Okay, so what am I doing wrong for box joints?

Folks -

So, I built the box joint jig in Shop Notes from a while back, but damned if I can get the spacing to work out. I have a Freud Dial Dado Set (works GREAT in laminates and ply, so far) and I have measured the thickness of the stack at 3/8 and the space between the right edge of the dado stack and the register pin - also 3/8... Yet after make a set of test cuts of more than about 6" pins and sockets are out of whack enough that I've just made some dentil molding kindling...

I have measured everything with calipers and over about 8" the dado'd sockets get progressively wider, on the order of .015. I kept a tight hand on everything while cutting, forcing the work against the registration pin, so I am at a bit of a loss. I know that the pin spacing and dado width have to be spot on, as any error is going to accumulate. The whole process seems to be alot fussier than it needs to be...

Unfortuneatly, I've disassembled the dado stack, so I can't find out if the stack "opened up" somehow, or what. I guess I'll try it all again in a bit, but wanted to see if rec'rs had any suggestions.

Thanks in advance!

John Moorhead

Reply to
John Moorhead
Loading thread data ...

Are you starting either the sides or the back and front with a spacer block set to you dado width? So you have one set off set from the other.

Chris

Reply to
Chris Melanson

Chris -

I know about the spacer, I haven't even made it that far.... I am still working on getting the test cuts trued up and having fits...

My neighbor suggested box nails or corner brackets. Damn! Here I go spending all this $$ on tools and I could have had the Ace Hardware guy point me toward the bracket isle. Betcha this was covered in that most secret of societies... HCA!

John

Reply to
John Moorhead

Hum...Are you measuring the teeth or the slot that the teeth cut?

If the width of the dado cut out of the wood gets wider, then, it HAS to be that the dado blade is "self adjusting" as the cuts are made. After all, 0.015" is slightly under 1/64"...and it would not take MUCH to get to that. If the registration pin were moving, the width BETWEEN the dadoes would be increasing. I believe, though, that if the dado blade and the stop pin are not at the "same" setting, that you will not end up with progressively wider errors...but, a set of fingers that don't match the width or alignment of the slots they are supposed to fit into. that, at least, was my experience a bit ago.

I am not entirely sure that anything but a stack dado (or router) will create finger joints without a bunch of fiddling and trouble. I would recommend trying this...Set the blade to its minimal thickness...say...1/4", and set the finger joint jig to the same width. Cut a sample piece. If it shows irregularities, then, perhaps it is a problem with the jig flexing. If it works fine, then, perhaps the problem is with the adjustment of the blade walking. Alternatively, set the jig to 1/8" and use your regular saw blade. If all is "ok", that should produce perfectly fitting joints. Of course, the problem SHOULD only show up if you have adjusted the thickness of the dado by adjusting the dials away from zero. Sorry if I seem to be wandering a bit...I am kind of puzzling over this, and producing a stream of consiousness note on things I would look at.

Reply to
Dave Mundt

Try setting up your dado blades make a test dado. Then cut a strip of wood to fit your dado cut and use that as your spacer block from your dado blades and then attach your index pin on the opposite side of your spacer block then the dado blades and then everything should work. It sounds like your distance from your index stop is larger or smaller that your dado width they have to be the same distance and width or you will get some creep in your slots and pins.

Chris

Reply to
Chris Melanson

Dave -

I measured the width of the dado head, (.375) then used the other end of the caliper to set the alignment pin one "dado head" width away from the right edge of stack. Now, the pin itself may 3/8 or narrower... I assume... as long as I keep the work hard against the pin ( and I (*was* careful to do so) everything should be okay...

I do like Chris' idea about cutting a sample dado, and then a pc of stock to fit, one "cut width" away... That would certainly eliminate errors creaping in from my measurements...

Now, about accuracy - it seems that the adjustment of any box joint jig would have to absolutely spot on.... If your spacing was off by even 1/2 of a thouandth, by the time you can cut a dozen "step and repeats" you could be out by to 10 thou, a noticeable amount, especially since I'd be that .0005 tolerance would be a pretty high standard for WWing.

Gotta be an easier way....

Reply to
John Moorhead

Chris -

My slots and pins both start out at .375. In two sample cuts, the slots grew progressively wider by .010 to .0150.

I like your idea about using a strip that fits the dado cut, etc.... I'll try that in the morning...

Thanks!

John Moorhead

registration

Reply to
John Moorhead

There is. See my post about box joints that gets rid of the accumulation problem. And laugh while you read the responses that tell me there is no accumulation problem with the method you're using.

formatting link
pictures.

Reply to
Bruce

First off, You mention that your slots are getting wider. Are your pins getting narrower? If the pins are getting narrower also, it sounds like your indexing pin may not be exactly the width of the slot being cut. It has to be exactly the width of the slot so that the work will not slide left or right. Also your indexing pin must be exactly the width of the dado from the blade. If only your slots are getting bigger, it sounds like your dado set is not staying properly adjusted.

Reply to
Leon

First, the best way to do woodworking is to fit the pieces to one another, not by measuring with micrometers on tools.

This is why I prefer the router jig to the tablesaw, which works on the same principle, but allows me to set my distance to the fence (pin in tablesaw jigs) by cutting and fitting any two pieces of scrap. The router jig taps the fence closer to make a narrower pin (looser joint) and farther to widen, and I don't have a thing to remake. These guys sell one in slick plastic, or you can make your own.

formatting link
If you're interested, reply here and I'll drip you some pictures of the jig in operation that I use to teach middle-schoolers.

As to "accumulating" errors. You are making a pair of pieces. If you accept that, you can see that the errors of one are compensated for by the other. Plus, you are working with a material which allows a bit of creep, but also correction, based on compressibility. The one thing you must do is ensure that you have a pin equal to or slightly less than the gap, which is your dado or bit width. I find this easier to achieve with the router table than the tablesaw,though I've used both. I have even, a long time ago, used the spacer method, so it was neither recently invented, nor, given the popularity of non-spacer jigs, is it in any way superior to them.

Oh yes, if you persist with the saw jig, remember that you don't need to vary (or measure) anything except the distance from the groove cut in the backer board to the part of the pin closest to it. Best jig I had was one that used a brass pin, not a full tooth.

Reply to
George

Well, I do prefer to cut a sample dado then use THAT to adjust the width over to the pin. I just don't trust the dado head not to either splay out, or suck in a bit when it actually starts spinning and cutting.

Spot on there!

Well, this is a point. However, I have cut a fair number of box joints in my time, and, while most of them have been on pieces of wood 4" or less in width, they have worked very well. They are VERY strong joints, and, look pretty good (IMHO). Now...one thing that I did that helps quite a bit is that I standardized on a couple of widths (1/8" and 1/4"), then, built a jig for EACH of those measurements. I had tried the adjustable route in the past, and, while it worked ok, I always ended up spending about half an hour making a test cut and adjusting the jig over and over. Frankly, for pieces much over 4" or 5", I tend to use dovetails anyway, so I did not run into the progressive errors you were hitting. Good luck! Dave Mundt

>
Reply to
Dave Mundt

ONLY way for the SLOTS to get wider is the DADO blade changing it's setting. Concentrate on the DADO blade, the problem you describe is NOT a problem with your jig

John

Reply to
John

Why don't you just gang both parts of the joint together - offset by a pin width, obviously - and cut them both at the same time? Any accumulative discrepancies will cancel out.

Cheers,

Frank

Reply to
Frank McVey

Silly as it sounds at first reading, that error does _not_ accumulate in a way that affects the fit of the joint. It may affect the cosmetic appearance, but not the _fit_.

The 'critical dimension' is the measure from one side of the register _pin_ to the same side of the saw blade (i.e. left side to left side) this is the spacing between successive cuts, "whatever it is" in actual dimensioning units. If you want 3/8" fingers on both pieces, and get the 'pin left'-to- 'blade left' dimension at 0.8003" -- well they're going to fall at multiples of .8003". If the blade is cutting out 0.375 dados, then you *DO* get an un- equal amount remaining, 0.4253".

At which point, you have two options: 1) set blade to make a _wider_ cut -- to wit, 'half the difference' between what it is cutting off, and the 'wider' part that is left. then make a 2nd pass. You'll now have something very close to .400 cuts, and .400 pins. Not what you were originally aiming for, but 'equalized'

2) The other option is "don't worry about it".

When you go to making the mating part, set the blade so it cuts the exact width of the 'unequal' tooth in the first piece. that '0.4253"' dimension.

** IMPORTANT: ** ========== ** Assuming you've designed your jig so that the register pin is on the ** 'nut' side of the dado blade, the 'critical dimension' (from the ** _fixed_ side of the blade to the near side of the pin) is *UNCHANGED* ** by changing the cut width. i.e. it is _still_ that 'silly' 0.8003"

Now, some test cuts to get the dado with set right (0.4253") and you start cutting. Every 0.8003 inches. And you're chopping out .4253" sections, leaving a 0.375" tooth between each one.

Which happens to _exactly_ fit against the 0.4253" 'tooth' and 0.375" gap in the opposite piece.

Miscellaneous requirement -- you've got to process both pieces starting from the -same- logical edge of the joint. i.e, from 'top' to 'bottom', or 'front' to 'back', or 'stage left' to 'stage right'. Whether you go 'top to bottom' or 'bottom to top' does not make a difference, as long as you're consistent about the handling of both pieces.

Note: only one side of the register pin is 'significant', thus it _can_ be made very thin. A _metal_ one, that is thinner than the kerf of your regular blade offers some real convenience for setting up.

Set the 'critical distance' to what you think is right for the joint you're going to attempt. grab a piece of _scrap_ stock (as big as the joint), and run a full series of test cuts. If things are 'exactly right', you end up with 'half the critical dimension' from the last cut to the far end of the piece of scrap. If you actually have more (or less), adjust the 'critical distance' and try again.

ONLY after you've got that 'right' do you need to concern yourself with the dado blade. set it to 'half way', and "Awaaaaaay we go!"

Only if you don't understand the process.

_Effective_ woodworking is -not- 'science'.

You can -start- with science, but, in general, what is 'really important' is that the things 'fit together', so you 'build to fit'. :)

Once you understand that, and approach matters from the standpoint of 'making them fit together', things look a _whole_ lot different.

_What_ the dimension is (in some arbitrary 'units of measure') is not nearly as important as the dimension being "right" -- i.e. that it fits where it goes.

It's the same principle as underlies the use of a 'story stick' rather than relying on written dimensions and a tape measure.

There are several things that 'could' have been happening. Your description isn't detailed enough to identify which one.

First off, if you're measuring from the *end* of the joint, and noticing that 'everything' seems to be 'slipping more-and-more out of alignment' with where you _expected_ them to be, the answer is "don't worry about it", The fact that the fingers are every .753" instead of every .750" _doesn't_matter_. As long as the other side of the joint is built with the *same* jig setting, it _will_ mate up.

Second, as you measure from the 'left edge' of one finger to the 'left edge' of the next finger, does the measurement change? If so, either the register pin moved in relation to the saw-blade, or one was not getting a consistent fit against the pin. 'Slop' in the mounting of the jig to the saw table-top could also contribute. If the measurement _is_ constant, you can eliminate register-pin issues from the possible problems list.

Third, if the finger-to-finger dimension is constant, but the width of the cut-out is changing (with, obviously, a matching change in the opposite direction, of the finger itself), then the _blade_ must have actually been changing in the width of the cut it's making. Many possibilities here -- ranging from a slightly loose blade, to excessive play in the arbor bearings.

Comment: if a _visually_ 'exactly balanced' set of fingers is what you require, then the 'step and repeat' methodology is -not- appropriate. you need to use a 'divide and conquer' methodology instead. It takes a bunch more time/effort, but has minimal 'error' and that error is _visually_ evenly distributed across the joint, rather than showing at one end or the other.

'divide and conquer' methodology. find a 'simple fraction' of the entire joint that will have some number of complete finger sets (a notch and a tooth). Mark off those fractions (a half, a third, whatever) of the overall dimension. Repeat the process, for each of the sections thus created. Repeat again, as needed, until you have each individual tooth market out. Now, cut out all the marked notches. You do have to align, individually, each saw-cut to the marks.

Reply to
Robert Bonomi

Don't forget the miter gage in the groove. Slop there can certainly make a difference. Press right or left every time.

Love the router, don't you?

Reply to
George

HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.