365″= 9.065″, 65 times the .001 at a four inch height. The solution, assuming you do not want to adjust the nominal ⅜” step box joint used is to simply to revise your drawer side height slightly to make it so it will be a match of the steps in height demanded by the notches and fingers to be used.

  6. I will add to the comment. I guess I got a bit obsessed with this problematic detail of the box joint, but in doing so, there are some interesting observations that can be made. The thing one realizes after fooling about with a spreadsheet solution to possibilities of combinations of whole integer sets of teeth and notches is that all that is really possible is entirely dependent on the ACTUAL WIDTH of the cuts you can make with your Dado set. Granted, they can be large depending on spacers, shim and chipper combination selected, but nonetheless, once you have set your sights on a 0.006″ clearance, meaning the teeth will be 0.006″ less in vertical height than the notches in your box joints, then, as Caesar said, “The die is cast!” The clearance plus whatever dada cut width you care to cut fixes a set of actual drawer side heights, and one CANNOT just willy-nilly think that, for an example, using ½” nominal dado cuts will result in you being able to build a drawer of, say, 10 inches in height using 10″/ 0.5″, or 20 even notch-plus-tooth sets on each drawer corner. It will NOT happen. You will end up with a “snaggle tooth” at either the top or bottom of each side of the drawer unless you are just plain lucky. The best thing to do, I think now, is to play with your dado making some trial cuts noting the practical configuration for each of them and the resulting width that you measure with your micrometer, then using those ACTUAL dado cut widths in conjunction with your picked clearance, run a spreadsheet determination of the actual side heights that can be made with them. I will throw out that there is another way to attack the problem of unbalanced teeth/notches in the joints, and that is to add a second “peg” to your box joint jig that is to the right of the blade matching the one William uses that is on the left. Cut each joint moving toward the mid-line from each side, so that the unbalanced tooth or notch is now becomes located at the center of each joint. This leaves both the top and bottom fingers and notches the same and adds one odd ball width finger or notch at the center of each joint that gives the appearance of being planned and regular since all fingers and notches will then be “mirrored” about the center line of the sides and front and rear of the drawer side heights.

  7. OR, you can build your peg holder so that you can flip it by extending the one peg though the back of the peg holding board using a spacer behind it.

  8. It would seem that you (especially you) could custom fit the spacer with a block plane.

    Your thoughts please


    1. Hi Ray, Sure, there are more than one way of performing a job. I just did it in a way everyone is familiar and comfortable with.

  9. William; I have seen a video online by Jody of Inspire Woodcraft about a simple box joint jig. It is truly very simple and accurate. For years I have used your method and relied on it for really great box joints. But this new way is a solid take off of your jig that really makes setup so simple and accurate. I really suggest you look at it and see if you agre

Comments are closed.