Heavy Mirror Frame, make it strong (how?)
jedimasteraubie
| Posted in General Discussion on
Hi,
I’ve made a picture frame or two over the years, but now that I’m in the midst of renovating a bathroom, I’m looking to frame the 48 x 36 mirror before I put it back up on the wall. I should first say, the weight scares; I dread nothing more than waking up to the sound of the bottom frame letting go and shattering glass. Am I overly concerned? Is 3/4 x 2 going to hold just fine with biscuits or splines? Other recommendations for carrying the load?
Also, I seem to be finding more examples of people securing their mirrors to a frame using liquid nails instead of clips? Is that really the better way to do it? Or are these examples just lazy? Not being derisive . . . yet. Seriously asking.
Thanks!
Mike
Replies
I might consider attaching it to the wall with clips and building a frame that fits around it and using adhesive to glue the frame to the mirror in a few spots. That way, the frame isn't supporting the mirror.
You can get metal J-mold, designed to hold mirrors on the wall, and build the frame as Badger suggests. Just google J-mold or J-channel
That mirror should not be too heavy, and would be easily supported by fairly slender timbers. It is, after all, only about twice as heavy as a similar piece of picture glass. I support a 4' x 8' mirror on 12mm treated pine outdoors without any problems (3 years now) except when a large horse used it as a scratching post. Not likely in your bathroom, I hope...
As the mirror does not flex, the majority of the force is applied on the bottom edge. Should the bottom part of the frame bow, the force gets transmitted to the corners, then up the sides.
Provided your corner joints are robust AND you put your mounts on the sides, then you will be fine.
Don't put the mounts on the top (or if you do, make the frame a bit bigger) as this can bow, then fracture, dropping the mirror.
This all works because wood is extremely good at resisting ductile stress, and it is why picture frames can be so remarkably thin yet still work.
As for biscuits - possibly. I'd use a mortice and tenon or floating tenon rather than a biscuit, simply because the biscuit is not good at resisting shear stresses due to it's shape. If the glue fails (unlikely but...) then the biscuit fails. With a deeper, square peg or dowel joint, glue failure will also require a racking force of some significant magnitude to cause your mirror to drop.
You could also glue the mirror to the wall.
If you have a shower in the bathroom, a nu-clear mirror demister pad is a nice luxury touch.
I would do what Badger wrote above.
The joinery of the frame will make a difference. Rather than rely on just glued at the corners, you could make the corner joints mechanically as well as glue-strong.
A bottom and top rail mortised into the uprights would resist gravity better than would splines or biscuits in mitred corners, for example. Uprights with dovetail ends let into the top and bottom rails would also resist gravity mechanically as well as via the glue.
For even more assurance, you could pin either joint with dowels - or even decorative nails - as well. Lap-joint corners (mitred if you like) with double pins through them are another option. I have picture and mirror frames in all three styles, some of which are true heavyweights.
I have made and had fall apart one mitred-corner picture frame, glued up with just one inadequately sized biscuit in each corner. It just takes one such slack corner joint for gravity to gradually wreck such a frame hung on the wall. And the heavier the frame content (and frame itself) the easier gravity finds it to apply it's wrecking bar. :-)
Lataxe
Some examples of frame corner joinery with mechanical as well as glue strength, in the attached photos.
All of these frames are attached via either a French cleat or a string 'round a hanging screw into the wall via the top rail of the frame. The dovetail, pinned half-lap and/or M&T corner joints provide anti-gravity mechanical strength to go with the strength provided by gluing those joints too.
The pins in the half-lap corners are two ebony dowels per corner, made very carefully with a dowel plate. (Ebony is quite brittle). They only need to be an inch long then planed, sawn or chiselled flat to the frame where they stick out after insertion through the half-laps.
Lataxe
I have laid plywood backs into frames with heavy content. If you are making the frame, a rabbet to accept this would be no problem. The additional racking performance of this "sheathing" can be quite substantial. I have a 60# mirror with this construction hanging from a pair of FastCap Roc-Locs and doubled #4 picture hanging wire in SoCal where the ground does the Watusi on a regular basis.
The advice above is better than I can provide, separately.
That said, if you want a woodworking solution, regarding the strength of the joinery, I'll offer the following...
I had a commission for a window frame to hold a 4' x 2' heavy glass window (1/2" thick), and created the following floating tenon joint. I didn't want to rely on glue-alone.
Or you could do what I do and don't build anything that hangs. There are a few good mirror stands out there.
Mikaol
I don't want to single anyone out here, but why do people with obviously no engineering background insist on telling others about the physics behind how things work? All you do is misinform others.
I didn't see any posts discussing the physics of hanging mirrors...
Badger... I appreciate your point; though, it is a bit cryptic a response. I think, I assume, we're all trying to help within the limits of our own experience and preferences...
OK then...the details of response#3 as to how the forces are transmitted, the concept of "ductile stress", etc. are all hogwash, to put it mildly. If you don't understand a subject, why pretend that you do?
I remember someone in a home improvement forum advising someone about electrical repairs, claiming to be a master electrician. He advised the person to just connect the hot, neutral and ground wires together with a wirenut to get rid of a problem. When that person did so and reported back that his circuit breaker tripped, he was then told he had a faulty GFCI. Could have burned his house down.
This isn't as bad, but I have to call BS when I see it. This is sophomore level engineering indeterminate statics.
Ha ha - Badger by name and badger by nature.
May we look forward to more of your biting and growling on other subjects?
Lataxe
Engineering this question.
A 3ft high by 4ft wide mirror using 1/4 glass will weight 40 lbs. Assuming the weight will be distributed evenly it will apply 20 lbs on each bottom corner joints when hanging on the wall. We will not consider dynamic loads, unevenness of weight distribution, wall friction etc, these will be accounted for in a generous safety factor.
If we cut the frame at 45 degrees and simply glue it together we get a glue surface area of 1,85 in2 at each corner. Using 3000 psi of glue holding strength and a safety factor of 10, we still have 550 pounds of holding force. Using that same logic to a half lap joint brings us upward of 1000 pounds to hold up against the 20 pounds load.
Chances are that is you pulled down on the lower frame to test the strength of the frame, the hook on the wall would fail first.
Personally I would use a splined miter joint for appearance and peace of mind or a mitered, half lap.
Umm....there are no "dynamic loads" to be ignored when considering a mirror hanging on the wall. Dynamic loads only apply to something that is moving. Perhaps you are think about a mirrored disco ball?
I stand by my comments.
I would simply alter the gravitational constant of the universe in the area surrounding the mirror. Problem solved.
And a dynamic load would be applied whenever cleaning the mirror.
I am a novice woodworker who enjoys reading this forum over coffee every morning. Although I do not have the experience that so many of you have. I have learned a ton of things in the category of "things I didn't know I didn't know."
What I admire most about the forum is its civility. Questions from novice woodworkers are particularly well-handled. Comments are almost always respectful and non-judgmental.
This thread however has for some reason engaged a commenter who may be an engineer or physicist whose comments are not particularly respectful or helpful.
I am a retired dean of engineering. I have taught more than 5,000 engineering students who have gone on to professional practice. I have emphasized to all engineering students that they are privileged to have a formal education in how the physical world works.
But they should never forget that the world is full of intelligent people who have developed a remarkable grasp of engineering and physics without formal training. I advise students to listen carefully to the experienced craftspeople in whatever company they work for. Wisdom is not taught, it is learned from those with experience.
That's nice, Russ, but I think you miss the point. Read my posts from the beginning; in my first post, I simply gave what a few others agreed was a good option to supporting the weight of the mirror from the frame - there's no concern about how poorly made the corner joints may be designed, or executed, if you eliminate the need to support the mirror and simply set a frame over a mirror securely fastened to the wall. In my second post, I was merely trying to get the point across, without naming names, that there was some bad info here. It wasn't intended to put anyone down or build myself up.
As an Engineer yourself, don't you want to strive for correctness and accuracy, and not just let anyone talk about something they don't understand, thereby misinforming others?
Is it OK that the person posing as a master electrician was giving bad advice that could easily have caused a fire, or worse?
Is it OK for the famous MN politician to propose that the value of pi be rounded to 3.0, because it would be easier to remember and might help others to learn geometry? After all, his intent was good.
Is it OK for me to tell people to drink bleach to ward off the coronavirus?
Closer to home, is it OK for me to tell people that they should not worry about kickback because I don't understand the physics of it? Can I falsely tell them in a forum that in all my experience, I have never used a splitter or riving knife and have yet to be seriously hurt, so they don't need one because of some made-up misunderstanding I have of the physics of kickback?
AS an Engineer, didn't you also realize the faults in the analysis given in comment #15, which I wasn't even going to bring up? Even my wife, with no technical or engineering training whatsoever, would realize that a simple picture frame corner joint doesn't have 3,000 lbs of "holding force"? If that were the case, our local auto mechanic would be lifting cars and trucks with a pair of picture frames.
You should drop the simplistic, "holier than thou" comments - you know nothing about me or how I operate in my personal life. This isn't about people who don't have formal training but understand and successfully apply concepts of physics. It's about misinformation! I have always had nothing but respect and admiration for skilled craftsmen, for example the tool & die makers and pattern makers that I have worked with in the past; they have always been, for the most part, quietly knowledgeable and humble, even though they know more about getting tooling, prototypes, etc. made than many young and inexperienced Engineers. The same goes for the electricians, millwrights, pipefitters, etc. that I have come to know. I have to believe that they would also try to correct misinformation.
Sorry to spoil your morning coffee.
BTW, I'm not going to respond on this subject any further; I think I've made my point and can't add much more to it, so have at it.
Sorry Badger, but you were out of line. That one response that you were criticizing was in essence correct. Remove the terms "ductile stress" and "racking force," and there was no issue. I don't believe for a minute that those terms were misleading anyone.
You clearly have a lot of anger at the world. So do I. Some of the misinformation out there is absolutely dangerous. But no one here was teaching people to burn their houses down, or to drink bleach.
You didn't try to correct misinformation. Not here anyway. You've told us about specific misinformation in a bunch of other places though.
There were many ways you could have taken faulty information and used some expertise to explain it. You still haven't done that. You opened a door, lobbed in a stink bomb, and closed the door on your way out. That isn't helpful, instructive, or even kind.
I would ask Badger for his advice about good conduct in a forum ....but ..... he appears to have no expertise in this area. :-)
But really I think Badger should be allowed to practice his curmudgeonly mode as this forum needs some light entertainment. Of course, he comes nowhere near the expertise in curmudgeonation as seen in the olden days of Knots, when we had BossCrunk, Mel the malicious and various other biters at one's faux pas, character assassins and reedy-voiced pedants giving lectures on the REAL meaning of some WW thang.
To be frank, this forum needs some livening up. It can be very, very dull.
*****
Anyroadup, how will the OP make his mirror frame? I am hoping for the DT'd uprights, myself.
Lataxe
Badger,
Reading through the posts, you jump straight to the insults–right past the discussion. You can certainly disagree but lets find a way to point it out in a civil manner. I don't care what anyone's background is, civility is required.
Can we get back to woodworking... even Einstein was confused about the Gravitational Constant... He didn't like Quantum Mechanics either, for that matter, even though he originated the idea of 'quanta'.
let's let this one go.
..Even my wife, with no technical or engineering training whatsoever, would realize that a simple picture frame corner joint doesn't have 3,000 lbs of "holding force...
Where do you see 3000# of holding force ?
..Umm....there are no "dynamic loads" ...
Yes there are when the mirror gets moved and manipulated, these can weaken the joint, one has to account for installation , moving and unexpected mishandling.
Most failures of glued miters are due to wood movement. As the boards get narrower or wider the angle from the inside of the miter to the peak changes from 45 degrees. Glued end grain cannot hold against the same forces that crack solid panels in furniture for the same reasons.
Most "holding force" tests for glue strength are done under far more ideal conditions than end grain to end grain joints.
Thanks for the responses. I appreciate the constructive input! There were plenty of perspectives gleaned and I think the mitered half lap is going to be the way to go for this project. Not quite ready (at least not in the timeline I'd like to have this finished) for mortise & tenons but one day soon!
@Badger
This is I believe the second time you have made rude and unhelpful comments about my posts.
I feel that including at least some perspective of how the forces are transmitted is helpful in the discussion of any joinery issue. If you believe that I have erred, simply using the term 'hogwash' is not acceptable conduct - as a self-professed expert, it is beholden on you to explain why this is the case and provide educational correction rather than, as you seem to have done, stomp off and sulk under a bridge.
In future, I am more than happy to accept corrections, though in both this case and the previous one I am correct in principle at least, if not in terminology, however a civil approach is appreciated, as is the benefit of your knowledge, which so far at least, you have not felt it necessary to share.
Perhaps you might like to educate us all on how picture frames stay together, from your own extensive experience, of course.
Rob_SS,
Please don't allow a badgering by a curmudgeon to put you off making your often valuable contributions and perspectives. Also, please tolerate the curmudgeons as they are just .... different; and so are interesting, as are their viewpoints.
I learnt lots from curmudgeons (as well as from the less miserable folk) in the days of Knots, the old forum of FWW that was such a good school, a decade or two ago. One must simply play the curmudgeon on the end of a line until they eventually surface and emit something useful rather than just a snap of their wee teeth and a lot of splashing noise. :-)
It's likely that The Badger has a lot of useful stuff he could impart. His mode of doing so is merely the scratching at his various mental fleas before he tells us something of utility. You can ignore his growls as they are just indications of those various fleas jumping about his wetware.
It takes all sorts of personalities to make a good forum. We need to be able to tolerate and interact with all sorts of attitudes if we want to have the conversation. Those conversations can teach us a lot; and not just about the overt subject matter.
Think of the forum as not just a means to discuss WW but also as an anthropological or sociological exploration in exotic domains of "the other".
Lataxe, more like Mole meself.
To The Jedi,
Mitred half laps seem the best answer to obtaining a traditional look with a stronger joint, in a frame. They increase and improve the glue surface. I would still suggest a pin (dowel) or two through them. Even if the glue somehow turns to dust and fails (unlikely, but ...) the pins would stop a crash to the floor.
Overkill, perhaps. But also the opportunity for a bit of a decorative effect albeit minimal; and a new joint in your arsenal.
Lataxe
I think I am going to jump into this sticky conversation because I actually feel for Badger. Not in the way he responded, but in the fact that there is so much bad information given in this forum. I personally don't respond, but it does make me cringe. Sometimes I want to blow up like Badger did, but I just let it go and part of my engineering soul dies with it.
I will correct the problems with the Rob_SS, MJ and Gulfstar posts because there is some incorrect information in these post. I will try to do it in an educational manner.
Rob_SS mentions ductile stress. This is not a type of stress typically discussed in engineering. I think you wanted to say tensile stress, as in wood has good tensile strength (along the grain). Ductile means a material will stretch under tensile load and typically remodel and elongate. Wood is not ductile.
Also, forces really are not "transmitted." The weight of the mirror is supported by the end joint always. This is not affected by any bowing of the bottom support. Theoretically, the bottom support has to bow to some extent under load and this will change how the joint is loaded. This is a very small deflection and can typically be ignored in all but the rare case.
I'm going to skip my opinion on joint strength of various types of joints, but I think the biscuit joint is strong and gets label inappropriately as weak.
Gulfstar's post on joint strength is both pretty accurate and really wrong. As in, a good engineering approach and some really bad data.
"A 3ft high by 4ft wide mirror using 1/4 glass will weight 40 lbs. Assuming the weight will be distributed evenly it will apply 20 lbs on each bottom corner joints when hanging on the wall. We will not consider dynamic loads, unevenness of weight distribution, wall friction etc, these will be accounted for in a generous safety factor."
This first paragraph is awesome. Spot on engineering. As for all the people who mention dynamic loads, they can be ignored. A factor of safety of 2 provides resistance to a dynamic load with 1g of deceleration. Typically, way more than enough for most small hanging objects, unless your cleaning person is crazy aggressive with mirror cleaning. Dynamic loads are treated differently than cyclic loads, which is not mentioned in this post and in the case of a mirror frame, can likely be ignored, but in the case of a chair, should be at least considered.
So Gulfstar starts of with a well set up engineering problem. The second paragraph didn't past the sniff test with me when I read it and a bit of research lead me to the appropriate data. A 3000 psi holding strength for an end grain to end grain joint would mean a 1"x 1" block end grain to end grain glued would hold 1.5 tons in tension. No way will that joint hold that load.
Some research while drinking coffee led me to some really good testing data on wood glues. PVA (specifically Titebond) has a edge grain to edge grain strength of 300 psi in shear. This seems to be a more realistic number. Same testing revealed an 200 psi strength for edge grain to end grain, again loaded in shear. I could not find any end grain to end grain testing, but I found a journal article with testing in both shear and tension. It appears there is a 50% decrease in strength when loaded in tension and this is what I would expect from previous studies I have done on adhesives.
So, what is Jake_W predicted strength of the mirror frame glued miter joint? Well, there is a 33% reduction in strength when we went from an edge grain to edge grain joint to an end grain to edge grain joint. I am going to estimate that it would be another 33% reduction with and end grain to end grain joint. There was also a 50 % reduction when the loading went from shear to tension. The miter joint will see a combination of loads, both tension and shear, so I am going to assume another 25% reduction. Starting at 300 psi and reducing 66% for the end grain to end grain joint would put us at 100 psi and another 25% for the combined load would give us an final strength of 75 psi.
Using your joint surface area of 1.85 in^s would give us a strength of 138.75 lbs per joint. This would be a factor of safety of 6.9. And this calculated strength does seam more reasonable.
And for the non-engineers, this load is calculated at the joint. If you place this same load in the middle of the frame, it creates a "lever" or in engineering terms a moment around the joint, increasing the stress.
MJ mentions the change in dimension with moisture content. It is absolutely 100% true that the joint angle will change with moisture content and this places stress on the joint. For a narrow frame, this change is very small and I am not going to go through the math on this and for almost all common woodworking glues the strain is within the flexibility of the glue. This analysis changes as the with of the board increases. At some point, it will cause a problem.
Sorry for the long and technical post, but sometimes an engineer has to do what and engineer has to do.
I appreciate the engineering discussion on glue strength. But It's on paper. I've seen glue joints fail. I've seen a couple of glue joints that I've made myself end up failing, and I would have sworn they are sound.
The only time I trust glue by itself is either on a wide panel glue up, and on purely decorative elements. Neither is structural. On my response at the top of this thread, I seconded Badger's suggestion to mount the mirror to the wall, and put a decorative frame over it. Glue alone for that frame is ok, as it is purely decorative.
On anything structural, I will always use some kind of joinery to provide strength, and the glue is only used to hold the joiner together. In the case of 40 pounds of glass hanging on a wall, I wouldn't dream of holding it up with a few square inches of glue alone. It's just not necessary to use only glue -- there are too many better choices.
Why not use metal reinforcement? There are tons of metal brackets and corner braces that would give you more assurance without complicating the design. I like to use these on heavy frames:https://www.amazon.com/Hillman-Group-122728-Joiner-40-Pack/dp/B000JPEGBU/ref=sr_1_4?crid=1RLA2TQ1MQQCO&dchild=1&keywords=picture+frame+corner+fasteners&qid=1599140366&sprefix=picture+frame+corner+%2Caps%2C169&sr=8-4
I'll use half-laps, with pegs, and add some of these to the back corners.
My son had a big mirror that was glued to his bathroom wall come off one morning (with the frame attached), just after he had been using the bathroom. The glue pulled the paper right off the drywall. Looked pretty dried out.
Now I know remember why I left Lumberjocks, WoodWorkingWeb, and the miriad of other woodworking forums that degraded into personal attacks among those that contribute.
The fact is ... I'm not learning anything from this thread.
That's too bad. Depending on how advanced you are, there is a bunch of good information here. And this is the only recent thread I've seen where there has been any squabbling. And I wouldn't call any of it personal attacks.
I took the 3000 psi glue strength from the titebond data sheet, I did not study ASTM D-109 for the claimed value on hard maple. Titebond 3 actually claims 4000 psi. I also taught the values were high. There is nothing like a peer review to make the values more in line with what one would expect. I would expect the 45 degrees miter joint To be somewhat stronger per square inch than a end grain but joint . More like a short scarf joint if that data exists.
I realize this thread has drifted far from the original question, but there are some good mechanical engineering concepts being discussed that relate to woodworking.
I wanted to respond to John_C2 because he posts lots of good stuff. Yes, engineering happens on paper. All design happens on paper. It drives me nuts when people act like this type of design is "not real world applicable." Engineers design millions on things and failures are occur at a low rate. And a good engineer will study the failure to determine why is happened.
You mention you only use glue on wide panel assemblies and decorative elements. This is where tradespeople can drift into dogma. Your answer is correct, but your logic is flawed. Glue strength increases with surface area. A panel glue up has a large surface area and this allows glue to really demonstrate its value.
On the other end, you mention decorative items. These typically have a small surface area and the glue provides minimal strength, but they only need to resist gravity and the occasional wondering toddler.
Glue is pretty strong. A 300 psi measured strength or a 3000 psi impact strength is really good. My whole house is built on glue lam joists. They are plenty strong and glue is a huge part of that strength.
The flaw in your logic is that in certain areas glue is "weak." It is not the glue, but the low surface area does not allow for adequate strength. Or there was a part manufactured out of tolerance. While jointery is one way to solve the problem, increasing the surface area of the glue bond can increase strength dramatically.
What I'm trying to say is that when people classify a design as weak or strong or good or bad or whatever, it drives this dogma that actually limits design.
There are infinite solutions to this mirror problem and many will work well. Would I trust a glued miter for this design, heck no. I would probably stick a dowel in it. But a mitered half lap joint would increase the glue surface area to allow a strong design. Or a mortise and tenon. Or a steel bracket. Or just glue the mirror to the wall (strong option because of the large surface area), etc.
Every tool in my shop has its pros and cons and understanding this allows me to execute my projects well. The design process is a huge tool in any woodworkers shop and to limit the design by putting artificial limits on a method because it is "good" or "bad" limits your options and ultimately limits your abilities. Understanding why a glue joint works or fails is very important and not "just on paper." You can take this knowledge and design a joint that has the right strength for the design. It is freedom.
I know these have been long posts, but there have been a couple of threads where strength and assembly method have come up recently and there were lots of statements on good and bad and whatever. It is frustrating because it causes woodworkers to close down their design space. I just had to vent. And it is not just in the forum where people do this, some of the authors and editors at Finewoodworking do it as well.
Thanks for letting me vent.
I have nothing against paper. All of my designs are still done on a drafting table and paper.
But I think people get hung up on the theoretical part of woodworking. That's what I meant by "on paper." The piece still needs to be built, and my preference is for something to stay together even if there was no glue. Or if the glue fails. Dovetails, mortise and tenons. I peg joints whenever I can.
Some glue joints fail because of poor technique. Other times, I just don't know. But when I see someone insist that you can make a perfectly strong joint on a rail and stile door using just glue on the end grain of the rail, I roll my eyes. They are welcome to try it, but I never would dream of it.
I often wonder how newer homes will fare over time. So much of the construction depends on adhesives. I wont live long enough to see if there are issues, but I do wonder about it.
The engineering discussion is indeed a very interesting one. The "on paper design" point is particularly interesting as it's one way of looking at the theory vs practice issues that all crafts involve. It's also an aspect of the tension (forgive pun) between a pure science approach in deriving technologies and the longer try-fail-improve traditions found in the history of technology going back to the Egyptians and probably beyond.
The great advantage of technologies, including construction practices, that have evolved with a minimal amount of pure theory (or on-paper) design is that that they come to be refined to work very well through lots of trial and error. The dovetail, mortise & tenon and no doubt many other joints evolved in that way, with little theoretical a-priori input.
Their drawback is that, like all traditions, they can be ossified or moribund - stuck in a particular good design because no one has the motivation to improve it if it works very well.
The on-paper or theoretical approach that waxed and waxed from the days of The Renaissance, the rise of Rationalism (the early beginnings of the scientific method) through the Industrial Revolution provided vast amounts of possibility, much of which was rendered into reality as real new technologies that would have been unlikely to arise through a wholly practical and empirical trial & error process.
But there are issues with the Rationalist approach in that theories (being made by humans) are often flawed - but don't exhibit the flaws until rendered well into the technologies they indicate.
Obvious examples abound in the modern world - technologies that have been produced by scientific and engineering theory that work ..... but not exactly as intended. They often have a host of unintended and unforeseen consequences.
At the low level of wood working tools, materials and design, the theoretically-derived techniques and tools can often be seen to work .... but only if the user obeys often stringent rules, using exact materials and methods with a perfectly rendered design of the tool. The history of the biscuit joiner is a good example. .....
Lamello, Mafell and high end manufacturers made well engineered examples that were fairly easy and effective to use albeit not foolproof. Many other manufacturers climbed on the bandwagon and made biscuit jointers that were not so good. They made slots that were often too wide because of blade or bearing wobble, canted to one side a bit because the fences were floppy. Biscuits got a bad reputation in many WW circles as a result. Many biscuit joints failed.
Pure-theory on-paper designs can produce very good real objects and technologies; but also some very poor objects and technologies. They are not derived in a cultural vacuum out of pure logic and math but are also subject to design inputs such as profitability and fashion. When a manufacturer ceases to be motivated primarily by making good things to become largely motivated by making lots of money, the designed objects can suffer greatly.
Consider the planned obsolescence of most high street furniture (and planned obsolescence as a deliberate design input); the TSOs (tool-shaped objects) of the post war eras that sold cheap in their millions but didn't actually work as a plane or whatever; the landfilling "new, improved!" effects of constant iterations of i-things and similar that cease to work as the manufacturer changes the design parameters of the supporting tech to enable the sale of a million "better" models.
****
Personally I've come to prefer tools. methods and designs with a long history that have become those good traditions we often employ in woodworking - the various joining methods that John_C2 mentions being one of them. This often means "belt & braces" - a mechanically strong joint as well as one further improved with the right glue applied properly.
Lataxe
JakeW: "MJ mentions the change in dimension with moisture content. It is absolutely 100% true that the joint angle will change with moisture content and this places stress on the joint. For a narrow frame, this change is very small and I am not going to go through the math on this and for almost all common woodworking glues the strain is within the flexibility of the glue. This analysis changes as the with of the board increases. At some point, it will cause a problem."
A 3" wide maple board (a reasonable frame size for a 3' x 4' mirror) will move about 1/16 in width going from 12% to 6%, about right for seasonal swings in the NE. that changes the miter from 45 to 44 degrees or to 46 degrees, depending on when you built the frame. That 1 degree change becomes 2 degrees when you add the same change to the mating board, and the stress being compounded (my non-engineer experience / impression) by the same thing tugging at the other 3 corners.
I don't think it should be dismissed quite so quickly when a rather large panel of hanging glass is involved.
What "incorrect information" was in my post exactly?
I strongly considered letting this thread die. I don't know how to respond to your question MJ. The original post mentioned a 2"x3/4" frame. Gulfstar used the same dimensions in his assessment. (I totally think Gulfstar is right on the scarf joint for strength by the way). I made my assessment on a 2" wide frame and then you assume a 3" wide frame. If you change the problem, my answer very well may change. Heck, lets make it an 8" frame, determine all engineers are idiots and go on our way.
But, your numbers are wrong. Assuming a 3" wide frame and a 6% moisture change and the data I could find on expansion and contraction for hard maple, I get a change in width of 3/64. Close enough to 1/16. For 2" wide maple I get 1/32, as expected for a decrease in width.
When I plug this in to my calculator: atan (2/2.0324) I get an angle of 44.6. Only a change of 0.4 degrees, not 1 degree. Total opening angle of 0.8 degrees. This is approximately 1/32 of an inch of opening on a 2" fram. If you increase this to a 3" wide board it is the same open angle, but the width increases.
You discuss stress in the joint because of the expansion. The change in width generates a strain in the frame. Unfortunately, I cannot find a modulus of elasticity for hard maple measured in tension, only bending. For an anisotropic material, I am hesitant to use a modulus not measured in the direction that I want to use it in and therefore I can't calculate the expected force induced in the glue joint for this strain. I could find a modulus measured in bending and this was higher than expected for my knowledge of wood behaves in tension. But I used it anyway (1.83 x 10^6 Pa) and calculate a stress of 1500 psi assuming half the joint is loaded in tension and half in compression on the 36 inch side of the mirror. This is half of the strength listed in on the Titebond site, but much higher than the strength found during testing I used earlier.
Every the math nerd and thinking the modulus for bending was significantly larger than expected, I calculated the expected radius of bending that would occur in the 36" side that would be generated by this stain and it was approximately 1100 feet. This seems to be a very reasonable small deflection that would be generated by the change in angle with maximal moisture changes and within the range of plasticity of hard maple and the plasticity of PVA glues.
Going back to my original posting, I never mentioned "incorrect information." In fact, I mentioned that you were correct. I said that you could ignore this small change in angle for the 2" wide maple frame. While I did not show my work for all of this, I fully calculated all of this out and my analysis fall inline with my initial statement that the small strains and stresses fall within the plasticity of the wood and PVA glue.
If you change the glue, wood, moisture content, width, or any of my other assumptions, my recommendation may change.
If you want to use the general non-engineering response that this is all theory or "on paper," I would point out that as an engineer I can put a number on my assessment of strong or weak. I can determine within a reasonable and definable degree of accuracy whether a joint, structure, or any other object will fail. If you still don't believe my analysis, please don't drive across any of the 600,00+ bridges in the United States. They were all designed on paper and trial and error is to deadly for bridge building. The math for this simple mirror frame nearly identical to a bridge.
On an aside, dovetails, M&T joints, etc were not developed by trial and error. They are clearly well designed joints that used the available knowledge (at the time of their development) about wood and glue (mainly hide glue) to achieve a solution to certain fabrication problems. These are engineered joints, not something that occurred haphazardly. They have stood the test of time because they are well designed and rely on the material properties of wood for strength. This was important when adhesives were in general poor. They are not some fabled method past down from generation to generation by word of mouth. They are just a joint that has good mechanical strength.
I think it is time for this thread to die
The time for a thread "to die" in not within your purview, Jake old engineer. Even a Jake-argument will not necessarily contain the an irrefutable calculation as to why it should die "in theory". :-)
******
Personally I prefer to approach woodworking problematics (and everything else) with an understanding of what actually worked and what actually didn't. This takes into account what could have worked theoretically but didn't because the design parameters were too exacting for many human makers to achieve.
In short, belt & braces, so the trousers don't fall down when the poorly-made buckle of the belt snaps.
The width of a crack in the glued edges of a mitred frame when humidity changes flex it are not that relevant really. The point is, glue-only mitred joints (of any width) can be cracked by humidity changes quite easily, by a lot or a little, reducing their already relatively weak bond beyond failure point. Why risk it it with nothing but a trust in the glue and the perfect application of it to perfectly-made mitres that are not subject (ever) to damaging humidity changes?
Biscuits, splines and similar will improve things, certainly, if also well made and applied. But a mitred lap joint pinned with two dowels will be more resistant to collapse again. It might be "over-engineered, on paper" but why is that any kind of problem? Theoretically-sound glue-only joints that have a long history containing numerous failures are the problem.
****
To a designer, everything looks designed. To an engineer, everything looks as though someone used engineering theory to design it.
Many technologies have arisen out of pure trial and error with subsequent on-paper designs after the fact; or as refinements of an already extant design. And trial & error design is not "haphazard", merely a method not using formal mathematics and mechanical theory in a fundamental way.
It's an old trope in Rationalist circles, especially those of Academe, that the theory comes first in designing, followed by the design. This certainly occurs now but is a relatively new mode of designing technologies in human history. It's a reflection of the scientific method, another modern phenomenon.
As an aside, you will find this point turned on it's head within the history of science. The notion that "pure facts" were observed which then suggested theories used to be commonplace. Since Thomas Kuhn, its now recognised that theories tend to come first with the scientists, followed by a definition of what are facts so they fit the theory. And so arise all those on-paper perfect designs that in practice - aren't. Consider the history of modern medical drugs and their many damaging side-effects.
******
It's a modern attitude, this, "Everywhen was really like now only not mentioned in the written history of the time for some reason". People argue that science and capitalism "have always been with us". .......
No they haven't. Other modes of thinking and behaving were possible, extant and sometimes rather better in many respects. The past is "a foreign country"; and not all the foreigners living there with their different ways and means were lesser folk than "us, now".
Lataxe
Nothing ever truly dies. 14 years from now someone will respond to all this.
One possibility to strengthen a miter joint is the use of corrugated metal clips that are inserted from the back to close and hold the joint together. We used them in a factory to assemble 3 1/2 in. Wide hard maple frames using a press but they can be hammered or pressed using a manual clamp or better still a vise. https://www.finewoodworking.com/forum/picture-frame-miter-fasteners
@JakeW: "Going back to my original posting, I never mentioned "incorrect information.""
Well, yeah, you did: "I will correct the problems with the Rob_SS, MJ and Gulfstar posts because there is some incorrect information in these post. I will try to do it in an educational manner."
I love your approach to the numbers; assuming the ones you "found" are superior to the ones I found.. but I'll call yours "close enough" without declaring them wrong.
I plotted your 1/32 change in a 2" wide frame on graph paper (scaled up) and measured a 1/2 degree change in the mitered cut, or a 1 degree opening in the joint. My numbers for the 3" frame were also plotted and measured, just like most of what we do in the shop.
No one here would attempt to assemble a frame that is out of square by a total of 4 degrees, even if we could get it to fit by clamping an 1100' radius into it knowing it would be straight come winter (summer?). I believe the numbers you are spewing to be accurate, I just don't see how they are useful as anything but something to ponder while you are actually building a frame in the shop, or as feathers on the end of a spear to confuse your opponent.
How can the "opening angle" be the same when the change in width for each board differs as your earlier math points out? My graph paper shows a change in the angle.
I admit that for my first reply I did not go back to the OP for the specs, but I think we've arrived at the same place anyway. I don't believe the jump to declaring ALL engineers to be idiots is warranted because of it.
Long live the thread!
I am only responding to this thread because MJ asks some interesting questions and appears to engage in a discussion that may be fruitful.
I hear you on finding good numbers for the materials properties of wood. I think we are both using numbers in the reasonable range. When I do actual engineering design (and some of this stuff has ultra low factors of safety), I list all of the assumptions I make, many of the assumptions used to derive the modeling tools I use and all of my constants. I then check all of these with my final design to determine they still hold. I frequently include expected variation in manufacturing. These are not some pie in the sky perfect designs. It is a structured process that allows my team (as outside consulting engineers) to nail a design in short order with rare problems.
Now to address your question on why this analysis even matters. Here is my opinion. In my bathroom there is a framed painting. About 12"x 14" with glass and matting. The frame is about 1/2" of some unknown soft wood. The joints are just glued together. It hangs on a nail just in the drywall. None of this is overly strong. There was no engineering analysis of how strong this is. Mostly a decorative item, installed based on my past experience with all of these fabrication methods.
In my study, I have a 6" wide mitered return on a top below a bookcase. The joint is glued and reinforced. Half of the joint floats. There was extensive thought in how this will move overtime and construction methods were used to reduce the risk of opening.
No formal engineering analysis was done on either of these designs and if the joint fails, minimal damage is done. But in the first case, I made the assumption that seasonal changes in the width won't affect the joint and in the second I made the assumption that seasonal changes will. I can do an analysis at anytime to check my assumptions.
We all make these assumptions. At some point, we all use a rule of thumb or prior experience to make a prediction. These rules of thumb all have a limit and you can use some basic analysis to determine the limit.
In this problem, there is some width of the frame, below which seasonal changes in moisture don't affect the joint. The glue is strong enough to resist the forces. This width can be calculated. That to me is useful. As in a rule of thumb where you can ignore seasonal changes or where you need to account for them.
As for the joint angle being the same, using trigonometry the calculation is atan (width/(width + width * constant used to calculate expantion)). The width completely factors out and the angle is calculated at atan (1/(1+constant)).
I'm only responding to this thread now because it amuses me. And hopefully others.
Heh. It do.
I know for a fact that the folks over at http://www.finephysics.com/forum are hugely amused, ha.
@Jakew: Having lost track of my trusty TI-55 I went back to the graph paper (see attached) to run it again. It made no sense to me that the joint angles on two boards of different widths, both starting at 45 degrees, changing size in one dimension and remaining fixed in the other could result in the same joint angle post-expansion.
I scaled the boards to give each a clean 1/2 box for expansion and used a protractor to measure the results. I invite you to reproduce this work.
I never asked the question of why this analysis matters, just the practical uses for knowledge like a radius of 1100' to a woodworker. I DID ask about what incorrect information was in my original post, and while the detailed description of your business, team, process, and past successes are impressive they do not serve to answer anything we are discussing here.
Finally, you might want to choose between math & science and rules of thumb. The way you bounce between them is convenient, but confusing.
MJ, why is the wood shrinking from one side only with the outer edge in a fixed position ? What if the fixed position in the center of the board ?
In my original post I stated that the change in joint angle for a mitered frame made of 2” maple could be ignored. In your post you stated this needed to be taken into account in the construction of the frame. My opinion is that the need to account for the change in joint angle is incorrect information. You have taken great offense to my casual, definitely not scientific, acquired through trial and error, rule of thumb, that dimensional changes caused by seasonal moisture fluctuations can be safely ignored for board 2” in width. I didn’t realize this was such a grave insult in the woodworker culture. I give you a thousand apologies. While I continue to use my mathematically derived rule of thumb for furniture design, I will no longer give a new woodworker this same information.
The math provides some justification for my statement that wood movement can be ignored for hard maple less than 2” in width, clearly this was not helpful in this thread. I will refrain from using logic and material science in future posts.
I would be happy to provide an algebraic proof for how the opening angle will stay the same in this situation because the change is proportional to the width, but I’m unsure if I can use algebra (invented in the 8th century) because it was written down and I’m currently considering different possibilities for proving this developed by all the cultures that didn’t have a written language.
I’m also far to busy redoing all the miter joints in my house trim using double blind mitered dovetails. It is slow going with my stone tools. I have forgone steel because it is a modern invention.
If you want to know where the error in your drawing is, it occurs with rounding errors. The calculations give you a bunch of decimals and you round to the nearest fraction. This is where some of the difference occurs, the rest is measurement error because you don’t get 3 decimal places with a protractor.
Well I’m off to use math and rules of thumb to rebuild all of my furniture using tried and true methods passed down by the master unicorn woodworker. Or maybe I’ll become a stone mason because I’m the history of earth, stone has been around much longer than this modern invention of wood.
My original reason for posting anything on this thread was to point out that using sound engineering principles in discussing woodworking design is frequently ridiculed on this board. I think this thread proves my point.
Again, a thousand apologies (and I would use the language of a forgotten culture, but I don’t know that language) for trying to provide some knowledge of the materials we use in woodworking.
In off to http://www.finestoneworking.com to learn more about cutting double blind miterdd dovetails in Devonian limestone
Gulfstar, It does not matter; you can think of it as expanding from center and then being slid back to the the edge to be measured. I drew it
that way so it would stay on the grid. The peak and inside corner only move apart, the wood does not move along its length, that's why the angle changes. The 45 degree angle is the starting point, like it was built in the dry season, the wider state of the boards is August in DC.
My thinking is that if, for example , the wood is held fixed in the center of the width, then by shrinking or expanding in width by say 1/32'' from each edge, the angle remains constant.
The angle can't remain constant, the lengths of one leg of the triangle and of the hypotenuse are changing.
Jake, as I said before, I know your math is sound, it is just impractical to "educate" a woodshop situation out to three decimal points. That fixed set of expectations fostered my resistance, and to be honest I was enjoying our exchanges.
I have seen narrow frames fail, possibly because they were not built by a team in a lab where perfect glue applications yielding perfect bonds due to the perfect application of clamping force. Stating absolutely that the change in a 2" frame carrying a large slab of glass can be ignored with no knowledge of the builder is in my opinion irresponsible.
The other contributors understood the goal of the OP and were trying to help him with his project, not impart a greater understanding of his materials.
My point is not that the science is flawed. I used the numbers you said were "in the reasonable range" because for an artisan working at a bench they are, and using your numbers produced my results.
In my shop I round measures to about 1/32", sometimes refining with an up or down arrow, or maybe a "split the line" or "take the line" notation.
If you are reworking your home with a dial caliper you'll be busy until well after the pandemic. Apology accepted.
You caught me at a week moment, after a good mint julep. Your problem is trying to solve a trig problem with paper, a ruler and a straight edge. This is like trying to cut dovetails with a hammer and screw driver. Possible, but inelegant.
Just so we are on the same page, I can tell you that the angular change that is produced by seasonal changes in moisture of the joint are the same regardless of the width of the board.
Assumption: Total seasonal moisture change is 6% and rate of expansion is 0.00252 per percent per inch.
The formula for calculating the change in dimension is: width * % moisture change * rate of expansion
So the new width is : width + width * 6 * 0.00252
Now here comes some algebra, I am going to simplify this to:
width*(1+6*0.00252)
If the original angle is 45 than the two legs of the triangle are the same. With the seasonal moisture change, the new triangle will have one leg with a length of the original width and the other leg with the length calculated above. And I realize we could also subtract the seasonal movement change. Just going with addition like your drawing.
Now from trig we know the tangent of an angle is the opposite leg over the adjacent leg. I'm going to assume the adjacent leg is the longer leg after seasonal movement. This doesn't matter since both are 45 initially.
If I know the two leg lengths, I can calculate the new angle using the arctan function.
This would look something like this:
arctan of the angle = opposite/adjacent
plugging our calculations in yields:
arctan of the angle = width/width(1+6*0.00252)
The width on the top cancels with the width on the bottom, i.e. any number divided by itself equals one.
so simplifying:
arctan of the angle = 1/(1+6*0.00252)
As you can see the width fell out of the equation and all that is left is a constant term. I calculate an angle of 44.6 degrees.
If you want to think of this in geometric terms and skip the algebra, the initial triangles for the 2" wide board and the 3" wide board are similar (have the same angles or the ratios between the sides are the same) The new triangles will also be similar be similar because the ratios between the two sides remains the same. This happens because the amount of change in width is directly proportional to the width.
Trying to draw this on graph paper produces small errors and they compound themselves to yield your results.
Yes, as I said earlier and you cannot seen to grasp, I was baiting you. I understand the math. The actual issue is that while the rest of us were working on a question involving a ruler and a bit of wood to help Mike build a frame for his mirror, you decided we were all somehow freshmen students in your Trig 101 class.
I will order an arctan from Harbor Freight on Tuesday. I hope it comes in brass, I don't want a plastic one.
I miss Badger.
MJ , you are right, I was not keeping the lenght the same.
MJ, of course I knew you were baiting me. I called you out by name because you have been a bit of a jerk on the board recently and there has been a bit more trolling behavior than I think appropriate for a board designed to educate new woodworkers.
I then dropped a bunch of troll bait in my posts like: I think this thread should die and that I am here because you seem to want to learn something.
And you bit.
And you had a bunch more jerky, condesending, know-it-all comments on my posts. And I acted like the engineers you know-it-alls love to troll.
I have enjoyed trolling the trolls and in the process I found some guy who did some amazing, masters thesis level testing on wood glues. Nothing like entertainment and education.
My problem with your baiting me is that you purposely put out bad information and if I don't correct it, some poor new guy may take it as fact. If I respond, you get to try and make me look like a jerk.
And I was not trying to give a less on trig, but you went there. I have no clue if your little "it do" response was that my first post explaining it was understood or not.
I make my post to clear up some information that seems to float around on these boards that can be safely ignored my new woodworkers. So I too was helping Mike design and build a frame.
For the record, I'm not an engineer anymore. When I was working as an engineer, I did software design. I'm just a guy with 30 years of woodworking experience, some basic engineering knowledge, and a good BS detector.
I know a good troll likes to have the last comment, so put one more up there and I will not respond.
Jake, old troll (interesting admission you made there, that) ....
You boast that:
"For the record, I'm not an engineer anymore. When I was working as an engineer, I did software design. I'm just a guy with 30 years of woodworking experience, some basic engineering knowledge, and a good BS detector".
I feel you do yoursen a disservice, as you are much more than that! Your fine posts encapsulate much that is, er, interesting about a certain type of yank fellow blessed with a large dollop of that self-esteem that's so popular over there. Should I wish to perfect a mode of braggadocio and reedy-voiced, closed-ear pedantry with which to impress lesser folk in a fails-to sort of way, you're my prime exemplar!
What a fine specimen you are in this anthropological laboratory of interesting experiments. Keep it up, especially the inchoate math gobbledegook.
Lataxe, loving this thread.
PS I hope the OP is happy with his extraction of a framing joinery from all this noise. :-)
Thanks Lat_axe, really all an act. Just playing the role of a know it all engineer. My actual day job cuts you down on a regular basis. I pretty much expect the worst, hope for average and if it looks better than average assume I misinterpreted the data. I’m like belt and suspenders and suspenders on my boxers. Just in case my pants fall down, I don’t want to flash anyone.
Lat_axe, one of these days I would really enjoy a conversation with you. I wonder if you are as "real" as you write?
Mr 268,
Conversations can be had here, at any time we wish. Whilst they might commence with the WW topic who knows where they'll lead?
One feels that some lads (and also any lasses about the place) would do well to drop their inhibitions and just come out with it. Well, that's what I do. Out it all comes!
We might discuss what the "it" is. Jake will tell me it just needs a couple more letters for that to become clear. It can be difficult, sometimes, to distinguish between pearls and rabbit-dottle, mind. :-)
Lataxe
So, to be sure I've got this.. you invented a person to "single me out" me for posting correct information, and I'M a troll? I'd suggest that you have a sit down with Badger, but I think that's as likely as Batman having lunch with Bruce Wayne.
I'm Batman.
Ha.
This forum post is now archived. Commenting has been disabled