Official Luthiers Forum!

I just got a Red Spruce top set off of Ebay.

The growth rings are over 20 degrees off of vertical - it's not as obvious in the photo but I measured with a machinist protractor.

The set feels pretty stiff. If the it was rectangular I'd test it with the Gore methods to get some objective data.

What are your opinions & experiences with this? Doesn't Spruce loose a lot of it's stiffness as the grain goes off vertical?

Thanks,
Kevin Looker

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

It does lose stiffness yes. But that doesn't mean it can't make a good guitar. Hopefully you got a good price for it at least. I personally would not use it unless perhaps it was super stiff for some reason. I do deflectin testing so I would test it with that method. I am restoring a '40's Martin 000 right now and through the soundhole it looks like it's about the same off quarter as yours though it does show some medulary rays on the outsides of the lower bout so it probably goes to quarter there. The guitar has a great tone.

Deflection testing, , why didnt' I think of that.

I have never done it before but I have enough other sets to use as references that I've tested using the Gore methods.

Thanks

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

I'm of the opinion (along with Serge DeJonge, I'm told) that cross grain stiffness is over rated. You want a certain degree of cross stiffness, of course, but it's lenthwise stiffness that keeps the top from folding up. I have found that I get better results by reserving the stuff with the higher cross stiffness for the wider guitars; Jumbos and Dreads. As you go toward a narrower aspect ratio you don't need as much. That top would probably work just fine for a size 1, 0, or 00.

When I was over at his shop talking to Sergei about tops, he told me a few surprising things about
his view on tops. Outside of doing some deflection testing, he'd like to see the word stiffness removed
from the discussion altogether. He also mentioned he like tops that are off quarter by about ten degrees because
they don't crack as easily. Can't really argue with a man of his reputation. One of the nicest people you'd
ever want to meet too.

I didn't know that about not cracking as easily, but that could explain why the instruments I've built with 45 degree wood (maximum cross grain floppiness; it stiffens back up as you approach perfectly flatsawn) haven't cracked, despite the fact that they should have higher humidity movement than perfectly quartered wood. They also sound great, so I would have no qualms about using the wood in the original post, at least on narrower instruments as Alan says. I haven't built any large guitars with off-quarter wood yet, so I can't say how well that works.

If the wood was advertised as quartersawn, then I'd ask for a discount just to keep the seller honest.

I also agree on the point about stiffness being a confusing word. People often say "this is a really stiff top!" as if that somehow makes it better, but that statement has basically no information in it. Any wood will be stiff if it's thick. And are they talking about long grain or cross grain stiffness, anyway? What I consider to be the primary measure of a top's quality is the ratio of long grain Young's modulus to density. Tap tone/damping is another factor which is important for nylon strings, but steels seem to sound good and perhaps even better with thunky wood.

The top on my 1937 D-28 has 45 degree grain in the center. It does get more vertical on the outside edge, but it is nowhere perfectly quartered (no straight-across silking visible). There are no top cracks.
It is one of the best D-28's I have ever played. and the top is close to 0.130" thick in places.
One of the best dreads I have built also had off-quarter grain in the red spruce top....almost identical to the OP's top.

_________________
John

Deflection testing only tests the long grain stiffness but I think stiffness is an important word. The way I think of it, the stiffer the top is the thinner it can be made which should also reduce the weight considerably. So if you start off with two tops say 3/16th inch thick and one of them feels floppy and the other one is stiff in the long grain then when you deflect each one to a target deflection the stiffer one will be thinner and lighter.

Off quarter, thick, sounds great?

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

You can test cross grain stiffness using deflection. I measure the density and use vibration testing to find both the long and cross grain Young's moduli. I basically use the density and long grain E values to decide what sort of guitar to use the top on, a,d how thick to make it, and the cross grain E to figure out whether to make a wide one or a narrow one. That is, Classical guitars get the tops that have lower density, or an unusually high long grain E value that will result in a light weight top. The really dense tops go into steel strings for flat pickers, who want 'headroom'. Jumbos and Dreads get the tops that have the lowest stiffness ratio: the cross grain E value is high relative to the long grain.

Just for reference: perfectly quartered softwoods can have a long/cross stiffness ratio as low as 8:1 or so. This drops off fast as the ring angle relative to the surface goes above 5 degrees or so. At 45 degrees the long to cross ratio can be 100:1. It will get back up to as much as 12:1 on perfectly flat cut wood. If you look at the shape of the cells on end grain, softwoods tend to have square or rectangular cells: they look like a stack of boxes. Most of the resistance to bending comes from tension and compression in the cells at the surface. When the sides of the boxes are all along the surface, is it is in quartered or flat cut softwood, then you have to stretch or compress the material in the wall of the box to get it to bend. When you go to a 45 degree skew cut you only need to deform the squares into diamond shapes, which is a lot easier. The somewhat higher cross stiffness of quartered wood compared with flat cut is probably due to the medullary rays, which are small bundles of cells that run in and out along the radius of the tree, so that have fibers that run across the grain and help tie thing together.

Skew cut wood has the highest resistance to cracking. I cut it that way for harp soundboards, which don't need cross grain stiffness and sure benefit from the extra splitting resistance.

John Arnold wrote:
"The technical name for the 'hard lines' is compression wood, which occurs under large limbs and on the downward side of leaning trees. Compression grain is more common in younger trees, which is why it is seen in a lot of red spruce. The red spruce trees being cut today are younger, since many of the older trees were logged long ago.
All spruces can exhibit compression wood, but it is often discarded as being low value. It can make outstanding instruments....if it is handled correctly.
Compression wood is denser and stiffer than normal wood, and can be worked thinner without worrying about structural issues.
Steve Gilchrist only buys compression red spruce when he buys from me or the Hampton brothers. He contends that it is more consistent than the more expensive 'white wood'."
- and-
"The top on my 1937 D-28 has 45 degree grain in the center. It does get more vertical on the outside edge, but it is nowhere perfectly quartered (no straight-across silking visible). There are no top cracks.
It is one of the best D-28's I have ever played. and the top is close to 0.130" thick in places.
One of the best dreads I have built also had off-quarter grain in the red spruce top....almost identical to the OP's top."

I value John Arnold's experience and common sense.
Although you can't make a silk purse out of a sow's ear, at least it's good to know you can make a fine guitar out of some stripy , wide grained , thicker and somewhat off quarter wood. Myths and Holy Grails are dying every day.

While we are discussing great guitars that on first sight would be suspect, if it is not my subconscious playing tricks on me I think I read that Martin joined many tops of this vintage with the bark side out. If I were to ask anyone if it was true it would be John. If it were, it would put the wider grain in the center and the tighter grain (assuming the grain pattern is not even across the board) at the outside. I wonder if there might be some relationship between the sound of these older guitars to the way they were joined?

I sometimes join tops with the wider grain in the center. It can sometimes allow you to cut out flaws that would make a top unuseable, and sometimes it just looks better. The center section of the top has a large brace in the form of the bridge (plus the X crossing) and half of it is cut away by the soundhole or constrained by the fingerboard extension, so I don't worry about stiffness issues.

If you test a bunch of softwood samples for Young's modulus along the grain, and plot out the results against density, you'll find that most samples will fall close to the same line, regardless of species. The major exception to this is wood that shows heavy latewood lines, which will tend to denser than 'normal' wood, but not commensurately higher in Young's modulus. Prominant latewood (what John calls 'compression wood') tends to add more density than stiffness.

I think there's a nomenclature problem here, so I'm going to tread carefully. All trees form 'reaction wood' in response to bending stress, whether it's from the tree trunk leaning, or in a branch, or from an off-center load, such as having more branches on the south side. In softwoods reaction wood is formed on the compression side. The latewood lines will be very wide, and include interlocked grain, with the cells running at least somewhat diagonally. I interpret this as an effort by the tree to butress itself against shearing stresses in the wood which would result in 'cold flow' due to the thermoplastic nature of lignin, but that's just my model.

I've got a really good example of this in a European spruce plank that came from an old barn in Austria, or so I'm told. The plank came right from the center of the tree, and has clear reaction wood on one side, with wide grain spacing and very heavy latewood. Since there is no 'sweep' to the wood (the trunk was straight in that section) I suspect this was the south side of the tree, and had a lot more branches. The wood on the other ('north') side shows a more normal grain structure, but still has heavy, although much narrower, latewood lines, and tighter spacing. A sample cut from the tight grained side is the almost the densest piece of softwood I've seen, at a specific gravity of .54, and has a lengthwise Young's modulus of 19,000 megaPascals. When plotted out on the E vs density chart it's right 'on the line'; the density and stiffness relate in the 'normal' way. The sample from what I'm taking to be the south side of the tree, with the reaction wood, is actually somewhat less dense, at .51 sp.g., but the E value along the grain is only 9400 mPa. This matches the E value of an 'average' piece of spruce with a sp.g. of .38. If you were making tops from these two pieces of wood the one from the south side, with all the reaction wood, would end up being more or less normal in thickness, but quite heavy. The 'north' side top would be thin, and also somewhat heavier than a top made of more 'normal' Euro spruce. IME, heavy tops tend to favor bass response and 'headroom'. (In the event, it was not possible to make any tops from this plank, alas.)

I'll note that 'reaction wood' normally has high built-in stress, which causes it to curve (usually away from the center of the tree) when it is re-sawn. Despite having been cut something like 300 years ago, the sample from the 'south' side of that plank has warped slightly since I cut it a couple of years ago, indicating that there is still stress in the wood. The more 'normal' sample from the other side is still straight.

Here's where the nomenclature issue comes in. We often see softwood, particularly from low down in large trees, that has heavy latewood lines, but is not 'reaction wood' in the usual sense: it doesn't have obvious interlocked grain, a lot of built-in stress, and the sort of 'fuzzy' look of true reaction wood. Archer, in his book 'Growth Stresses and Strains in Trees', says that the combination of weight and built-in stress can actually exceed the compression strength of the wood in the center of large trees at the base, leading to things like 'brash fracture' due to microscopic compression failures. It's at least plausible that very large trees could react to high loading near the base by reinforcing the structure, say, by thickening up the cell walls in latewood, to avoid compression failures. Doug fir and Redwood both commonly have what John Arnold calls 'compression wood'. This is not 'reaction wood', which is also commonly called 'compression grain' since it occurs on the compression side of the tree.

So you see there seems to be a nomenclature issue here. I hope we can resolve it: it seems that a 'rectification of names' is in order.

"So you see there seems to be a nomenclature issue here. I hope we can resolve it: it seems that a 'rectification of names' is in order."

Good luck with that! Like so many lumber related terms, it is somewhat regionally and "field of endeavor" dependent. I've heard the term "rift sawn" to mean radially sawn with grain at 90 degrees to the face of the board (vertical grain) and at other times used to mean boards with grain closer to 45 degrees to the face of the board.
You are right - it would be nice to agree on the meaning of terms, so as to avoid confusing reaction wood from wood that may be more dense but otherwise fine to build with.

Certainly the sample needs to be rectangular to tap test accurately, but as long as you can find some part of the panel that will give you a rectangular section, deflection testing will get you Elong and Ecross. The equation to use is Equ. 4.4-6. For example, suppose the panel has a corner missing but is otherwise rectangular, do a three point bend test but make sure the supports are on the rectangular part of the panel (just let the bit with the missing corner overhang the supports), then load in the center of the span. Works for both long and cross grain measurements.

The pictures below show my traveling 3 point bending rig, set up for firstly long-grain stiffness measurement and then cross-grain stiffness measurement. I only have pics showing fully rectangular panels, though.

Long- grain stiffness measurement:


Cross-grain stiffness measurement:


It's worth looking at the sensitivities to variations in Elong, Ecross and G. This is done in detail on page 4.64 (Design) and shows that 50% variations in Ecross and G give very small variations in the modal frequency of the plate. This leads to two immediate conclusions: 1) Things aren't very sensitive to cross grain stiffness (as many above have pointed out) 2) Whilst G is not measured by this deflection method, a typical value (e.g. 1.0 GPa) can be used in the standard equations with little error, as the modal frequencies are rather insensitive to G. (A 50% change in G gives only a 1.35% change in modal frequency).

So what is the "best" top when making comparisons across a number of tops? For best responsiveness, it's the lowest mass top (sample density x target thickness x soundboard area) that gives you the modal frequencies you are after (governed by the parameter "f"). As Alan and others have pointed out, if you want "headroom" rather than responsiveness, choose a higher mass sample.

So stiffness (in engineering terms) is important, because it allows you to compute the top panel mass and so judge responsiveness.

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

Thanks Trevor.

Keep in mind as well that the properties of the wood will vary from point to point. This is often particularly the case with the cross grain stiffness of a top that is off quarter: usually one edge is noticeably better quartered than the other. If you flex a short section of it across the grain the well quartered side might be quite stiff, while the off quarter edge can be pretty floppy. In a test of the whole top you're getting some sort of average value, and that's useful, but two tops with the same average measurements can make very different sounding guitars. In many respects these measurements are best thought of as ways to stay out of trouble: they might not help you make 'great' gutars, but they can keep you from making bad ones.

That section deals with E for a beam or brace. I don't see why it would make any difference but to be clear that will work for a panel as well right?

That's good to know because most of the tops I have in stock I'd say are cut around knots and as such are not perfect panels. I guess it doesn't matter as long as you measure the distance between supports. Makes sense.

Put it this way, it's the same basic equation that everybody else uses for panels when measuring E by deflection (static) methods. However, it does contain a few assumptions buried in it, the major one being that long-grain stresses don't cause cross-grain strains. But they do, and they're related by Poission's ratio. For low aspect ratio prismatic sections like beams and braces the effects are generally regarded as negligible, but not quite so negligible for thin panels. The difference is quite small and I've had quite good agreement between measurements taken using deflection methods and those using tap methods. The tap methods take into account the Poisson's ratio effects and using tap (dynamic) methods all the way through the process (measuring a frequency to get E then getting back to a frequency via the target thickness) keeps systematic errors to a minimum.

So the errors introduced are small, and the answers you get are very usable and certainly better than no answer at all.

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

I've done quite a few mandolins with quarter approaching that with the grain running \\\\////. They have come out just fine, but a mandolin has curved structure, so tell me where it is quartered...
I have never used that kind of "quarter" on a guitar, but it would probably be fine. I always approached instruments from the "thick side" and kept the bracing about the same on all, relying on sanding and tapering the lower bout after the box was assembled.
I guess if it was for a customer, I would not use it...

It's normal on carved tops, such as violins, to start with a wedge of wood split from the trunk, which is split in half lengthwise along the radius of the tree. These are book matched, yielding quartered faces on the flat 'down' side of the piece . On a 'perfect' piece of wood the grain curves upward and outward on either side, and the upper faces of the spit also show perfect quarter. As you carve the arch the grain never gets more than a few degrees off quarter anywhere in the top, preserving the stability and high cross grain stiffness. You don't always get that, but it's worth some effort.

The arch at the ff hole notches on an F5 mandolin is 11 degrees off quarter, if you need perfect numbers...