Home Articles Forums Gallery Marketplace Videos

WoodCentral Forums

Knowledge You Can Build On — Est. 1998

Compare cordless drills on the numbers that matter

fitting dovetails question...

Posts

fitting dovetails question...

#1

fitting dovetails question...

Bill Tindall, E.TN

>Yes, I know you are supposed to saw them perfect and fitting is unnecessary. At my skill level I can often do it if the situation is compressible poplar tails into rigid maple pins, for example, like on a typical drawer. But this trunk is oak into oak and this oak is not compressible. I might as well be trying to fit brass pins and tails. In fact, maybe I should be asking the infill plane makers!

Anyway, with great care, since they will be on a trunk corner for all to see, I have the pins and tails sawed and they are crisp........and in most cases a few thousandths of an inch from going together. In most cases I can just make out the remains of the scribe line on the pin top, so I am less than 0.010 from perfection.

Let me preface the questions by saying I have never had great success in diddling with pins and tails to make them fit. What once was nice crisp sawed pins and tails get rounded and out of shape as I attempt to remove material by various means.

Questions: 1.What is the best strategy for fitting- remove material from pins or tails? I think pins as they seem easier to get at with the wider spacing between them. 2. What is the best way to remove a few thousandths of an inch of material so the pieces will slip together. 3. In a row of 14 fairly closely spaced pins, what is the best way to locate some that are too big?

Re: fitting dovetails question...

#2

Re: fitting dovetails question...

Dave Mount

>Hi Bill--

With the precision with which you work (don't give me any of that crap about your work being sloppy) I'm guessing that all your sawcuts are square and that the problem is truly that the pins/tails are slightly oversized. If so, my suggestion is to do your exploratory paring in the inside edges of the tails. I appreciate that you may have less access there, but that part of the joint doesn't show when it is assembled, so if you overpare it will not be visible, except at the very outer edge. So I would start adusting the fit on the inside side of the tailboards, doing very little at the far end (the broadest part of the tail). Fussing around this way you should be able to get the tails at least started onto the pins, and that should help you identify where trimming is needed the most. Just be careful you don't split the half pins off the edges.

In terms of determining where the poor fits are -- one thing you can do is mark (or re-mark if you're a pins first guy) the tails from the pins using a pencil. It'll be subtle, but I'd guess you'll be able to see which pencil lines are more than a pencil line away from the cut. I use a 0.5 mm mechanical pencil for this because it marks very close the to work. Another possibility is a dial caliper. Before I get a hail of nasty comments about not needing 0.001" accuracy, I recommend it not because of the accuracy/precision per se, but because of how well designed they are for measuring both inside and outside dimensions with ease (OK, and accuracy). I used to be someone that scoffed at dial calipers in the woodshop. Then I bought (a pair of? is one tool a pair?) dial calipers, as much as anything to have the standing to say with authority that they were not that useful. Though I was a doubter, my experience was just the opposite -- I cannot believe how useful they are, they now live on the most accessible shelf to my work area.

Best,

Dave

Re: fitting dovetails question...

#3

Re: fitting dovetails question...

MikeW

>Hi Bill,

I often used Bubinga and While Oak, so I know pretty well of what you are going through.

I don't know if you have slightly beveled the inside of the tails starting just below their tops and ended alomst at the baseline or not. I do as even a slightly wide pin will begin to enter before things come to a too tight situation.

I then look for the slightly glazed/compressed ends of pins. When I do have to fit DTs in such a hard wood, I typically use a chisel a little less wide as the side of a pin. I pare as thin a shaving as I can stating below where the pin will protrude--if your pins sit proud after assembly, do account for that--and pare that thin shaving down to the baseline. The chisel I typically choose to use has about a 1/16" clearance on each side.

This leaves the basline and top of the pins alone and I find that even in hard woods, those edges will begin to compress and fit. However, this only works when one is still pretty darn close to a fit off the saw.

At times my pins have looked like a drunken sailor has cut them, and I can see a pretty good bit of meat on the waste side of the marking line, I do pare to the compression marks written about above if I can even get them to start once the inside of the tails have been lightly beveled.

If I cannot even get them started, I first pare the entire width of the pins where I can see meat on the waste side of the knifed line.

I line up the chisel next to the knifed line on the waste side in the off chance it has been mismarked (I never trust tryong to line things up and remarking), I pare slightly beveled--so the chisel runs out somewhere on its way to the baseline and check and pare as needed once they begin to go together.

Don't know if any of the babble above makes sense. Hope so.

Take care, Mike

Re: fitting dovetails question...

#4

Secondhand information

Bill Houghton, Sebastopol, CA

>Someone or other at some workshop I went to years ago said you could tell the tight pins by a different sound when you tapped them with a hammer. Can't remember what the sound difference was supposed to be now.

Or maybe I'm just constructing a memory...

Re: fitting dovetails question...

#5

0.001 accuracy....may be needed

Bill Tindall, E.TN

>That is not far off when fitting oak to oak. Fits that I could push together if this was a poplar drawer can't be hammered together. Did I mention incompressible! The whole assembly rings like I was tapping metal together. A scribe line width left on the pin, which can't be but 0.005, makes the tail too tight to fit. The English worked oak for furniture. they must have developed some approach to fitting it (maybe 5 years apprentice!).

I have digital caliper and used it constantly for all manner of handy measurements. I would recommend everyone owning one. As an indication of how often it gets used, it sits in the front of the drawer by the bench vice. For example, a precise caliper enables screw shanks to be measured precisely such that a number drill can be selected for just the right sized pilot hole. I haven't wrung a brass screw off in years. I'll try it for finding fat pins or fat tails as the case may be.

Re: fitting dovetails question...

#6

very helpful

Bill Tindall, E.TN

>I'm not off much, not more than the width of the scribe line. But as you know in these wood it might as well be the width of a pencil line in terms of not going together. thanks

Re: fitting dovetails question...

#7

Re: fitting dovetails question...

Derek Cohen (in Perth, Australia)

>Bill

I think that you have already been advised to chamfer the inside edges at the rear of the tails before fitting?

Then further to all that has been recommended, when finally fitting everything together (and hoping it is still a tight fit), add a clamp across the tail board to prevent splitting as the two pieces join together.

Regards from Perth

Derek

Re: fitting dovetails question...

#8

Re: fitting dovetails question...

Marv....... Northern California Lake Co.

>Hi Bill,

For short distances, soft metals can be press fit, meaning that there is an interference from zero to .001 and sometimes more. Harder metals are less forgiving. I consider dovetails as short distances. The amount of surface to surface contact has a lot to do with it as well.

Most wood types will compress and is far more forgiving than metal. Tighten down on a piece of oak with a C-clamp and you can easily leave a permanent indentation.

I would cut the pins long, at least 1/4" and taper the ends so they easily slip into the mating piece but no more than the 1/4" taper. This will allow you to get the dovetails at least started. Once fitted together, remove the excess. Attempting to get each and every dovetail exactly to a slip fit can drive ya nutz.

Wood fiber will relax under pressure or compression. I recently put a new handle in my 12 pound maul. When I first fitted the handle in the hole, it was super tight. For several days after, while I was using the maul, I would tap the butt end of the handle on a stump. (I always do this for awhile before using wedges) Each time I did that, the head of the maul would move a little farther up the handle. The handle is now sticking out about a half inch. The handle is Hickory, at least as dense as Oak. The point here is, after you press the tight fitting pieces together, after a day or so, the wood will relax a bit and they will not be as tight as you thought.

My preference when fitting any piece of wood to another when a close fit is desired is to make them slip together easily. When glue is applied, the wood will swell a bit. The glue will cause the wood to remain mostly in the swelled state. Even with a slip fit, I would still taper the ends of the pins or tails, depending on how you think of it.

Marv

Re: fitting dovetails question...

#9

.001" Accuracy

Steve Elliott

>At one of my former shops I built dovetailed drawers out of maple using a router-based system. The fit of the joints was determined by the height of the router bit and I found (using a dial micrometer gauge) that the height of the bit needed to be set within .001 of an inch.

I have great admiration for those who can cut dovetails by hand and hope to develop that skill myself after I retire. I have samples of hand-cut dovetail joints that I made at various stages of my career, starting in 1978 when I was still in training. The joints get better over time but I never came close to the speed and accuracy necessary to make them a viable option for work that I needed to make a living from. In ten years or so I hope to "graduate" to a life phase in which I can focus on the kind of fine hand craftsmanship discussed on this forum.

Re: fitting dovetails question...

#10

outstanding idea

Bill Tindall, E.TN

>That is a tip I never saw before. While I have not split a half tail or pin recently I must admit it was not uncommon at first and I still fear it.

Re: fitting dovetails question...

#11

Re: .001" Accuracy...Hogwash!!

Marv....... Northern California Lake Co.

>Having worked as a machinist for many years of my working life, to hold a tolerance of plus or minus .001 required some of the most sophisticated high end machinery used in an a temperature controlled atmosphere just so the machinery wouldn't contract and expand that much. To hand cut wood joints and expect that kind of accuracy is unbelievable. From a practical and logical standpoint, when working with wood that can move in all directions by a factor of ten beyond a .001 tolerance from temperature change is a pure waste of time and expectation.

If you were given a drawing of a dovetailed joint with tolerances of plus/minus .001 for each dimension and told to hold those tolerances, you would end up with nothing but a pile of scrap that most people couldn't even measure to that degree. It isn't practical even with a digital micrometer or liniar caliper and a height gage.

Come on folks, lets get real.....

It's the wrong message to leave with a beginning woodworker, to make him/her think to become a good woodworker, they will have to get so they can hold tolerances of .001 and to buy the sophisticated tools to measures those kind of dimension.

Marv

Re: fitting dovetails question...

#12

Re: Secondhand information

Denis Ch�nard, Orl�ans, Ont.

>Not sure about the sound, you're probably right, never really paid attention, but the "feel" as one is trying to drive in a recalcitrant tail/pin section is definitely different.

I use a scrap with a hammer to drive the joint home, and that method tells you right away what area is binding (the joint might just pivot around that point).

Usually, unless there's a lot of resistance I just drive the joint home, with glue, no pulling out to pare again (same method that Cosman uses). With over a hundred joints done I've had only one split, and all the joints were with the same wood, either birch or maple. I also use a pencil for layout and this doesn't cause a problem. I leave the line, and just a smidgen more.

DC

Re: fitting dovetails question...

#13

partially missing the point

Bill Tindall, E.TN

>The issue is how much gap is ideal for fitting the mating surfaces for, in this case, a dovetail. This issue is some different from holding tolerances. Joints are typically fitted, and when "right", assembled. We all know that a joint made today may not fit right if assembled later.

I have at hand a study published in Furniture Design dealing with tolerances in mortise and tenon joints. A tenon 0.000 to 0.002" smaller than the mortise is considered perfect. It is implied that factories strive for such a fit. Joint strength begins to degrade in the range of 0.003 to 0.005". By my own measurements a tenon 0.005"under is beginning to feel loose. I don't question one's ability to feel a difference of 0.001" in a joint made in a wood that machines well like maple.

I doubt Steve meant that 0.001" was the tolerance required. On the other hand many of us would find a joint loose by 0.010 unacceptably loose. A joint 0.005" too big in a wood that resisted compression may not go together with out risky force. I never made good M&T's until I learned to fit them by planing shavingings from the tenon till the fit was right. I could never mass produce them on my machinery, for all the reasons you imply. But by hand fitting I am sure my goal is less than 0.005" under for the tenon.

When it comes to sawing dovetails, that is another issue. Unlike a tenon, there is no easy way to use a machine like a plane to remove a reliable shaving of some predetermined thickness. I think in most cases we rely on compressibility to get dovetails to fit snugly. That is certainly my case when I use poplar for a drawer side. I certainly can't saw to 0.005" tolerances, especially so repeated 14 times down a board.

Re: fitting dovetails question...

#14

Re: the real answer

ThomD

>Mostly when I cut them, they do come out perfect. One thing I do as a check or if problems appear, is go in there with a little straightedge and try the surfaces to make sure they are correct. I do this before assembly. The "straightedge" i use is a small Japanese keyhole saw, the type with the fine flexible blade. It works for trying the surfaces but also as a rasp, and to cut away little fiber that I may have missed.

There are available cheats. I don't use these moastly, I'm just trying to be helpful. Most of the cheats do not make for worse structure so there isn't anything wrong with them except some take more time. I believe that if one's work sucks it is better to cut and chop to the line but make the job easier with cheats, than to do bad stuff one won't learn from and roll around in it. Some cheats are:

- contrasting woods

-undercuts in the endgrain. Use the marking gage method to get your bearing surfaces then the undercutting of the rest won't hurt the structure. If the MG stuff stops your chisel diving you aren't going to need to worry about structural bearing.

- pare away some of the hiden corner on pins and tails

- concaving pins

- with correct clean-up of the overhang you can hide quite remarkable gaps, as in Krenov.

There must be something else. I;m real woosy after some bad seafood...

Re: fitting dovetails question...

#15

Never believed in Cosman myself...

Scott Burr in Ben Lomond, CA

>But I do believe in Kingshott. It's easier to cut a DT joint than a M&T.

You only nail them together once, otherwise you will loose the strength and integrity of the joint.

Re: fitting dovetails question...

#16

baseline

Adam Cherubini

>Bill,

Unless you made an error sawing (sawed the wrong side of the line for example), I would not recommend paring the mating surfaces. This is just asking for trouble. To pick up a uniform .005 of relief, work the baseline. This approach has several advantages:

1) it produces a uniform increase in slop.

2) because you are sliding wedges apart, it doesn't take much to fix the problem. You may have success using a saw file

3) you can easily remark the baseline and have an accurate line to work to

If you use white glue it's going to swell the wood. So you might want to take that into account. Personally, I don't think its a great idea to either use white glue or to beat these together like metal. I prefer the joint to fit more loosely and I prefer hide glue.

Good luck with your project.

Adam

Re: fitting dovetails question...

#17

Do I understand correctly?

Bill Tindall, E.TN

>Situation: a line of dovetails where most of the joints are too tight.

Solution: remark the tail base line and pare to it. This operation would thereby open the pin holes. Is this what you mean?

While this fix would not be suitable for a case where only some of the joints are too tight (it would leave some too loose) it would fix my problem nicely.

I may not have made clear that my problem is simply that most of the pins are too big by about the width of a scribe line. For oak on oak there is not enough compression to enable the joint to go together. The sawing is straight and orthogonal.

Re: fitting dovetails question...

#18

Hogwash!!

Steve Elliott

>My reference was to the height of a dovetail bit in a router-based system, and I found after repeated batches of drawers that in maple a bit .001" too high produced a joint that was too loose, while one .001" too low produced joints so tight that the corners of the tails would break when the joint was tapped together.

You're right about not wanting to send the wrong message to beginning woodworkers. Traditional joints like dovetails can be made without measuring at all. Their fit, though, can be sensitive to very small differences in dimension.

At my current job I regularly fit wood parts to already-fabricated metal parts made by welding. The tolerance held by our highly-skilled welder is around .030" but the fit needs to be within .005". I measure the metal using a template of known length and feeler gauges, then trim the parts on a CNC router that has accuracy of about .002" plus bit deflection. The resulting piece sometimes needs adjustment using planes, which are slower but much more accurate.

Re: fitting dovetails question...

#19

Re: partially missing the point

James Watriss

>Bill, much respect, but I think you're obsessing over the numbers a bit much. And a dovetail is much different than a M/T joint. AND a joint will swell a bit sometimes when you're using yellow or white glue. So the numbers all lie.

One of the first things I'd check is flatness of the board. On wider DT joints, it's not unusual to find that one board has cupped just a wee bit, and that this is really messing with the mojo.

Unlike some, I won't take anything for granted... use a 4-6" ruler (or the flat of a chisel) and hold it against the sides of the pins. You'll be able to see by eye if the sides of the pin are out of square, by even a little.

If there's only small amounts, and you're really just trying to fuss this together, then I guess the old standby method is probably best. Use a soft pencil on the end edges of the pins, and pare, file, or sand off whatever gets rubbed onto the spots in between the tails. Or, pencil on the mating edge of the tails... whatever floats your boat. Pencil marks go on the edge of whatever you know to be square, to be rubbed off on whatever needs to be pared away. It's a little time consuming, kinda tedious, but if you're trying to get it "just so," then it works. Do this with boards clamped flat between cauls if there are flatness issues. It's similar to the compression mark idea, but there's not as much force required.

It's an old standby technique, it can be a hassle, and the slow going ca make you feel like you're a beginner again. Truth be told, I think this is important sometimes... I'm pretty sure I was a lot more careful and attentive when I first started out. As I got more experience, I got more cocky, and I was worried about just getting everything done. Slow and easy, worry about the work, not how it makes you feel. There's nothing wrong with going slowly if you're in a fussy mood.

Don't sweat the numbers. You're not a machine. Get things just so, and if there are goofs along the way, that's an entirely human and understandable thing. Not every antique has DTs that are wholly gap free. Keep the end goal in mind... you're building a chest. DTs are just a way to join the boards, and even if it's not 100% machine-tolerance perfect, the joint will hold.

Re: fitting dovetails question...

#20

I disagree with both points

Bill Tindall, E.TN

>"It's easier to cut a DT joint than a M&T." Well, I sort of disagree. I cut the tenon over sized and plane it to fit the mortise. By this means I get a perfect fit every time. If I felt obligated to simply cut and assemble, then I would agree. philosophy follows: My objective is to get stuff I want built. I don't have the ability or experience to do every thing the way a professional in the 1700's would have done it. And I don't fret this deficiency or eschew an opportunity to seek a crutch. I regard the trunk made with crutches better than the one never made for lack of 18th century professional skills.

The problem, for me, is that there is no equivalent way (planing to fit) of getting a perfect dovetail in a drawer or trunk. Ideally, they are sawed perfect to begin with and that ain't easy in oak. Editorial: I was in a shop that has a famous name and does the best work. I noticed that their sliding dovetails for drawer dividers were in fact half tails. That is because they can be planed to fit. I made note of this crutch for future projects.

"You only nail them together once, otherwise you will loose the strength and integrity of the joint." Again I sort of disagree. If the joint is perfect it will slip together without compressing the wood. So it can be taken apart with out irreversible damage. Additionally, a water based adhesive will re-swell some compression. And more additionally, if the humidity goes up and down the joint is going to experience the equivalent several fittings.

Typically, dovetailing is done where at least one part of the joint was done with a wood that was compressible so close tolerances are not critical and compression damage is possible. Recall the Tage Frid parlor trick, performed with pine, not oak. The statement is a worthy goal, but I'll bet seldom attained by most of us making stuff. On second thought, it is a silly goal that will prevent some people from using dovetails.

Re: fitting dovetails question...

#21

to clarify

Bill Tindall, E.TN

>I'm obsessing over a too tight joint that has some otherwise perfect pins uniformly some number of thousandths too big. I am focused on the joint not the measurement. The measurement is just a quantitative estimate of the size of the problem and an indication of the precision required. If preferred, I could describe the problem in terms of pins uniformly too big by a fine shaving thickness.

Everything about the execution of the joint would have yielded a near perfect fit if the boards were poplar. But they are oak, and if you have never dovetailed oak into oak you may not appreciate its a different world than poplar and mahogany. It does not compress significantly and it is difficult to pare accurately through its alternating hard and porous regions.

The sawing is straight, the panels were clamped flat for marking, etc. In other works this time I took great care and I avoided all the normal dovetail screw ups and defects. The problem comes down to the fact that if I can see even the faint remnants of the scribe line on the pin (I do tails then pins), the pin will be too big. So, it is a fact that I must remove a uniform shaving width of material (if one prefers this measurement to 0.005" or whatever) from the too big pins (or too small tails), by some means.

James this is where I am at........."If there's only small amounts, and you're really just trying to fuss this together, then I guess the old standby method is probably best. Use a soft pencil on the end edges of the pins, and pare, file, or sand off whatever gets rubbed onto the spots in between the tails. It's a little time consuming, kinda tedious, but if you're trying to get it "just so," then it works. "

I need to get better at the "paring and filing" part so as to keep the pin face flat and orthogonal. I am finding that a new, very sharp file helps.

Re: fitting dovetails question...

#22

Re: to clarify

James Watriss

>Oh, yeah. I'm familiar with oak. That's why I don't use it as much. :)

Not trying to talk down to you at all, I'm just starting from scratch with what I know, because, well... it's what I know. Sometimes DTs fit, or can be made to fit, and there's no fuss. Sometimes they don't, and you have to suffer through obnoxious fussy work like this, and it's not fun, and there's not usually an easy way to do it. I've been lurking lately, and infrequently at that, and I'd rather point everything out that I can think of than assume you know about _____.

I hear what you're saying about trying to keep everything straight, flat, and orthogonal. The joy of the pencil technique is that you're not generally removing huge amounts of wood at a time... You're basically trying to scrape off the graphite from the pencil. Some wood will go with it, too.

I can relate to the stress of "Dammit, this pin board is perfect, everything's straight and flat and square, if only..."

Sharp tools do help. Files, chisels, etc. I know it's intimidating to go in and mess with a pin that's straight. So it may be worth leaving the pins be, and go to work on the tails. Or not, it's your call. I you have a thin enough chisel, it's not so hard, it's just fussy.

If you have scraps, it might be worth cutting a few DTs, and practicing the paring bit until you're feeling happy enough to come back to the finished piece. It might be less nerve wracking if practice left you feeling more confident in your ability to pare and shape.

Re: fitting dovetails question...

#23

Re: Do I understand correctly?

Adam Cherubini

>I think you have my meaning. The pins are wedges that fit a socket (tails). It's not unlike a morse taper, Bill. The fit is effected by the penetration of the wedge. If the angles are correct, why touch them again?

If nothing else, make sure that sure the baseline is straight.

The other thing you can do is plane the inside of the pin board. Sometimes a single pass straight across the end grain with a smoother is sufficient.

It's funny how much talk there is about dt's and rarely a mention of the relationship between the baseline and how you mark or saw. Most times, I mark the pin board as if there is no gap at the baseline. I mark, then saw out the mark. Theoretically, this should result in gaps. I should have left the mark theoretically. But the fit is always tight because there is always some gap at the baseline. And that pushes teh wedge out and makes it tighter.

I say work the baseline and don't pare the mating surfaces unless you have a local problem.

Adam

Re: fitting dovetails question...

#24

Re: Hogwash!! Sorry, missunderstood....

Marv....... Northern California Lake Co.

>Hi Steve,

Sorry, I think I related your .001 adjustments to holding those kind of tolerances in wood. Making an adjustment of only .001 is not easy, is it? Even with a digital micrometer. Tilt it a bit one way or the other and you are up to .005.

When you go from plus .001 to minus .001, you have a difference of .002. But even .002 should press together in Maple. The problem is probably accumulative when you have several dovetails. With just one dovetail it would be tight but wouldn't break but if it builds up to .010, yes, that could cause a crack or just way too tight. I'm thinking it might accumulate, but is it actually possible with dovetails? It feels like we are over analyzing this issue. Like you said, just mess around until it fits.

Marv

Re: fitting dovetails question...

#25

thanks for the commizeration

Bill Tindall, E.TN

>It helped my frame of mind. I need to be more bold on the next set and saw the scribe line away. I practiced a bunch in oak before starting but chickened out taking the scribe line out when I came to the real pieces.