A Suggestion for a Research Project or a Thesis Topic
I have been unable to find even a rudimentary analysis of the strength of ordinary woodworking joints compared to the stresses those joints are intended to tolerate in various piedes of furniture. We can say with confidence that a glued dovetail joint is stronger than a nailed butt joint, but what does that mean in practical terms? When is a butt joint strong enough, and when is a dovetail joint necessary? I have some vague notion of cases in which stronger joints are needed (chair rails to legs, for example) and when less-strong joints are probably o.k. (coffee table apron-to-leg attachment, say). But I really don’t know if the joint I use is barely adequate or ten times stronger than needed.
I would really like to see a systematic study of the stresses on a range of pieces in ordinary use. A coffee table just sits there for most of it’s life. But if it is moved by shoving along the floor and the legs snag on a doorway threshold, then a much more severe set of stresses must set in. So I would like to be able to look up that condition in a table or something and find that the most severe stress is 8.0 on Brown’s Scale of Furniture Stresses. Then I would like to be able to look up, say, the strength of glued mortise and tenon joints and find that joints of the dimensions I have in mind have a strength of 63, so I don’t need to worry.
This sort of thing has been done in the building industry to a fare-the-well. The code books will tell you all sorts of things such as what kind of rafters to use for various structural and environmental conditions. I have seen a PhD thesis on the strength of timber-frame joinery. But I have not seen anything comparable for the broad range of furniture and the loads to which various items are subjected.
So you researchers out there, get busy with your grant applications. Tell them Don Brown sent you.
Replies
Don,
It's a fascinating subject to study. The question is would you approach it from the total design of the piece, or just from the 'joint' perspective? Is a tapered leg on that coffee table less threatening to the joint as it catches on the threshold? What does the double stretcher on a cherry shaker ladderback chair and the tilt mean to the joint on the seat....or the weaved seat for that matter, Impact?
Fascinating!
BG-
My favorite Einstein quote is (translated): "Make the explanation as simple as possible, but no simpler."
The manual I postulated would treat joint strength alone, clearly stating the factors that affect the strength (length of legs (lever arm) might be one, but I doubt if taper is a first-order consideration). If there are important excursions, leave that for the next grant. Modulus of elasticity, shear strength, fiber stress in bending, and all of those properties have already been studied for most important furniture species and are readily available.
The formula for the behavior of a rectangular beam under load is the model for my suggestion. Given that grounding, a designer can then at least have a solid starting point for exploring, for example, curved undersides and things like that.
Edited 8/26/2002 9:25:58 PM ET by Donald C. Brown
Don,
I'm sorry I can't offer a text title for you, but I have seen woodworking joints cover in some Strength of Materials text books. From there wouldn't what you are asking about be basic physics problems based on forces?
Don
Don C.-
I've not seen what you describe in the one Strength of Materials book I have on hand. But even if such exist, that solves only half of the problem. What stresses are woodworking joints subjected to in ordinary use? is the other half.
Don B.
Edited 8/26/2002 10:17:52 PM ET by Donald C. Brown
Don,
Did a quick google search and got this on biscuit and M&T joints. Unfortunately, it's all metric!
Biscuit Joinery Abstract, Page 1... large savings on working space, tooling, materials, ... enough for building heavier architecturalwoodworking, ... two different configurations to compare the strength ... http://www.woodworking.org/WC/GArchive98/ Abstract/abstract1.html
Don
Don C.--
Thanks for the reference. FWW, also, has published several studies of the relatve strengths of, e.g., biscuits, long tenons, short tenons, etc. But that's only one type of joint.
Maybe it's all out there and simply scattered. But I still would like to see a clear comparison of joint strength vs. needed joint strength.
Don B.
Don,
That's where I think the basic physics problems come into play. For example, a short legged coffee table compared to a long legged dining table with both having exactly the same joint of type and measurement would easily be understood that the coffee table would have the overall stronger joint. It's just the physics of force and leverage. Do you agree?
Don
Don C.--
I certainly agree with your interpretation of the physics involved in the example, but that still leaves it in comparative and not quantitave terms. Is a 1/4" x 2" x 1" deep mortise and tenon adequate for one and not the other? Should the dining table have a 1-1/2" deep joint instead? I don't know and haven't found a way (short of doing a structural analysis with a lot of pure speculation about stresses) to find out.
Don B.
Don,
Wouldn't you think that for what you are asking, standard engineering comes into play? You determine what the needs are and then design accordingly. As for building a table leg to have the same strength as a coffee table leg you would adjust for the variables. If you are figuring for the joint, whether your factor variable is from a table or input from an equation, you would still just be balancing the equation to meet the predetermined need. Or you can adjust for other variables. Maybe use a stronger wood, better adhesive, additional or different fasteners, change the legs shape, etc. If you think about it, I'm not sure if what you are asking for is available for much of anything. Take concrete bridges. There's no table available that states traffic volume, speed, approach lengths, bridge length, deep of footing and resting material, etc., where that table ultimately leads to some specific density of concrete, or type, or anything. Each individual variable is determined from some desire to build to a predetermined need. There are just too many dependent and independent variables. Do you agree?
Don
They were pouring the basement for my house and this fellow was trying to explain to me thrust loads, psi strength of the concrete etc. All this , because he wanted to justify not using $50 worth of rebar.
Sometimes you can spend alot of time thinking about these things or just go on and overbuild it.
Frank
Absolutely! Many many things are standardized by worst case scenarios.
Don
I am going to go off on a bit of a tangent - though still focusing on joinery - and apologize in advance.
When you take a historical ("an" historical, for Sgian and other Brits) perspective, what we do today seems sort of strange. By that I mean that the joints we use and revere today, the mortise and tenon and most especially the idealized dovetail, were invented a trillion years ago by woodworkers using handtools. Today we have all these marvelous machines that do everything for us, and what kind of joints do we make with them? The very same ones. Every day you see another jig for making dovetails with routers and every issue of FWW includes an article for making dovetails with tablesaws, bandsaws, your dishwasher, whatever.
So the question is, is this the limit of our machines? Or if we didn't have our heads stuck in the past, would we use machines to create new and better joinery techniques?
I claim this is relevant to Don's original post because. . . .well, because maybe there are a whole set of new joints we should be making and should therefore be included in the thesis.
Several years ago I read an article on the relative strengh of several joints including m&t, loose tenon, biscuit, and dovetail. I have it somewhere but I fired my libarian.
Interestingly the loose tenon came up a little stronger than the standard M&T. I'll try to find it, but there are no promises, I have a ton of magazines in no particular order. and I lost some in a basement flood.
Peter
Have you looked in "Forest Products Research Laboratories Handbook"? It's out of print but fully available in PDF on the website.
Just curious, but are you just curious or really concerned about some problem which MAY (never) occur?
The code books will tell you all sorts of things such as what kind of rafters to use for various structural and environmental conditions. I have seen a PhD thesis on the strength of timber-frame joinery. But I have not seen anything comparable for the broad range of furniture and the loads to which various items are subjected.
I doubt if anybody would die if a coffee table leg fell off. There's a reason you can't find any research (or very little).
I can imagine the look on my customer's faces if I told them I was substituting a nailed butt joint for dovetails on the drawers for their Secretary "because the nailed butt joint is plenty strong."
Don, you really aren't serious are you?
Donald:
What you are looking for is not as simple as you think. The construction industry code is based on grading standards that mean materials are consistent with a given range. There are only a couple of species of wood covered by the code - that's why you need an engineer if you decide to use an odd species like poplar to frame a house. Span lengths, stud spacings, etc. may all change and there are no tables to accommodate that.
When someone creates a piece of furniture, it could be made from one or more of hundreds of species of wood. And each of those species can and do act differently in joint situations. And to add confusion to the situation, most pieces of cabinet wood are not sold with a grade stamp - hence there are no guarantees of consistency (within a range).
Testing joinery in the multitude of species (there is not one species of cherry, for example) would require many man-years of work and would create volumes of reports.
It is far better to design your piece of furniture, anticipating what would be normal stresses and then using joinery that has a proven track record (look at something that has been around a hundred years or so see what joining methods have survived). And if you are gearing up for a production run, then you need to invest in some engineering assistance to assure you have the most reasonable joints for the money/product being produced.
Ed--
What you are looking for is not as simple as you think.
I don't understand how you got the impression I think the task is simple. I suggested it as a research project or thesis topic.
Nevertheless, I don't think it need be as complicated as you suggest. Serious question, because I don't know the answer: Is joint strength really highly dependent on wood species? It's just not obvious to me that it is. Even if there are significant species-dependent variations in joint strength, would it take more than two or three categories (weak, medium, and strong, e.g.) to encompass them all?
Great thread!
To grossly oversimplify - the connections in framed buildings depend on the number, type, size of fasteners not the wood itself. Most joints are stacked, lapped or butted. Nails, lags, thru-bolts, Simpson connectors, etc. are consistent and repeatable therefore predictable through testing and emperical data collected over time. That why it takes 5 #16's to nail a stud. Adhesives are far more subject to variables, especially in workmanship, which is why using epoxy achor bolts requires special inspection, pull testing, or in LA - observation during placement by a deputy inspector.
A "simple" M&T has far too many variables irrespective of dimension to test and therefore predict. M/C, expansion / contraction, fit, adhesive type and application, temperature, grain orientation, time, static loading, abuse (nobody ever sits on a coffee table), and so on. I don't think we're ever going to get past a 1/3 or 3/5 wide tenon x as deep as possible x as wide as possible.
I for one am glad. Standards would eventually lead to regulation. They'd be codeified or economically enforced via product liability losses or both. So much for freedom in our craft.John O'Connell - JKO Handcrafted Woodworking
Life is tough. It's tougher if you're stupid - John Wayne
The answers from the grant committee so far seem to fall into two categories:
1) Why bother?, and
2) It's too hard.
I find the first one much more persuasive than the second.
My question started out from a notion that most furniture joints are much stronger than needed. And then I realized (as has happened so many times in the past) that I didn't really know what I was talking about. How strong a joint is needed? I don't know.
I can't say that I've ever seen a joint fail for structural reasons (tenons break off, dovetail pins shear, things like that) in any furniture other than chairs. So let's gather some data.
Who has witnessed furniture joint failure (not including glue failure)? What were the circumstances? What failed? Would the failure have been prevented by: A stronger joint of the same type? A different kind of joint?
Having just finished designing a timber frame building and finding that even with the various tables, etc. available, I still had to go and do manual calculations to determine certain timber sizes based on the actual load for the building. I had to calculate the load to ensure I was sizing the timbers correctly.
The average cabinet maker/furniture maker does not deal with loads - simply stating low, medium or high or moderate, strong or very strong is meaningless - particularly when you've designed something that is clearly moderate in strength and you sell it to someone whose concept is very strong. These ratings are all opinion - you need numbers and I dare say hardly anyone in the cabinet business walks around knowing loading values - and not just at 90 degrees but also at 45 degrees and perhaps other angles. Your reference to a coffee table and someone bumping into it - what angle are they coming at that table? Is it a child or a 300 lb wrestler?
There usually 3 answers why something has not been done - 1) Nobody's thought of it or 2) Nobody wants to take the trouble to do it or 3) There are easier ways to achieve the end point. I think the answer here is #3. However if you really want to pursue the data, the Forest Products Lab at the University of Wisconsin (Madison) has already published numerous tables, etc. for wood and possibly wood joints. Also, the trade journals like "Wood" do discuss loading factors, strength levels, etc. routinely - this is not the kind of stuff usually discussed by hobbyist woodworkers.
I've suggested in the past (to students for instance) learn how to cut and understand the two basic joints 1 and 2, and you'll find that the majority of common joints used are merely a variation, then learn the less obvious ones, 3 and 4, and you'll be in pretty good shape.
1.Mortise and tenon, and its derivatives, e.g., biscuit, bridle, tongue and groove, lap, loose tongue and groove, housing, dado, dowel, finger joint (a complex bridle) and oh,--- almost forgot, the mortise and tenon, including double, forked, haunched, etc..
2. Dovetail, e.g., carcase, through, lap (half blind) secret mitred, secret lapped, slot, double twisted, machine, finger joint, sliding, tapered sliding, etc..
3. Combination, e.g., wedged mortise and tenon, showcase joint (aka 3 way joint) tapered keyed loose tenon as between stretcher and leg in refectory tables, etc..
4. Miscellaneous, e.g., mitre, scarf, three way joints, again, but here as used in hay rake stretchers, etc., knuckle joints, butt joints such as table tops, veneering and laminated work, and so on.
I don't know about a technical analysis, and charts, and stress testing being of much interest to me, Donald. Those sorts of figures might be interesting if you want to approach furniture making like some kind of 'by numbers' engineering project. Those people might as well work in plastics and metals and paint it all as far as I'm concerned. The best furniture makers seem to possess a combination of (instinctive?) technical understanding coupled with an appreciation for (predominantly) timber as a sculptural or 'artists' material. (That should get Charles active again, ha, ha.) I can't put numbers on the joints I use, but I know a well made M&T is darned strong, and ditto a corner dovetail (in a box, for instance.)
Two old maxims I'll pass on that I've found useful, which were passed on to me by old cabinetmaking farts many years ago. I think they learnt it from earlier old cabinetmaking farts, and so on going back a generation or two.
If it looks strong enough, it probably is.
If it looks too flimsy, it probably is. Slainte, RJ.
For a while, within the Forest Products Department at Purdue University, some staff members were doing some joint strength research, which if my recollection is correct, was published in the Forest Product Journal (Forest Products Research Society [FPRS]).
If you wish to pursue this issue further, I would suggest you contact Dan Cassens at Purdue. He is a faculty member and an Extension Forest Product Specialist at Purdue. No doubt if you do a web search on Purdue University Forest Products Department (or some appropriate variation of key words) you will get to their site and within that you should find what you need to contact Dan.
As wood technologist, I don't think that calculations based on the moment arm will provide accurate results. As I think about this issue there are quite a few variables that need to be considered. The first is adhesive type and strength. Based on my practical experience repairing furniture, glue joint/line failure is real. Joint strength will also depend on machining accuracy, how tightly/loosely the joint is constructed, grain direction, adhesive spread rate, the range of seasonal changes in moisture content (swelling and shrinkage) and the commulative effect of loading over time. And if you want to get really complicated, glue joints assembled with freshly machined wood (eg <24 hours) are stronger than glue joints where the machining was done previously. It has something to due with surface reactions with oxygen. And which is more important as an accurate predictor -- MOE or MOR? Maybe shear strength is a better predictor!
And then it gets all buggered up if you need to establish allowable design limits. What do you consider to be an appropriate safety margin? Chairs are used regularly as step stools; do you have to design the joint for all possible use applications (including for instances, some really grossly obese individual who leans back on the hind legs of a chair) and possible mis-use. I don't think that you can write enough fine print to exclude any and all liability exposure. Does the very act of "engineering" a joint increase your liability exposure; or is it safer to stand up in court and indicate that you based your design on some historic example?
I consider your interest in this subject laudable. Without some formula that will allow designers to predict the strength of a joint, design becomes a guessing game. Likewise there is no realistic way to calculate deflection of shelves. I tried to calculate some deflection and quickly realized that the formula presented in the Wood Handbook: Wood as an Engineering Material (USDA Forest Service -- Forest Products Laboratory [Ag Handbook #72] were derived from data collected on larger structural applications associated with building construction. To use these formula beyond their intended range is not a good idea in my mind.
It's an interesting subject and I am pleased you brought it up!
"Of all the woodworking joints in the world, you had to pick this one."
Rick, in Casablanca
Just to get it visible again to see if some answers turn up.
I was asking a similar question about joint strength and I found my answer in two Fine Woodworking issues. The first article was/is in April 1995, No. 111 and the second article was/is in April 2001, No 148.
This forum post is now archived. Commenting has been disabled