Official Luthiers Forum!

Alex,
Moisture content can tweak your density numbers. I am sure your sets are all seasoned and stored in a tightly controlled environment and that will allow you to evaluate your stock, but you should make sure that information you are gather from others accounts for moisture content in their stock. If you look at two possible conditions, space maintained at 80deg.F and 30% equalibrium is around 6%, space maintained at 70deg.F and 55% equalibrium is going to be closer to 10%. Most of our shops are going to fall somewhere in that range, but the range of moisture(6%-10%) could modify your density by more than .65 in/g (about 40Kg/m3). Actually, that would be a pretty big consideration if you want to build a database of shared information.
For what its worth I have found overlapping density numbers amounst the Lutzi and Sitka stock I have. I don't see as clear a distinction between typical density numbers as I do with say Englemann and Lutzi, or Western Red Cedar and Sitka.

Rich

Yes!

_________________
"An adventure is only an inconvenience rightly considered. An inconvenience is an adventure wrongly considered." G. K. Chesterton.

Rather than trying to tie everybody to some standard methodology, I would think we'd get more useful data simply by having everybody find the Young's modulus along and across the grain by whatever method, and density. There are several good ways of measuring these things, and all of them have about the same uncertainty if they're done carefully, so they'd be comparable. Anybody doing deflection testing could find the E values pretty easily. The density would require a good scale of some sort. It would be _really_ nice if we could circulate a set of samples that everybody could measure, so we'd have some idea of the range of errors we's be likely to see from different 'labs', but that might not be possible.

Alan

How do you measure Young's Modulus?

Todd,
What are the unit of measure for this formula? I've seen them written as Kilograms and Meters. Is that what used here?

It's looking like we're onto something here! Keep the ideas and opinions coming!

_________________
Jim Howell
Charlotte, NC

If we would put both E and thickness in a xy chart, the graph might probably differ a lot from top to top?

If the above has some truth to it, and if we are going to put together some info, perhaps group the data by species, and into several thickness intervals?

_________________
Build log

I actually typed the E and density chart var too then edited and deleted it.

My point (blured inside my head) is that i fear wood's natural ability to defy logic and math trends. For example is it not possible to have two raw thick (the usual ~5mm) plates that have the same E, but when thinned down to near bracing thickness they now have a very different E? If that is so, then we need to measure and compare E when the wood is much closer to the final thickness, say not more than 3.5mm.

How far off track am I now?

_________________
Build log

The modulus of elasticity should, and I stress should, be the same throughout one piece of wood. The modulus is a material property and therefore solely dependent on the characteristics of the material being used. Size and shape will have little to no affect on E. Now assuming that E is constant throughout the material is also assuming a homogeneous material, which wood is not. The changes in grain pattern and defects in the wood will alter material properties slightly.

So if you wanted to obtain a value of E(longitudinal) that was the closest representation of your top then yes I believe you should have it at final thickness for calculations. However, the value should not change much if calculated before the top is to thickness.

Also when the top is vibrated it will experience stress in more than the longitudinal direction. The moduli in the tangential and radial directions will come into play. those may not be calculated through deflection testing though. I believe that many mechanical properties of wood books have ratios relating E(radial) and E(tangential) to E(longitudinal) so if anyone was feeling overzealous they could calculate all three based on their deflection results.

_________________
--Kyle Burner--

Good idea... so a useful data set might include:

species
moisture content
density -- using corrected scales
specific gravity
calculated E
calculated Q

I'm also wondering out loud if we could use something like the Lyman shooter's weight check set (under $25) for scale calibration. This would keep people from having to pass calibration weights around.

_________________
Jim Howell
Charlotte, NC

It is interesting that (among the few) good classical guitars i inspected so far, the stiffness at the bridge was all over the place, from near rubbery to hardly moving. I guess it would have helped to know the weights of the top and of the bridge before drawing conclusions from the press-the-bridge test.

_________________
Build log

Agreed, a simple and easy method for people to have and use makes good sense, and may actually make the information more usable. If your test rig becomes so complicated and expensive that very few could have it in the shop and use it to test tops, the data would be nifty for debates, but not of much value to a person who can't run a comperable test on the top in front of them.

Maybe a simple sketch of the test rig (with significant dimensions) would be good, and spec. out other significant features. Then we could start testing and passing around a few sets. I would be willing to pass a few of my sets around.

Rich

Say a person is will to put $100 or so into materials and a weekend building a test rig. Is it possible to build something suitable inside these parameters, even for point load testing? In a past life I have spent many, many hours (years) doing materials testing on the civil engineering side and the 'trick', if I may use that word was to be consistent in technique. Most of the apparatii were pretty simple.

_________________
Jim Howell
Charlotte, NC

Jim Howell wrote:
"In a past life I have spent many, many hours (years) doing materials testing on the civil engineering side and the 'trick', if I may use that word was to be consistent in technique. Most of the apparatii were pretty simple."

Any test you do HAS to be simple or you won't do it.

There is no such thing as a 'perfect' test, but reasonably accurate ones are quite feasable for a very low outlay in setup. Deflection testing requires a dial gauge, which you can get for about $25, a weight, and a simple rig that you can build in a couple of hours. If you want to find modulus information you'll need some way of measuring the length, width and thickness of the piece with reasonable accuracy. These tests are best done on top half blanks that are rectangular: the math is easier.

I use my signal generator to find the frequencies of the lowest bending modes along and across the grain, and a triple beam balance to find the mass, so that I can calculate the moduli and Q values that way. It makes sense for me since I already had the stuff. I just do a whole batch of tops or whatever when I get them.

Finding the moduli is the way to go: they are thoeretically independant of the sizes of the pieces you use, and the units don't matter, since we can convert them easily enough from inch-pound-second to meter-kilogram-second, or whatever.

What you do with the information is up to you. I use it to try to keep the weight down on my tops: figuring out how thin I can get away with on _this_ piece. I just use the lengthwise E value for that: it simplifies things and can be justified logically. The ratio of lengthwise to crosswise E value says something about how easily you can get 'closed' Chladni patterns, as Mark Blanchard pointed out. He uses mode shapes on shaped top plates, but I'm starting to relate that to moduli.

Q values are certainly a good thing to look at. There are ways to do that with a computer and sound card, BUT there are some pitfalls. There are always pitfalls. I've begun to wonder about whether a combined shear modulus might tell us something. Remember that the crosswise E value depends very sensitively on the grain angle, since it relies so much on the medullary rays. Crosswise values of top samples I have are all over theplace, with no relationship whatever to density or lengthwise E.

Like Jim, I spent years doing civil engineering materials testing.

A simple jig is fine but the devil is in the details.

I would suggest the following-
-Knife edge supports, one side fixed, with one side free to rock inwards
-Mount dial guage underneath so that it does not prevent you from placing the load at the true centre of the span.
-Load must be in contact all the way across the board.
-Preload lightly to seat the board before test load

I would suggest an easy way to do the preload/load would be to use a steel pipe about 1 1/2" dia as long as the maximum board width, cut in half round side down.
This is the preload, place across the centre of the span and zero the dial guage.
Then place a solid steel bar of known mass inside the half pipe and read the guage.
Use a spreadsheet to input variables such as span, thickness and width to calculate E

Knowing E for a particular piece of topwood will guide you in deciding to go to 125 thou or 100 thou.

You could also use it as Brock and others have been doing to obtain consistent deflection readings as you thin the board, but the results are not meaningful to share unless everyone has EXACTLY the same rig and method of testing.

When using a small concentrated weight, the result will depend a good deal on the crossgrain stiffness as well, not just the along the grain one, right?

_________________
Build log

Yes, concentrated loads rely on the stiffness of the material to distribute them.
For the purposed of testing , you want to consider the plate as a wide beam or one way slab
and support and load it appropriately to achieve this
The simple beam formulae rely on having a uniform load per unit width applied at the midpoint of the span all the way across the width of the slab.

If desired you could also test crossgrain stiffness by changing the orientation of the test piece ( probably reducing the span to 1/2 as well)

Jeff Highland wrote:
"You could also use it as Brock and others have been doing to obtain consistent deflection readings as you thin the board, but the results are not meaningful to share unless everyone has EXACTLY the same rig and method of testing."

And, I would add, builds exactly the same guitar too! There are lots of good ways to build guitars, and what works for my design and desired tone might not work for you. In the beginning this sort of thing will help you make small departures from your norm and understand why they end up the way they do. Once you get calibrated your work is likely to be more consistent. 'Better' is a big word in this context.

Alan you make some good points about us all needing to use the same apparatus and method to have the data be really useful. I don't have a deflection testing setup yet, so I'm game to build whatever is a reasonable for all of us.

Jeff also seemed to have some good points about the actual rig and method. It looks like his points could be worked in to the rigs that people like Brock have with only minor tweaking. Attaching the dial from the bottom, installing one knife edge and so on.

Consider this... If we could agree on a standard rig, I would be willing to make them up for everyone, thereby eliminating any differences in testing equipment. I could source out all the parts and figure out a cost per unit that everyone could live with. I would also be willing to donate a few sets of tops that could be circulated around and tested by everyone under equivalent atmospheric conditions (temp and humidity) to see what the error between "Labs" would be.

Any thoughts?

Greg

That has been EXACTLY my point about sharing deflection data etc. It is meaningful to me, but taken in the context of another builders style or methodology it may have wholly different results.

_________________
Brock Poling
Columbus, Ohio
http://www.polingguitars.com

Greg E wrote:
"Alan you make some good points about us all needing to use the same apparatus and method to have the data be really useful. I don't have a deflection testing setup yet, so I'm game to build whatever is a reasonable for all of us."

Actually, I see no ned for us all to do the tests in the same way, just so we get comperable results from our different setups and, more importantly, express them in ways everybody can use. There are several methods for finding the Young's modulus along and across the grain. Each of them has it's own drawbacks in terms of accuracy, cost and time. It would be nice if we could get everybody to 'calibrate' their own setup, using some 'standard' pieces of wood or something, but realistically that's not likely to happen. This will necessarily limit the usefulness of the comparative data. However, I don't see this as being a big problem for us. None of us is working so close to the line, as it were, that a small error will make a large difference. It would be one thing is we were building these things to go into orbit, say, where a few grams one way or the other would make the difference between success and failure. With a bunch of builders making non-standardised items, and judging them subjectively for tone for the most part, we're going to get more from watching trends than exact numbers. Knowing that, say. less dense tops work 'better' for classicals, and more dense ones for Bluegrass Dreads, will help us all to sort our tops for best utility. Some comparitive data on stiffness ratios will help us to figure out just how floppy a top you can get away with and still make a good Jumbo.

Just wanted to know if anyone has gotten any further on this matter ?

Lars.

Lars:

I'm still interested in pursuing this topic, but [as most small time builders] lately my personal life has taken priority over my building projects.

Perhaps this thread can be revisited by others with more time and resources, so an actual sharing of data can occur.

Thanks... Greg

As soundboard shape affects the results is there a standard size they should be trimmed to, to achieve consistent results?

_________________
Formerly JJH

I learn more from my mistakes than my successes