Official Luthiers Forum!

Hi everyone,

I am about to make my deflection testing jig. I have a dial indicator. I plan to use that .250" with 5 lbs 4.4. oz. model. Any recommendations on designing a successful set-up? Pictures would be helpful if you have them on hand.

Thanks,
Peter

Here's one I just made. The dial indicator is on one of those magnetic stands clamped to the saw table. I made the bars 17" apart but meant to put them 18" which is what I think most everyone else is using. The bars are a 3/4" dowel cut in half and glued to the top of some 3/4" stock. The weight is temporary until I can get a 5 lb barbell but is a box of nails for a gun I don't have any more that weighs 4 lb and a 1,2,3 block that is 1 lb.

_________________
Steve Smith
"Music is what feelings sound like"

Thanks, Steve for the reply and Todd for the feedback. Todd, do you use rails thinner than 3/4"?

Check out Mr. Hurd's deflection test set-up. I've built one similar and find it very useful because you're not ltd. to a set distance.
http://www.ukuleles.com/Technology/statmeas1.html

Cal

_________________
Remember, there are no stupid questions, only stupid answers!

Now that I've actually tried to measure some of this stuff I can see that a) I need a weight in the center and, b) I need to measure from the bottom. And not too hard to set up a smaller pivot point. Easy enough to rebuild.

Two questions that comes to mind are:
1. If one is building acoustic guitars only why would you want to have slots at multiple distances? IIRC deflection varies directly as the cube of the length so it would seem you would want to take your data at some fixed length appropriate for acoustic guitar tops.
2. How do you normalize your readings for the width of the top?

_________________
Steve Smith
"Music is what feelings sound like"

About your question on adapting to different sample widths, if you use the Young's Modulus equation
in Mr. Hurd's article Cal referenced above, the width of the sample is accounted for in the equation, so
different widths are accounted for in the final Young's modulus number for the sample. Note his comments
on having the sample too "square" (graph showing the effect of Lx/Ly). By having different weight's
available, you can get enough deflection on thin or thick samples to get an accurate measurement.

_________________
Gene

Politicians and diapers must be changed often, and for the same reason- Mark Twain

So do you guys use your set-ups with 5 lbs or more or less? Do you test bracing stock as well? If so, do you cut bracing stock to 1/4" wide (finish thickness) first or test the whole piece of bracing?

Thanks for all the input!

I should have read Mr. Hurd's article as it makes it quite straightforward to calculate Young's modulus and, of course, answers the question of the multiple lengths. My next quest will be to determine how to use the data to dial in my guitars but that should probably be a different post as I don't want to hijack the OP's thread.

So Peter, it is obvious you don't want to build yours like I did; everytime I turn a corner in this obsession there is something new to learn

_________________
Steve Smith
"Music is what feelings sound like"

Don't take this the wrong way... because I have a serious question about this.

Realistically what would the margins of difference be (qualatatively speaking) for off center point loading or large radii rails? And, does it really matter? Provided that the same jig and the same testing methodology is used from sound board to sound board wouldn't the final results be similar? It seems less important to me to know what most accurate deflection numbers are as much as it is to produce soundboards of a consistant stiffness. In my non engineering brain it seems to make sense that if the testing methodology was the same every time you should get consistant results.

I am not trying to be agrumentative. I am curious.

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

Brock, you ask a really good question and the answer is important.
There are two main types of deflection testing being carried out by those on this forum

-testing a standard width top plate and gradually thinning it until you obtain a certain deflection under load that has given you good results in the past.
This type of test only requires that you use the same setup each time.

-testing a top plate at current thickness to determine its material properties (modulus of elasticity E) which then enables you to decide a target thickness for the plate.
This test gives a Value of E which should be the same whoever tests it, and whatever the test thickness is
So I test a piece of spruce at 180 thick and find it has a high E and decide to bring it down to 105.
Or test a piece of cedar at 200 find it has a low E and opt for a finished thickness of 135

This requires that support conditions and loading patterns etc be optimised so that the results can be plugged into the appropriate formula and give a meaningful result.

I do not yet do deflection testing, if I did I would use the second method but there is nothing wrong with the first once you have a few under your belt.

Yes, that is exactly what I was driving at. If all you are going for is a point of reference stiffness without worrying too much was E actually calculates out too it seems like point loading and wide rails wouldn't matter so long as your testing methodology was similar.

But I am not discounting the points you and Todd are making of having accurate data.

But that begs this question? What is the value knowing E if I can hit a stiffness I am targeting by using the first methodology? Unless I am missing the larger point all it is really doing is giving me a way to calculate how thin my final plate will be.

Again, understand these questions aren't meant to be argumentative... I really am trying to understand what I might be missing out on.

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

I'm hoping that if I can calculate E to a reasonable accuracy then I might be able to compare notes with others who are doing the same. That could provide us with some standard of comparison. Since I expect I'll always be a low-volume builder I would hope I might be able to use the information I get from others to improve my understanding of structural components of the guitar and how they work together. Maybe I can build better guitars.

I suppose someone building a lot of guitars and who was already building some great sounding ones might have a different, equally valid, approach.

_________________
Steve Smith
"Music is what feelings sound like"

A couple of reasons why you might want to use a more accurate method which can give you E.

-As Steve said, comparing with others.
-Rejecting a really floppy top before you go to the trouble of thicknessing it
-Choosing a really stiff top for that high string tension project (12 string?)
-Dealing with different size top blanks
-speed- not having to keep testing as you thickness
-ability to test brace blanks of all sizes
-Ability to test alternative materials

Makes sense.



I've got a couple 5/16" rails I place on my bench a set distance apart. I see how much weight it takes to make the top touch the bench, both along and across the grain. The distance between the rails, and the desired amount of weight is determined by size, and what I'm looking for.

_________________
I didn't mean to say it, but I meant what I said.
http://www.brackettinstruments.com/

This came up last time around: When deflections are significant, the length of the span will be increased under load. Having a large-ish radius on the rails may automatically compensate for this.

Question becomes (and I admit that this time I truly can't remember my engineering well enough) in the calculations of E, is the increased span under load significant or not?

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

Ok, that makes sense.

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

A typical deflection of a 1/4" over an 18" span would increase the plate's length a very insignificant amount. The extra work of calculating that increased length would not be worth it, IMO.

My memory of statics class says the same thing as Todd. The distance L that gets plugged into the equation is the length between the supports which is fixed.