## Tuesday, December 3, 2013

### Can't See the Forest for the Trees (II)

I'm not a logger or a sawyer, so I can't speak exactly from those perspectives. I have logged and I have sawn logs many times in my life, not enough to be any sort of expert in either matter, but I'll offer what I can on the subject.

Let's say you find some logs, like what you see, and you want to buy them. Or, perhaps you have some logs and want to sell them. What if you wanted to sell logs internationally? How is the price to be negotiated?  By number of logs? By length, by width, by volume? By the amount of usable timber or by the total volume of the log including bark? Or without bark?

Saw logs are conventionally sold on a 'cubic meter' volumetric basis, and this has been the case since the late 18th century.

You would think that a 'cubic meter' - 1m. wide x 1m. deep x 1m. tall  -would itself be a fairly well-defined and universally agreed sort of thing, but it is not so when it comes to logs. A cubic meter here is not the same as a cubic meter there.

It boils down to the one question: how do you measure a log?

What comes into consideration though, before the matter of measuring of volume, is the assumption as to which geometric shape the log conforms as we calculate volume based on a geometric form. The candidates are:

1. Cylinder

2. Cone

3. Paraboloid

'Paraboloid' itself can be considered a simplification. We might also assume the tree trunk has varied forms according to which section of the trunk is considered:

Those seem to be the choices in front of us. Notice in all four that the base or horizontal plane cutting through the geometric form is a circle. Any horizontal plane cutting through the log would also render a circle. Trouble is, trees are rarely, if ever, perfectly circular in cross-section, and some are a good deal non-circular. We could have added in an illustration of a cone with an elliptical base, for example. Not that any tree has a perfectly elliptical cross section mind you!

No tree trunk is a pure cylinder, however old Sitka Spruce trees do look a lot like gigantic pencils in the forest and the logs therefore come very close to cylinders. Some trees are much more tapered in form, like a cone. Cedar trees come to mind. Some tree trunks taper evenly, others stay fairly cylindrical and then taper radically at the crown portion. Some trees flare strongly at the butt, others not so much.

Between the forms of a cone and a parabola, neither is an absolute ringer for any tree I've come across, however the cone is certainly a simpler form upon which to base calculations. Assuming a conical shape for a log was the standard way to do things way back in 1765, as noted by Oettelt in Germany.

Assuming that a log is a section of a cone is a simplistic one, however the fact is that no tree truly conforms to any geometric form, not precisely at least. So we proceed from a simplistic assumption and nothing more, simplicity taken as a virtue over ultimate accuracy. Refining the accuracy of the measurement is seen to be more expense in term of the time involved than is realized as a benefit resulting from the extra effort. Relative accuracy, therefore, is what is sought, and that is simply the maximum accuracy that is profitable and possible to obtain in practice.

Right, given that, we can at least start from the basis of assuming each log is a portion of a cone, and calculate volume from there, yes? The formula for determining the volume of a cone is easy to remember, it seems to me, IF you find the formula for the volume of a cylinder easy to remember. How's that for a bit of, er, circular reasoning?

The volume of a cylinder is that of the base circle (π r2) times the height (h):

A cone is simply 1/3 the volume of a cylinder, or 13 π r2h:

An easy way to think about how a cone and a cylinder relate to one another is to imagine a pair of paper drinking cups, one a cone and one a cylinder. Both have the same total height, and the same cicular base. If we filled the conical cup with water or salt, say, and poured it out into the cylindrical cup, it would fill that cup 1/3 of the way:

One can see thereby the vastly different estimate of volume one would obtain whether one were to consider a log as a cylinder or a cone.

One niggling point though in regards to cones: when we are looking to buy logs, we're not typically looking for the pointy end of the stick (i.e., a pointed cone), as that is the end with all the branches. That bit is perhaps best left behind in the forest. We are looking instead at a section of the trunk, which is a truncation of a cone. Another name for a truncated cone is 'frustum' cone, a frustum being that portion of the cone laying between the two parallel planes cutting it:

The word frustum comes from Latin and means 'piece', or 'crumb'.

Now, the formula for a frustum of a cone is a little bit trickier than the formula for the volume of a cone. A simple way to do it, if you had both portions of the original cone before truncation, would be to perform two separate calculations, one for the entire uncut cone, and a second for the piece which was lopped off on top. Subtract the volume from that upper piece from the total and you will find the volume for the frustum. That's an inefficient way to do things, and it is also the case that when looking at a log you only have the frustum to consider. So we need a formula just for that portion:
The volume of a frustum of any cone is equal to one-third of the product of the altitude and the sum of the upper base, the lower base, and the mean proportional between the two bases.
Put into mathematical language, the above sentence becomes:

The uppercase 'R' in the formula denotes the radius for the base of the frustum, and the lowercase 'r' is the radius of the upper cutting plane. The area of the base is given by π R2 while the area of the upper plane is π r2 - you can see those formulas as part of the larger formula above. The piece under the square root symbol in the formula above, “√ π R2( π r2)” is the "mean proportional between the two bases". The 'mean proportional' of two different quantities, let's call them a and b, is a quantity r such that a is to r, as r is to b (a:r = r:b). for instance, the mean proportional between 2 and 18 would be 6 (as 2:6 = 6:18). The Golden mean, 1:1.6180339, is a particularly famous proportional, known to many readers I’m sure.

Consider then that the formula (a:r = r:b) can be simplified algebraically to r2 = ab. Solving for r, we get: r = √ab, and that is the core idea of the portion of the formula above we see as π R2 wh(π r2 wh). Of course, when we see a formula in which we take a square root of an entity that is squared, in effect we cancel out both the square root and the squares, as √62 wh = 6. Therefore, π R2 wh(π r2 wh) becomes simply πRr. We thereby produce this formula:

This formula can then be cleaned up a bit by extracting π out from the parentheses:

Well, not quite so easy to remember as the formula for the volume of a regular cone, but it does fall short of what one would call a complex formula. The math geeks out there will possibly be yawning at this point.

So, let's say we have a log, and it measures 300 centimeters long, has a butt end of 100 centimeter diameter (R of 50 cm), and a skinny end of 80 cm diameter (r of 40cm). By the above formula we would write:

V = ⅓ π ((502 wh+ 402 wh + (50 x 40)) x 300

That gives us:

V = ⅓ π (2500 + 1600 + 2000) x 300

Simplifying to:

V = ⅓ π (6100) x 300

Pi (π) equals 3.14159 or thereabouts. To complete the solution, we obtain:

V = ⅓ (3.14159 x 6100) x 300
V = ⅓ (5,749,109.7)
V = 1,916,369.9

The answer, since we reckoned the problem in centimeters, is in cubic centimeters. Logs are sold on a cubic meter basis, and 1 meter is 100 cm. Therefore, 1 cubic meter (1 m³) equals (100 cm/m)³ = 100³ cm³ = 1,000,000 cm³. We therefore divide the above obtained value in cubic centimeters, 1,916,369.9 by 1,000,000, to obtain the volume in cubic meters: 1.916.

Keep in mind that when we performed the above calculation we were working with nice round numbers, however lengths and diameters of actual logs are more often given with nominal or rounded forms, possibly in 2-foot length multiples, for instance, or the calculated answer may also be rounded, the 1.916 above being rounded to 2 cubic meters.

If you measured the log in inches or feet, it is much the same process, however at the end one would need to convert over to cubic meters from cubic feet or inches. One cubic meter equals 35.315 cubic feet. The conversion is simple enough: take the volume in cubic feet and divide by 35.315 to obtain cubic meters. If working from cubic inches, divide by 61024 to obtain cubic meters.

If only things were so simple as plugging data into a formula, but the situation turns out to be very much different in the world of log scaling today. Because taking a simple mathematical calculation of a log to obtain volume does not give us the actual likely amount of convertible timber, does it? Logs vary, and in some species by quite an amount. Between species it becomes more varied yet. Some trees, for instance, have thin bark while others have thick bark. In the US, log volume excludes the bark but in other countries the bark is incorporated in the calculation. Some tree trunks contain enormous amounts of sapwood with little commercial value (Gabon ebony would be a good example). American Black Walnut is a tree which also has a hefty band of sapwood, however the case is different from the ebony as the wood can be steamed so as to homogenize the color of the sapwood with the heartwood. Young trees have a higher proportion of juvenile wood than do old growth trees. Some species are prone to defects of various kinds, or have characteristic areas of rot, also limiting the amount of convertible timber. Some logs are almost exclusively converted into veneer, while others are nearly always sawn for timber. And on it goes. Log scaling is simply the process for estimating the weight or volume of a log while allowing for features that reduce product recovery.

One could come up with different scaling methods to calculate the useable extractable volume of timber from a log on a species by species basis, however it is fortunate that things haven't gotten quite that fractionated - but it is close. Welcome to the world of log scaling, where complexity, as they say, rules.

In reckoning log volume, some systems average the log end areas, while others average the log end diameters. Some assume different geometric shapes for the log. Here are 7 basic systems, with the actual formulas left out:
1. Smalian (assumes a parabolic log shape, and is the standard log rule (in metric form) in British Columbia and the basis of the Interagency Cubic Foot scaling System)
2. Bruce's Butt Log (as Smalian's formula assumes a paraboloid shape, it tends to overestimate volume of butt logs. David Bruce came up with his rule in 1982 to adjust for the butt portion of the tree trunk)
3. Huber (this formula assumes the average cross section area is found at the midpoint of the log, an assumption which is not always true of course)
4. Sorenson (derived from the Huber formula and assumes a taper of 1 inch per 10 feet of log length. The taper assumption is not always correct however)
5. Newton (the most accurate, requiring measurement of both end diameters of the log as well as the midpoint diameter, however this measuring is more time consuming to execute)
6. Subneloid (similar to the Smalian, however configured so as to allow a multiplication by 12 to obtain the board foot measure, becoming the "Brererton Log Foot Scale")
7. Two-end Conic (assumes the log shape is a cone, and is the basis of the "Northwest Cubic Foot Log scaling Rule")
There is also the Hoppus rule, derived in Britain. It is the most widespread system internationally; assumes an assumption about processing loss. Sometimes called the "quarter girth formula", it considers the volume of a log in cubic meters to be:

V = (C/4)2 wh x L/10,000

Where, V is volume, C is log circumference in centimeters and L is the log length in meters. The result obtained by this formula can be compared to the math formula for volume we worked earlier, using the same numbers. We would have to pick a circumference though, so I'll average the small end (r = 40cm) and the large end (R=50cm) to obtain a radius of 45cm. Radius converts to circumference as 2πr, so a 45cm radius gives a circumference of π90, or 282.743...

Plug that into the formula and let's see what we get:

V = (282.743 / 4)2 wh x 3/10,000

V =  4996.48722805 x 3/10,000

V = 1.49894616842 cubic meters

Compared to the strict mathematical estimate for cubic volume, the Hoppus rule gives a volume which accounts, on average, for about 78.5% of the cubic volume of the log-  the other 21.5% is lost as edgings, sawdust, and slabbed offcuts. Another name for the cubic volume measure obtained by the Hoppus rule is the Francon Cubic Meter, so as to distinguish it from the solid cubic meter.

As mentioned above, there is the Interagency Cubic foot System, developed in the US in 1991. This takes the Smalian formula and applies it to any log segment of 20' or less. If the log is longer than 20', then it is divided into segments and the taper of each segment is estimated as the difference between diameters at each end. This method aims to reduce the bias which results from the Smalian formula otherwise, which assumes a log to be a paraboloid. As you can see from the 4th picture in the series of geometric forms of logs shown above, the log usually does not conform perfectly to the paraboloid form.

That is an overview of the situation in western countries, however other places have different log rules yet. For instance there are four different log rules used in Japan (these rules are also used in Korea):

1. Japanese Agricultural Standard (JAS): the log diameter is measured at the small end only, and taken in two axes, which are averaged, towards obtaining what is called a 'scaling diameter' (D). The scaling diameter (D) receives adjustments which are done using a table. Scaling length (L) is in 20cm intervals. Volume = D2 whL / 10,000 if L is ≤ 6meters. If L is ≥ 6 meters, then the formula becomes V= [D -INT(L) -4) / 2 ]2 wh x L/10,000. "INT(L)" is the length of the log rounded down to the nearest meter, and the expression INT(L) -4) / 2 is a taper adjustment of 1cm per meter of length. This formula views a log as a square cant with a side equal to the scaling diameter.
2. Revised JAS: modifies the formula for logs 6 meters and longer with a factor to adjust the original taper assumption.
3. South Sea Log Scale (SSL): also known as the Brererton. Measure the long and the short axis of each log end and round them down to the nearest 2cm (i.e., 77.1 cm becomes rounded to 76 cm). then average the two rounded measures to the nearest whole centimeter. This gives the 'D' (diameter) value. Volume = 0.7854 D2 whL /10,000. This formula is the metric equivalent of the Subneloid formula described in the earlier table.This formula is typically applied to hardwood logs from tropical Asian sources.
4. Hiragoku or Heiseki scale: Uses the same procedure for finding the 'D' measure as the SSL method above. Volume = D2 whL /10,000. This method considers the log as being a square cant with each side equal to the recorded small end diameter, and there is no adjustment for taper.
In fact, in Japan the traditional measure of log volume has not been the cubic meter until recently. Much the same as Japanese carpentry, which is done to the measurement system of Shakkan-hō using shaku 尺 (30.3030cm) sun 寸 (3.0303cm) and bu 分 (3.0303mm), logs are reckoned by koku 石, a unit of volume equal to 10 cubic shaku. A koku was originally defined as a quantity of rice, supposedly equal to the amount required to feed one person for a year. The unit for koku, 石, is in fact the same character as is used for stone - if you find this confusing, welcome to the world of Kanji. For internal purposes, I would suspect a lot of log scaling in Japan is done on the basis of koku, however when importing or exporting, the koku value needs to be converted to cubic meters, one koku equals 0.2782 m³ (9.826 cubic ft.).

Indonesia, Malaysia and the Phillipines often use the SSL system, however there are regional variations in how dimeters are taken and recorded which lead to variances.

British Columbia, larger than many countries, has its own log rule system, the 'BC Metric Scale' which is the metric form of Smalian's formula. Here it is: V = (r12wh + r22 wh) L 0.0001570976, where r1 and r2 denote the top and bottom radii, rounded to the nearest even number, and L is the length in meters recorded to the nearest 0.2m, with the convention that exact odd numbers are rounded down (i.e., 13.5m becomes 13.4m).

Let's see how the BC Metric Scale compares to our previous reckonings using a 3m log, 40 cm radius at the skinny end and 50 cm radius at the fat end:

V = (402wh + 502 wh) 3 x  0.0001570976

V =  (4100) 3 x  0.0001570976

V = 1.93230048 cubic meters.

This formula lead to an answer in which the volume of wood was slightly greater than for the geometric form of the frustum of a cone, which is not surprising since the Smalian method presupposes the log is a paraboloid.

Chile uses the JAS method. Russia uses a system called GOST 2708-75, which has similarities to the JAS and Hiragoku systems, and to the Huber formula. New Zealand has its own systems, one for logs used domestically, and three other systems specially for logs exported to other Countries. For export to Japan they use a system which is an Imperial form (that is, in feet and inches) of the metric Hoppus formula described above. Or they will use the JAS system. For logs exports to China, likely to be the bulk of business these days, the William Klemme system is used, which calculates the volume a if a log were a simple cylinder.

In practice, a given location and a given mill will likely be using one scaling system all the time, so once a log scaler is familiar with the particular scaling system, it is fairly straightforward process. When international trade gets factored in, the picture can be a good deal more complex and buyer beware. Those selling logs may well have incentive to use systems which tend to bump up the apparent net volume of timber converted.

This post has been but an overview of the fairly convoluted world of log scaling. Not wishing to write a book on the topic myself, certain details have been omitted or simplified in the interest of brevity. There is plenty of reading to be found online in regards to log scaling, and I recommend interested readers have a look around.

Next post in this series will be a look at board foot measures for lumber and we'll find that the picture there is no tidier than the ones for cubic volume measure. Stay tuned, and thanks for visiting.

1. In Oregon, the log is assumed to be a cylinder based on the small end of the log and the excess is called "overrun" which the mill gets. I thinned my forest and sold logs from my property. After cutting, trucking and the low price the mill was paying, I got very little. If I ever did it again, I would have the logs milled into lumber on site and use or sell the lumber.
'http://nnrg.org/files/pubs-and-resources/measuring_timber_products.pdf

1. Andy,

thanks for the comment and the link. I see that Oregon uses the Scribner board foot system - a member of what's called a 'diagram log rule' type of system. Anytime you see systems based on board feet conversion you are seeing a system where the object is cutting 1" lumber. These systems, board foot log rule, Doyle rule, and diagram log rule, have their own particular assumptions- the log is a perfect cylinder, the boards are sliced at 1" thick, the sawkerf is a generous 1/4", and lumber widths are the standard 4,6,8... inches. Taper is ignored or crudely dealt with. The most common form of Scribner used in Oregon, is the Scribner Decimal C, which simply means that volumes are rounded to the nearest 10 board feet. The matter of overrun - - this is often used as a managerial tool, from what I understand, to monitor mill efficiency, however it can be manipulated by a shrewd operator, covering up real inefficiencies by creaming the most advantageous logs. You're certainly going to lose a lot of wood from butt logs and highly tapered logs.

More to the point though, by assuming a log is a cylinder and treating the taper as overrun, people will saw a log in a way which does not yield best fiber quality, as I will take up in the next post in this series.

~C

2. In my father's papers, I found a description of a logging operation during WWII. Part of the description included complaints about how some white oak logs were set aside as "mast wood" - 3+ logs long and paid for by the Navy based on the diameter at the top. The W.Va. locals were not amused - except for watching the loggers trying to move those pieces out of the mountains.

Regards,
Mike

1. Mike,

thanks - good to see you comment. It's curious to me that they would still be sourcing 'mast wood' in WWII, when the wooden ship era was all but over. Perhaps that wood went into something else besides 'masts'.?

~C

2. The three main log scales used in the U.S. is Doyle, Scribner and International. You always measure length and diameter of the small end. If the log is 30" at the butte and 24" at the top, you measure inside the bark on the small end and this is your number you work with. You need to cut the "conical" shaped log into a cant or cants, four sided sticks, that work on the mill. So the flare at the butte is taken off in triangular pieces that will usually just be waste, depending on size and all. The pith may or may not all be waste an the size of the pith can be quite large in some trees/specie. So some types we deduct for pith as well.

Doyle is used for hard cut logs, logs that need much chainsaw work to fit on the mill, defects to cut out or around , just plain pain in the butte log. The Doyle scale will show a few BFT less that the other to compensate for the added labor expense. Of course what you describe in the post is mostly straight soft wood specie scaling. Also scaling and grading are two separate functions. You grade the log to get the scale. International will yield the most bft and might just be called international because it will be used internationally.... Scribner is kinda in the middle. This is used for big straight easy cut logs. Cutting POC cants in to dimension lumber I did 800bft in three hours, mango three hours gets me250bft some days some days less :^(. I had a 24x160 Koa log that was pretty straight and yeilded 25% over scale giving 250bft in two hours. So each individual logs will give drastically differing results. What else could you expect from nature eh?
For logs of only 100~150 bft there is only about 5 bft or so difference between the three. deducting for wane and other defect is the real area of expertise as this is where you can get burned if you do not deduct correctly for the defect. Any log with any wind or bend will either be on the Doyle scale or counted as two shorter logs depending on the value of the material. Shorts not being worth much unless they are of an expensive hardwood specie. i.e. Curly Koa which holds it's value down thru 2 and 3 foot lengths which is way under grade for select, walnut allowing the most defect in the US right now for a select board, only needing one face good. I use Doyle most as most of the tropical trees are hard cut logs, bendy, twisty, a ten foot log often comes down to 3- 36" logs. Still good furniture and cabinetmaker stock.
It can be very difficult to determine where to begin busting a log into cants. The pith very commonly can be to one side of the log, diagonal thru it, or bend thru it. It's a crap shoot. You end up with loads of waste when you start trying to cut boards with grain parallel to edge when the pith runs diagonal thru the stick. Beautiful long straight log may have that diagonal pith running and all the lumber comes out with a slope grain whether it be to the face or the edge. You won't know if the tree is still standing but obvious after you drop it. Not good choices for a furniture shop but if it were a N. American softwood bound for the construction industry then all the material is milled over sized to compensate for defect as per International building codes. Allowing the industry as a whole to conduct businesses as usual instead of grading each individual 2x4 and 4x6 for strength and grading each for differing uses. i.e. load bearing,post, fencing,pallets etc....
I have never met anyone in the US that does not use one of the three scales above. The one of the things I have learned thru milling logs to lumber is you can never judge a stick of lumber unless you were the guy at the mill working the controls and squaring the log for cutting.
Good wood cost alot of . The process to get tree to board is a lot of sweat and hard work. The very best select is often a small portion of a tree. The really amazing wood is even a smaller percentage , that material never makes the market. This is why it is necessary to mill your own logs if you want unique or high quality material.

3. Correy,

thanks for sharing your expertise - it sounds like you've been doing a fair amount of milling yourself.

~C

3. Thanks Chris, Hardly an expert . The industry is so juxtaposed to what a small woodworking outfit might think the proper way of processing wood should be. I could never find wide boards, poeple do "not want to lift them" no one has the means to process them or much over 12" wide. Drying was a big issue, and I was tired of buying wood with tension set and uneven drying. Not to mention it is harder to go looking for wood for a project than wood telling you to make something with it. I just spent 6 month looking for two 19x120" boards for a project. To do nice work you can not wait for the project and then search for the wood. You need that inventory of material on hand to look thru. Theres no 銘木店 in my area LOL. I enjoy the lumbering, I find wood that I wouldn't otherwise. Wood is like $it needs to flow. I think it is good to process your own material . You get the best sell the rest, meet all sorts of nice people into woodworking. It's all good. 1. Correy, did you by any chance check into that guy who is based in Hawaii and has an enormous clear board of Eucalyptus robusta for sale on Ebay? In case you hadn't come across it: http://www.ebay.com/itm/HOLIDAY-SPECTACULAR-SALE-MOST-BEAUTIFUL-SLAB-ON-EBAY-I-do-NOT-want-to-sell-/111229628055 No idea about the working characteristics of that wood, but it looks like a nice slab. ~C 2. Yeah that's a big robusta board for sure. They get real big. Some on my road that are over 5' diameter. Really hard probably .85. Tenacious reversing grain. Takes a polish o.k. Also known as red gum, it's from the AU. Will leave gum corns on wooden plane block bottoms. I have to wipe the dai with oil alot when planing. Typically really unstable. Tress have huge piths 6" or more diameter so the bole needs to be really wide to get quatersawn boards. Flat sawn are just a mess when the get dried. Back in the day, they would let the log set for a couple years on the ground prior to milling, now it just get dropped , milled and stuffed in the kiln without even much AD. End checks over 12" are just considered acceptable loss. Quarter sawn boards will crook heavily, a 1x6x10 will often only yield 1x5x8 when processed and dried in that manner. The ropey curl in the ebay listing isn't uncommon . Most of what you see is quater sawn ribbon stripe. Starts out a deep red color for old growth robusta, like a jatoba. Younger trees are more pale like pink. The old growth fades to a nice brown after a while in use like on a threshold covered but gets some partial sun. So cabinets will probably change a bit over time. Good for exposed places. Can take ground contact and the fromosian termites don't put it on top of their lunch list. They will eat it if that's the only thing around. At one mill site there was a bench with robusta legs and a fir top. The mud trail from the termites went from the ground and up along the robusta leg to get to the nice soft new growth fir top. Robusta has been used a little bit for some post and beam homes here but PT fir is more widely used for that. I am building my vanity from robusta currently, I have some of that curly figure for the door panels. I like"swamp mahogany " over "robusta" for some reason. Red gum even sounds better. The latin names never held much romanticism for me. Robusta at the mill is about$2.50 green rough dimension.Dried and rough planed to 15/16"x random widths and lengths are about \$4.40 most days. It's one of those woods where it's probably cheaper to buy it at the mill instead of buying a log and processing it yourself. I think the mills get much of it for free from tree cutters, private and state. There are plenty of good places to use it on a house of other building projects.
If you need some......