Monday, October 12, 2009

Bracing Situation II: Tension and Compression Considered

I let a few more days pass than usual between blog posts, in the interest of giving readers a chance to respond to the poll attached to the previous post. Not being a professional pollster by any stretch of the imagination, I am left to wonder whether the way I set the poll up, or the range of answer options I provided, was entirely useful. For instance, over the past 5 days, the blog has been visited some 1000 times, and the total votes were, at the end of the poll period, 20. Since the poll mechanism is completely anonymous, I was in fact a little surprised at the apparently 'low' number of votes (though I must confess I have no real idea whether that number of votes, given the context, is 'low', 'high', or 'average'. I wonder if some readers declined to answer the poll because they were unsure, or had no idea what to answer - - thus having a poll answer selection of 'Don't Know' might have been helpful. I'll leave it to readers to let me know in that regard, if they feel like doing so.

It was fun to do a poll nonetheless, and I thank those of you that voted and the one reader, Don, who opted to comment as well. So, now that the poll is done, it seems that a slight majority thought the 'up' brace type was structurally the best, with a second large contingent backing the 'X' brace arrangement. Both 'down' and 'it makes no difference' received but a couple of votes each.

I find it a bit odd, to tell the truth, as I look around New England at the braced gates and doors I come across, that there is such a variety of brace forms used. One would think, upon seeing such variance, that the builders of these gates were not entirely sure which way to go about bracing their structures. Even if one was to presume that the makers of the gates had no idea whatsoever about how a triangulated hinged structure like a fence gate should be braced, one would imagine that, just based upon chance alone, one would observe roughly equal appearances of the 'up', 'down', and 'X' -form varieties, however such is not the case. From my observations, and they are hardly extensive or scientific, it would appear that about 70% of the gates and hinged doors I see are built using the 'down' brace. And the down brace is, as Don put it in his comment, pretty much the worst method of all:

"Building a door or gate other than with the brace stemming up from the hinge side would be seen as a measure of incompetence which I would agree with."

Sharp-eyed readers who examined the photos a bit more closely, would have observed the typical result of 'down' brace construction, namely that the gate will inevitably sag. Here's a close-up of the cemetery gate:

Note how the space between gate stiles is open at the bottom and non-existent at the top. Notice how the lower end of the brace has become partially separated from the stile. While it is possible that the gatepost itself could be the culprit for the sagging gate (which would be a sign of poor anchoring/bolstering of the post base), a view from further back shows clearly that the gate has sagged and not the posts:

And a closer view of the other down-braced example, this one with doubled sets of braces:

It is apparent that there is rather more space now at the bottom of the door interfaces than when the door was first hung. The craftsmanship of the doors is decent enough, to be sure, however the builder fundamentally didn't understand how to brace a structure like that. I think such a misunderstanding stems either from a failure to consider how a gate is loaded, or, and altogether more likely, stems from a failure to understand the basic nature of wood as a material. It may also be a misunderstanding, rather than a failure to understand -however at very least it would appear that the builders of gates and doors employing down braces had not spent much time looking at older examples of such construction to note which were the successful designs and which were the failures. And, it may also be said that sometimes people do not understand that what works in one material may not work so well in another, and braces are a very good example of that.

For instance If these gates had been built with metal rods/cables, in the down direction, instead of wood, the construction would have been just fine. the same can be said for using 'X' form bracing, though for slightly different reasons which I will detail soon enough.

Metal, in whatever form, is an isotropic material. That word, isotropic, for those unfamiliar with it, combines the prefix iso-, meaning 'same', and the Greek root tropos, meaning 'turn', or 'direction'. Thus isotropic means to have the same properties irrespective of direction. If you pull on a piece of metal (i.e., produce a tension load), or push on it (a compression load) the behavior of the material is identical. Put another way, metal's ability to resist loads as a material is the same regardless of the direction of the load. Of course, the shape/configuration of the piece of metal has much to do with how well it might resist a load from a given direction, but that is another matter.

Wood, unlike metal, is anisotropic. This word has the opposite meaning from 'isotropic', the prefix aniso- meaning 'unequal'. Anisotropic materials, like wood and concrete, show differences in property or effect in different directions. The reason for this, naturally, is because wood of the type generally employed by man for structural applications grows in an orientation perpendicular to gravity: straight up. Of course, wind and snow impart loads on a tree as well, and these serve to shape the tree (the more so if the loads are frequent), but the one constant is gravity. The effect of gravity becomes more significant as the tree grows bigger and the mass of the tree increases. The fibers of a tree are aligned then to the direction of growth, which is a longitudinal one perpendicular to the ground in ideal cases.

Another term, and a slightly more precise one, that may be applied to wood is that it is orthotropic. Ortho- comes from the Greek orthos, and means "straight, upright, right". An orthotropic material, like wood, has elastic/mechanical properties which vary in the directions of three mutually perpendicular axes. Here's a picture showing those three principal axes in relation to the direction of grain (fiber) on a piece of wood:

The tangential and radial directions shown above are referred to as being perpendicular to grain, and the longitudinal direction is parallel to grain. Here's the important thing that a woodworker needs to know in reference to the above drawing: the properties of wood parallel to grain are higher than those perpendicular to grain. Not a little bit higher - a lot higher. The reason for this difference is that the grain direction of the wood is also the direction of primary chemical bonds and fiber alignments of the cells of the wood.

Typically, with a mature tree (as distinct from juvenile trees which have differing fiber characteristics as a result of their juvenile core), as you move from the bark in towards the pith, passing through what is termed the neutral axis (which would be halfway between the pith and the bark in a perfectly straight 'ideal' tree), you will move from a zone of tension to neutrality to compression. The outside band of the tree, as it experiences loads from the wind blowing, develops a tendency towards tension. This fiber tensioning is akin to using a perimeter of cable stays on a tall antenna or mast, the stays working to resist tension loads, and tensioned themselves to keep the mast held in a fixed position. As one moves from the tension-bearing outside layers of the tree to the pith, there is a shift towards tissue that exists almost entirely in a state of compression.

The differences in zones of tension or compression in a given tree trunk are manifested more starkly when the tree is cut and processed into timbers. At the outside portion of the tree, the cutting of the tissue it is a process somewhat akin to cutting the tension wires that fixed that antenna in place - as they are cut, there is a recoil of the wire back towards a non-tensed state. For those parts of the tree that were in compression, there is a push back, or expansion towards a non-compressed state. The result of these tension releases is often the bowing of timber towards the bark side, as in this dramatic example of a Walnut trunk which has been partially quartered by the release of growth stresses:

The above photo comes from an article by Daniel Cassens and Jose Serrano, as part of the Proceedings of the 14th Central Hardwood Forest Conference (1986).

While the tree grows with gravity as an omnipresent constant, it is a fact that loads from high winds and snow can impart severe stresses on the fibers of a tree, and these loads often exceed the direct to fiber loads from gravity. Thus the fibers which compose wood as a material are have evolved to be stronger in resisting tension than compression. Look at this table comparing the mechanical properties of 5 wood species:

My apologies for the small size of the text in that picture. As you can see, I hope, in each case (except for Balsa, and why is that listed in the table anyhow?!) the parallel-to-grain performance in tension is significantly higher than the parallel to grain performance in compression - nearly double on average, and in the case of Yellow Poplar, there is an astounding fourfold difference between parallel-to-grain tension and compression figures. Notice too from the above table how much poorer wood is in tension or compression perpendicular to grain compared to parallel to grain performance.

It should be apparent to anyone who has pulled and pushed on small trees or sticks of wood, that it is far harder to rip the tree trunk apart by pulling on one end (assuming the roots hold!) than it is to press down on it and break it. Of course, when one imparts a downwards compression load on a slender trunk or piece of wood, we never actually find out how strong it is in compression - directly speaking - since the gradually increasing load causes the piece under load to ultimately deflect sideways, and only then does it fail. When the wood buckles sideways like that, the load on the wood shifts direction, and what results in the piece being loaded unequally, yet still parallel to grain, one side of the material now in tension and the other in compression. The compression side fails first. If you bend a tree branch or small stick hard enough, you will note the ripples of compression failure that form on the inside of the bend, as the material fails sooner in compression.

Wood is very strong in compression parallel to the grain. In fact, a hickory chair with 1.25" square legs could support a load of 32,430 lbs (I take this example from Bruce Hoadley's fine book "Understanding Wood", pg. 112). In fact, a chair with only 1/8" thick legs, if they could be supported so as not to buckle when in compression, would easily support a 250lb sitting person. Therefore, a piece of wood which has to bear loads parallel to grain either needs to be supported in such a way as to not deflect under load, or, as is more often the case, needs to be 'over-sized' (in relation to it's strength in compression parallel to grain) by some margin so as to adequately resist bending.

From the above table it would appear that designing wooden structures around the superior strength of wood in tension would be the best way to go. However such is not the case, and in the next post I'll delve into the reasons why it is usually wise to design wooden structures around compression rather than tension, and consider the loads that act upon a triangulated structure such as a hinged gate and how they might be best dealt with in terms of sound design.

Thanks for visiting and see you next time. Go to part III.

No comments:

Post a Comment

All comments are moderated:

Anonymous comments will be deleted

Spam will be deleted

Comments containing links unrelated to blog content will be deleted