Copyright © 2015, Charles S. Greenwood, P.E., LLC * Do Not Copy without Written Permission
An Overview by Charles S. Greenwood, P.E.
Trees support vertical and horizontal loads by resisting the crushing forces acting on sap wood and the inner heart wood; very little load can be supported by the living cambium layer located just under the bark. Compression parallel to the grain supports gravity loads in tree houses and other tree attachments. Compression perpendicular to the grain resists horizontal loads such as zip lines, bridges, wind and seismic forces from structures, etc. Wood is much stronger parallel to the direction of the grain, (5 to 20 times the horizontal strength) and this presents a significant problem when attempting to support horizontal loads. Therefore, this engineer recommends reducing the externally imposed horizontal loads as much as possible; obviously, a significant amount of horizontal load will transfer in any case from the tree house itself into the tree. Typical mounting methods include an inserted pin of some diameter, adequate for the loads involved; this can be a nail, a lag screw, an arborist shackle mount, a special lag screw or stud with an integral shear collar, or a pipe or tube going part way or all the way through the tree. All have been used, and all work when applied correctly. All penetrations can involve risk to the health of the tree, so proper practice is to minimize the total area of insult to the living cambium layer, and to treat the injury. Cinch rings damage a much larger area of the cambium and are not recommended except for small horizontal loads and very small vertical loads, and never where the contact area on the bark is a significant proportion of the total circumferential affected area. It is recommended practice to use doubled fasteners – one over the other – either a lag screw attached to a turnbuckle tied to a typical TAB (Treehouse Attachment Bolt), or, in more sophisticated designs, a complete integral moment bearing beam support frame.
Remember an upper fastener there will be subject to a pull-out load due to the triangulation inherent in that design. Also, the upper TAB may support a large portion of the load and therefore needs to be at least as strong as the lower TAB. Finally, all loads in trees create a biological response that induces the tree to add significant “bolstering” at the point of stress; this can occur in as little as one year and may become the dominant fastener consumption mechanism.
One must decide how many years of service your tree house is going to provide before major re-mounting is required. Experience is showing that this becomes an issue much sooner than most builders expect. Extendable mounting schemes are now coming into use to provide for easier “planned” re-mounting; having one’s beam lines and joists laid out in an easily movable pattern helps to facilitate this. Of course, many older tree houses have simply been partially consumed by the host tree(s) but the jury is still out on the effect of this type of damage. One school of thought is if you make the beams out of the same material as the tree they may not reject the tissue as they are consumed. On the other hand it may kill the host tree.
Crushing support from the grain bearing against the TAB is analogous to a system of coil springs (literally each ring) supporting a horizontal leaf spring that is almost rigidly attached at its inner end. Thus at the inner end some of the compression force must be resisted at the top of the fastener to counter balance the upward support along the shank (or collar) as it nears the outside surface. There is almost no bearing strength over the outside inch of the fastener pilot hole, so it is up to the crushing strength of the wood several inches in to the tree to support your tree house. This why it is common practice to increase the bearing area by using large diameter collars of sufficient length – indeed, the same 3” diameter penetration through the cambium layer may provide three or four times as much support by using a long collar.
Metallurgical properties are as important for tree fasteners as any other critical use fastener. Specifications advocated by this engineer are to anneal after machining followed by quench and tempering to produce a Rockwell “C” hardness of approximately Rc = 35 up to Rc 45. With 4140 alloy this will achieve yield strengths from 100,000 psi up to 185,000 psi. Through- hardening is essential since surface hardening (“case hardening”) leaves the core of the fastener without spring steel properties. Since stress reversals often occur many times per day, it is predictable that without proper alloying and heat treatment, the steel will fail – just like putting a piece of metal in a vice and bending it back and forth until it fractures. Since the fracture will begin at a discontinuity on the surface, it is (or should be) absolutely forbidden to install a TAB with a pipe wrenches or other devices that can injure the highly-stressed fastener surface.
The use of the terms compression, shear and bending are often used almost interchangeably both by novices and engineers who should know better (myself included). Compression as described above is the dominant property for the support of loads attached to wood, but the load may also be placed where it acts in pure shear, or in bending, or in any combination of the two. Shear parallel to the grain acts when the fastener bearing surface has been overwhelmed allowing the grain to be split into two – something we really don’t like to see except when making kindling.
How do we know how much load a particular fastener design will support at a specific location in an individual tree? While we have developed reliable mathematically models to predict the probable performance, we really don’t know for sure until we follow the testing procedure presented at the end of this document.
TREE HOUSE SUPPORT SYSTEMS
An Overview by Charles S. Greenwood, P.E. ●
LOAD COLLECTION: How big is it? What do the components weigh? What does it all weigh? How is the weight distributed? What’s the seismic load? What’s the “sail area” (wind load)? How about the live load? Snow load?
● LOAD TRANSFER: How do you take that entire load you’ve collected and place that into the least number of fasteners possible and have it work properly? We are seeing a variety of problems with inadequate mounting methods. What can be done about it?
● LOAD SUPPORT: Now your loads are bearing on the trees themselves. How do trees react to external loads? How much is too much? When is a tree too small to act as part of the support system? What happens when we mix “crutches/helpers” and trees together?
● PERMITS: Do you need one? How do you get a permit if you need it? Why is the International Residential Code not very useful, but International Building Code is?
Tree Houses mounted in a single tree will use the classic “knee brace” or “hanging cable” solution; some of the vertical loads and typically all of the horizontal loads may be transferred directly into the stem of the host tree. The major disadvantage is that at some point the tree must penetrate the plan view, and probably (except in the case of very large support trees) the roof as well. Deflection of the host tree over the vertical distance of the triangle created by a knee brace system will attempt to bend the building frame and/or the knee brace itself as the tree is likely to be a great deal stronger than the knee brace/frame system. Something has to give. Cables of course are laterally non-supporting but all of the horizontal load must be resisted by the interface between the house frame and the tree itself; some kind of fastening system will at least prevent destructive “bark banging”. Very small backyard tree houses are good candidates for an adequate single tree.
TWO TREE SUPPORT SYSTEMS
A second support tree offers major advantages, along with a few added difficulties. Simple twin-beam support systems will often suffice, closely mounted at each side of larger diameter trees, or with the additional complexity of a moment frame (probably steel) with multiple fastening points. Robust joists or cross beams can cantilever outward in a balanced configuration on each side of the parallel beam lines, providing floor area quickly. Once again, something has to give, so it is typical practice to mount the beams solidly to one tree and with UHMW (Ultra High Molecular Weight Polyethylene) wear plates at the other. For commercial and public use projects it is recommended to completely enclose both the fixed and sliding connections with steel “clasp brackets” that includes UHMW on all surfaces; this will allow movement AND prevent moisture induced beam rot as will be the case in a steel-to-wood interface. While this arrangement favors a rectangular plan view, one may easily get creative. Adding knee braces to this type of support system can provide for greater spans, but remember that something has to give and it will probably be the frame itself – some amount of twisting and deflection is inevitable as the trees will rarely act in concert during wind events. Torsional loads also come in play with multiple support tree designs and this inevitability must be provided for. Fortunately, if the support beams are robust and not too rigid in plan view, and well-mounted with the beam mounts described above, there will likely be sufficient￼ flexibility to allow for normal tree movements.
THREE TREE SUPPORT SYSTEMS
Now we’re getting creative, as the building and structural design can become much more interesting. Usually one tree will act as the “master” and the others will act as “slaves” as far as differential motions are concerned. Beam connections must be free to move with respect to each other. This is where the analogy to early American railway locomotive design becomes significant. The European practice was to build very precise rail alignment on very stable roadbed, while the American practice was just the opposite – there was a lot of distance to cover, and the investors wanted to make the highest rate of return as soon as possible, so they made the crappiest track possible. Unfortunately, the engines in particular left the rails often. Rather than improve the track, designers realized that clever floating support linkages amongst the wheels and engine frame – suspension with load distribution and equalization – would keep the locomotives on the track most of the time. For tree houses, multiple sliding connections (with 4-bar knee brace links when required) provide load equalization during differential translational movement.
FOUR (OR MORE) TREE SUPPORT SYSTEMS
All of the above applies when you are blessed (or cursed) with numerous support trees. Remember, the complexity of the support system goes up geometrically with the number of support trees involved.
ARTIFICIAL TREES – GROUND MOUNTED SUPPORT POSTS
This can be a handy way out when you are unwilling to make your design fit what nature has provided, or your construction is too heavy to be supported by the available trees. A “tether ball pole” is a good starting point for this discussion, as most of us have put concrete around a car wheel and mounted a pipe to it and then rocked its world. The idea here is to keep the system from acting like a bowling pin; these pins are also a gravity stabilized system, but one that is likely to fall over if the horizontal force exceeds the vertical stability moment. Wind machine and utility towers, and for that matter, sky scrapers, are large examples of a ground supported and gravity stabilized vertical cantilevers. Tree house posts can be rigid with slider tops, or flexible if your design is elegant. FINITE ELEMENT ANALYSIS OF TREES AND TREE HOUSES For any but the most basic tree house designs, simple 2-dimensional statics based analysis, typically done by hand, is inadequate. Due to dynamic loads, trees and tree houses are best analyzed as 3-dimensional flexible objects akin to wooden boats or even space craft. This engineer has developed specific FEA techniques appropriate to tree house engineering that allows all manner of load combinations to be explored and demonstrated. The following screen shot shows the Laurel School tree house near Cleveland, Ohio, under maximum vertical and horizontal loading. The picture on the next page show this tree house under construction.
MULTIPLE FASTENERS BEING TESTED IN LIVING TREE TREE FASTENER TESTING PROCEDURE
Precise load rating of each fastener is possible by using the following methodology:
1. Install the fastener into the subject tree at the design location. Use Cut Guard or a similar product to ensure proper healing of the installation injury.
2. Attach a hydraulic jack with a calibrated gage BELOW the fastener. This may require a ground pad or a temporary attachment to the tree itself. Do not injure the roots nor the tree surface when doing this step.
3. Attach a dial indicator with at least 0.100” sweep so that it can measure the deflection of the fastener relative to the tree when the load is progressively applied. The point of action is typically placed 2” away from the bark for both the dial indicator tip and the load interface itself.
4. Observe and record the load at a deflection of 0.030” (0,75 mm). Relax the load to zero and record the permanent deflection (if any).
5. Repeat this sequence until the permanent deflection is approximately 0.030”; this is the rated maximum load for this fastener mounted at this spot into this tree.
6. If the results are as expected it may not be necessary to repeat the test for additional fasteners installed into this particular tree. Once multiple data points are obtained for the species of trees at this location it may not be necessary to conduct further testing. Designs of any size that use a cylinder (or several axially referenced cylinders) will unfortunately leave a 0.030” crush zone at the top of the fastener-tree interface. Treat this with Cut Guard or similar product to speed the healing process. The design load imposed upon the connection will act in the opposite direction from the test load, so support strength is not reduced by the test. The XL Tree Fastener (Patent Pending) overcomes this liability by allowing the tapered shank to be further torqued – about a half a turn – into the tree, thus closing the injury. It further solves the problem of fastener consumption by providing a method to infinitely extend the load collection point. If the design load is greater than the load maximum test value, then a second fastener can be added some distance above or below the subject fastener and the proposed design can likely be implemented.
Copyright © 2015, Charles S. Greenwood, P.E., LLC
Do Not Copy without Written Permission