Runout, revisited

[charangohabsburg]

Some of you may remember an old thread called "How to determine runout in spruce top billets" where a few years ago I posted pictures and explanations hoping to shed some light on the theoretical aspect of runout.

Last December I had to repair a nasty crack in a charango top that already had been repaired previously by a German guitar maker. He had done a very neat job, but one cleat for a crack that runs from the bottom of the top all the way up to the lower harmonic bar is not enough, not even on a small instrument like this (total string tension of a charango is about the same as the one of a classical guitar). Juan Achá, the maker of this little treasure is one of the most renowned charango makers of Bolivia, and all the tops of his charangos I have measured so far (maybe ten or so) are somewhere between 0.8 and 1.4 mm thin. I won't go into the details of the repair, but the following picture of my completed repair (the two transparent cleats) should illustrate why I wanted to use the lightest and strongest spruce cleats ever found on planet earth:

2014_P7700_0846-1.JPG (39.23 KiB) Viewed 13634 times

The dark cleats (mahogany I guess) are the ones used by the previous repairer and are about two or three times thicker than the ones I was using.

The crack holds up well and the owner says its sound and playability would be just the same as before.
I believe that not runout-free cleats of the thickness I used would not hold up. Not if that close to the bridge.

I had split completely runout-free cleats out of a discarded (free) spruce top with horrible runout. The cleats are about 25mm long and 0.4 mm thin. The bigger one brought 0.05 grams to the scale (one half of 1/10 gram), which I thought was acceptable.

A few days ago I got curious about the difference of behavior between the same spruce of a halfway comparable thickness (this time 0.95 mm), with and without runout. I cut and planed two longish cleats from the same horrible top (runout ratio 1:15 which equals about 3.8°), one just as the top is, with runout (B) ...



...and one from a carefully split piece of that top, ending up with virtually zero runout (A).

Then I proceeded to destroy them . I was filming the destruction (at 30 FPS) in order to get a picture at which amount of bending the cleats let go. First the runout-free sample (A), then the 1:15 runout sample (B) one frame after they let go:



And here the comparison just before they let go:



Between B1 and B2 some wood fibers started to crack audibly, B2 is the last frame before the sample cracked completely and flew across the bench.

Now you may imagine what kind of wood Vicente Arias was using when he built this guitar. Just saying.
If interested, here you can download an article about this Vicente Arias guitar and its history (nothing about runout in spruce tops).

Back to runout: some close-ups of the broken samples and fibers, again (A) = no runout, (B) = runout ration 1:15 (click to enlarge).

[Trevor Gore]

Nice work, Marcus.

So, the bending strain looks at least twice as much for sample A; therefore the stress at failure for A would be at least twice that of B? (Implying that in a pure bending test, A would take twice the bending moment of B?)

Impressive demo!

[charangohabsburg]

At first sight your assumptions may look reasonable Trevor, but I wouldn't go so far to draw any conclusions regarding the bending moment by only looking at this simple experiment and without measuring applied forces. The only thing I can say is that the runout-free sample is much more resilient, thus less prone to cracking and plastic deformation.

Let's have another close look at the breakage. In the pictures below it becomes evident that the breakage of sample A occurred in a different way than the one of sample B. While in the runout-free sample A apparently all wood fibers let go much in the manner an stretched rope lets go when stretched too much, in sample B all wood fibers (except those that still hold the two pieces together) separated, without actually being cracked themselves.





Also, considering that wood, when bent will deform plastically before breaking, and observing that sample A stays slightly bent over its whole length while sample B only stays bent within the few millimeter of the skewed cracking zone I venture to say that the stress distribution only works well along the wood fibers, and much less so between fibers. This may seem trivial because we know that a sliver of end grain wood almost can not be bent at all without breaking it, but it still may be a surprise that a runout ratio that can be found in lower, mid, and even some higher graded spruce tops make such a great difference.

With this I don't want to say that it would not be possible to make a decent, good or maybe even very good guitar with a top that has a lot of runout, but I dare to say that it will be impossible to get somewhere like close to replicate the sound of a guitar with a well performing very thin top, if trying to use spruce with heaps of runout.

[Mark McLean]

Nice work Markus. I am convinced. You have shown that the difference is between one piece that breaks when its fibres snap, and another which breaks when the fibres separate. Makes sense. I am going to spend the weekend re-examining my wood stack for runout.
cheers
Mark

[Kamusur]

Sure do remember your 'tute' Markus, you da man. That thread of yours plus one that Kim also wrote about runout amount to my total insight on this phenomena. Thanks again it shows very simply and graphically the effects.

Steve

[kiwigeo]

All very interesting but here in Australia there's only type of runout that really matters and its got nothing to do with wood.....

Sachin-Tendulkar-of-India-is-run-out-by-David-Warner-of-Australia-during-the-One-Day-International-match-between-Australia-and-India-at-Sydney-ground-420x288.jpg (26 KiB) Viewed 13451 times

[charangohabsburg]

That's nothing Martin, here in Switzerland we don't even have a word for "runout" (except the English word "runout", of course).

[Kamusur]

[quote="kiwigeo"]All very interesting but here in Australia there's only type of runout that really matters and its got nothing to do with wood.....

Yes Martin the worst sort of runout for Sacchin, and buy the way have you ever seen any photo's of his house? It'd be one i would runout of as well.

Steve

[Graham Long]

Hi All,
Just wondering if the run out in spruce tops can be correlated back to the tap (physical properties) testing as detailed in Trevor's books.
Ie Is there a substantial change is the resonant frequencies from a bad run out sample to a perfect one? and if so, would the top target thickness calculation offer compensation for the reduced strength of the top?
Cheers
Graham

[Trevor Gore]

Run out may increase the stiffness of the wood (unlikely) or decrease the stiffness of the wood (more likely). Either way, tap testing the wood will measure its stiffness (Young's Modulus) and then the thicknessing equation will tell you how thick to leave it. So, indeed, the process self compensates for run out, if run out has an effect. However, it would be hard to correlate changes in Young's modulus to just run out alone, as there are other possible causes for the changes.

Small levels of long grain run out generally do not change Elong very much, whereas small levels of cross grain run out can reduce Ecross significantly, especially, for example, in WRC. According to Schleske (1990), one significant downside of long grain run out was increased damping.

[slowlearner]

Ummm, hate to be the village idiot but here goes...



What's "run-out"?

[charangohabsburg]

What's "run-out"?

The wood fibers that "run out" of the wood surface.

[DarwinStrings]

The wood fibers that "run out" of the wood surface.

Why are they in such a hurry?

I have done this test quite a bit with my wooden go bars. I now find that with a humidity cabinet which runs at temperature of about 41C that I am getting more creep with my wood go bars so might be time for some fancy fiberglass ones.

Jim

[charangohabsburg]

Small levels of long grain run out generally do not change Elong very much, whereas small levels of cross grain run out can reduce Ecross significantly, especially, for example, in WRC. According to Schleske (1990), one significant downside of long grain run out was increased damping.

Thanks. Now I hunted down that article. Speed of sound and damping of spruce in relation of grains and rays, Martin Schleske, Catgut Acoustic Society Journal Vol. 1, No.6 (Series II) November 1990, pp. 16 - 20

Citing the paragraph where some numbers on runout are mentioned:

Even minimal deviations from 0° result in a decrease of the speed of sound and in an increase of the loss factor: At
the strip, for which [runout is] 5°, the speed of sound decreases by 7% while the loss factor increases by 19%. At 10°, the speed of sound decreases by 17% while the loss factor rises by 51%. At a deviation of 20°, the speed of sound decreases by 38%, while the loss factor increases by 145% compared with the longitudinal strip which runs exactly in the split direction. If we express this as a change of the Young's modulus, which results from speed of sound squared multiplied by the density (not taking into account the Poisson interaction) the Young's modulus has decreased by 62%.

A runout angle of 5° is what I call "runout ratio 1:11.43", 10° would be 1:5.67. Needless to say that most probably (hopefully!) nobody offers guitar tops with 10° of runout, but a runout ratio of about 1:12 (= 4.76°) is more common in lower, medium, and sometimes even higher graded spruce tops than one might believe.

[Trevor Gore]

For those with "the book", I converted Schleske's speed of sound numbers to Young's modulus numbers (SQRT relationship), Design, p 4-10. For a 1:20 runout ratio, (2.86 degrees) which I like to keep below, converting Schleske's data gives 12.6 GPa for zero runout compared with 11.8 GPa for 3 degree runout, which is a 6.5% reduction in Young's modulus. For comparative purposes, for a guitar top about 2.8mm thick, you'll get a 6.5% reduction in stiffness by taking off another 0.06mm.

As Markus points out, as runout rises, things start going bad pretty quickly.

These sorts of numbers help explain why, when working with guitars, the differences between good and really good are so small when working to linear dimensions.

[Mark McLean]

Pete
Have a look at this Frank Ford pictorial explanation:
http://www.frets.com/FretsPages/General ... unout.html