Official Luthiers Forum!

Hi all!

I am curious as to the various set-ups people are using for deflection testing. So, you data junkies (I'm guilty!) what thickness do you deflect at, and how accurate are you at maintaing that thickness from top to top (like +/- .002" or what?)? When do you deflect, prior to or after gluing the pieces together? How much weight are you using and what does your frame look like (spacing between cross supports in either direction)? What numbers are you shooting for and why? What have you based those on?

I am going to set up to do this and would like somethng that is at least in the range of what others are using so the numbers have some meaning. I do on occasion measure for desinty for some of my customers already but would like to include this in some cases and will also use it for my own builds. Al C, what are you using to measure MOE?

Thanks all!

Shane

_________________
Canada

I'd like to hear this as well, unless I get set up before someone suggests a good standard. The 100 Lutz tops to be joined here soon should give a good baseline...

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

I could make a good suggestion, but I'm certainly not an authority on top deflection. I agree that some sort of "standard" should be in place in the industry.

_________________
Reguards,

Ken H

Hi Shane,
When I do my deflection testing I do it prior to joining the tops, but they are thickness sanded, usually around .160-ish. Measure the top with a micrometer in many places and average the results. Try to get the top as even as possible of course and the averaging will help. .002 is pretty typical of what I see off the thickness sander.

I do simply supported, between two centers. I use metal tubing for the centers as it's pretty straight, much straighter than wooded dowels.

For weights I use some cheapo exercise weights that i've weighed on a gram scale. You don't want a whole lot of deflection just enough to get a repeatable, reading from. I usually shoot for .04 inches, but the readings will vary of course by about +/- .010.

I test both axes using similar setups, adjusting the supports for the dimensional changes. Make sure to load the top across the entire width, not just in the center.

Once you have this data and calculate the moment of inertia (b*h^3)/12 you can calculate the modulus of elasticity. This is the number you're after unless you just want to make all of your plates the same dimensions and just leave it at "this plate is stiffer than that plate type of thing.

I also think it's a good idea to keep a standard and check it at the beginning of your testing session to verify everything, sort of a calibration check if you will.

Hope this helps.

_________________
Jim Watts
http://jameswattsguitars.com

Thanks Jim.

How much weight are you using and what spacing are your metal supports? I know that you test your brace stock as well. How are you doing this differntly.

Thanks again!

Shane

_________________
Canada

Interesting i would like to know more on this subject as i would like to offer
more accurate info on the top woods i offer.

joel.

Shane, i'll pull out my fixtures and take a pic or two. More info coming a little later.

_________________
Jim Watts
http://jameswattsguitars.com

I deflect AFTER joining the top. I only deflect longitudinally. So the width of
the board has to be consistent. I make mine 16 inches wide regardless of
what the final width will be. I use a 5# weight and the supports are around
18 inches apart (I think, I will have to measure)

I have a basic target deflection for each body size. My feeling is smaller
bodied guitars need a greater deflection than a larger bodied guitars.

I will try to get some photos.

_________________
Andy Z.
http://www.lazydogguitars.com

Here are the only picts I could find. It would be better if my weight was a bar to evenly distribute the weight, but I have been using this so far and my results are fine.

Numbers for the deflection are kind of meaningless taken out of context of the entire system and bracing etc. For what it is worth, here is some data on my L-OO

Lutz L-OO Lutz/AAA 64     0.153     ; 360     6.68
Lutz L-OO Lutz/AAA 75     0.143     ; 348     6.91
Lutz L-OO Lutz/AAA 120     0.127   &nbs p; 303     6.77
Lutz L-OO Lutz/AAA 169     0.11     ; 268     6.92
Lutz L-OO Lutz/AAA 241     0.096   &nbs p; 234     6.92
Lutz L-OO Lutz/AAA 275     0.091   &nbs p; 223     6.96
Lutz L-OO Lutz/AAA 290     0.087   &nbs p; 215     7.02
Lutz L-OO Lutz/AAA 330     0.083   &nbs p; 205     7.01

The last 2 columns are weight and density measurements

_________________
Andy Z.
http://www.lazydogguitars.com

Wouldn't it be nice if we could all agree upon a single standard for measuring deflection?   For instance, start with a given thickness and then have a standard fixture and weight and measure for both longitudinal and lateral deflection...   How hard would that be?   It would give us a common language here.

Erik Schmidt, a newbie builder building his first classical guitar has a "build thread" (for those that hate that phrase, please substitute "online documentation of instrument fabrication") on the Luthierforum that includes a simple but effective deflection testing setup. I believe that Udi (a member here and there) may have provided the needed info to build the rig.

This link should save you some time, as it should take you right to the discussion of deflection testing, in the middle of the thread with the most hits I have ever seen on any forum. (While you're there in that thread, check out this guy's incredible photographic documentation, and feast your eyes on his rosette.)

Dennis

_________________
Dennis Leahy
Duluth, MN, USA
7th Sense Multimedia

Thanks Jim looking forward to them.

Andy, is that last number the deflection you are shooting for? 330? For a L-00 sized guitar based on your bracing pattern? It also appears that your density changes, that does not make sense but you are also dealing in pretty small numbers so sanding dust on the piece could probably alter this as the density should not change on these pieces.

Rick, any suggestions for a standard? I am indeed open for that as are others here. We could pick on of these as well and then the others who are doing this can calibrate and switch to the standardized system for comparison purposes. Although, at the end of the day some builders will seek different deflection values based on their bracing patterns and wanted tone/volume I am sure.

Do any of you redeflect as you brace to end up with a 'system' deflection number?(the top all braced up!)

Thanks

Shane

_________________
Canada

Rick et. al, The common language is stress and strain....these concepts are the foundation for about 50% of mechanical engineering. They're really not too hard to get a grasp of. Stress (in english units) is measured as PSI (pounds per square inch). Strain is basically the % change in any one dimension.

Here's a good link that I found:
http://www.engineersedge.com/beam_bending/beam_bending2.htm

I'd recommend paying particular attention to the Max Deflection Equation on this page. The value of "I" (moment of inertia) in this equation can easily be found using the Base and Height dimensions from the cross section of the part.

I = (Base * Height^3)/12 (all units should be inches)

W is the weight (in lbs) you are putting in the middle.

Then you can solve for E:

E = (W*L^3)/(48*Max Deflection*I)

E is variable used to represent Modulus of Elasticity...which is really just stiffness.

It would probably be a good idea to record the density of each board as well as the E values...

I hope this helps a bit....the beauty of evaluating E values is that you can test pieces of varying geometry. Not everything has to be "standard" size in order to get a meaningful reading of the stiffness.

_________________
http://www.PeakeGuitars.com

Shane-
I brought this topic up in a thread a few months ago.

There's a big difference between getting MOE data on wood samples (those numbers are independent of the testing setup to a great extent) and getting deflection data to answer the questions:
"How flexible is the top when you decide it is ready for bracing?"
"How flexible is the top in a great-sounding finished guitar with a body plan of xxxxx?"

Folks generally tend to 'play their cards close to the vest' on the latter topic.

Cheers

John

No doubt at all John. I have two reason's for asking this question. First is that a standard would allow me to tell people what some samples from a specific part of a specific tree were averaging out at. Secondly, it would be interesting to see what people are coming up with as an answer to the last of your questions. It is clearly understood by me that some (or all) may not want to state those numbers. But it is also understood that any numbers supplied should only be ball park for any builder. I think intuition will ALWAYS play a huge roll in building, there are just too many variables in the build process to control them all, like will you actually shave the braces proportionately on each build, I would suggest that there will always be differences, based on wood strength and just the process of shaving.

Thanks

Shane

_________________
Canada

Like a few others have suggested here, the meaningful number we're interested in is the modulus of elasticity (MOE) also known as youngs modulus. When you arrive at this number materials can then be compared regardless of dimensions, whether or not the top have be joined. This number can vary by 2x btw and it's different for cross grain and parallel to the grain.
This is the standard engineering property for comparing material stiffness.

I also want to point out that you could simply dimension everything the same, set that famous can of beans on it, sort by deflection, and keep records of it. This would be just as meaningful for most folks. You can't really share the data of with that method though unless like Rick suggest everyone agrees on dimensions and the brand of beans.

Modulus of elasticity is a way around that.

_________________
Jim Watts
http://jameswattsguitars.com

The variables involved:

Length, width, and thickness of blank
Span over which it's to be deflected
Weight placed on it
Placement of weight

You can account, in formulae, for differing physical dimensions of a board. That said, wood isn't a super-homogeneous material so it's far more desirable to just use the same dimensions every test which will give directly comparable numbers across all setups.

The span over which the deflection is tested can easily be held constant. As well, to make sure the setup is 'honest' the span should be reasonably sharp (ie: you've got a cleaner span if your two supporting points are pieces of angle rather than two large-diameter pipes since on the pipe setup the endpoints move as soon as there is deflection)

Constant weight is pretty easy with a digital scale. If the weight used is relatively large and accurate to within X grams then a good margin of safety can be assumed.

The last variable is ignored by a lot of the setups, and it's really important. The weight should be applied cleanly to the exact center of the blank and distributed in some agreed upon way (ie: On a 3" circle of MDF or somesuch). If the weight isn't applied the exact same way every time, the results are much less meaningful.

Based on all that science stuff, the setup I plan on using (barring someone proposing a simpler way to get good results):

-Two pieces of aluminum angle placed peak up a set distance apart on the test rig.

-A 5 lb weight placed on an acrylic or MDF base which is located dead center on the blank (so weight is always similarly distributed). A beefy piece of steel round-bar would self-center and do the trick, as well, though it might be a hassle for others to acquire compared to 5lbs of random weight.

-Blanks will be milled to constant thickness, width, and length. I'm weighing the pros and cons of testing pre or post join. It's easier to prepare the blanks to set dimensions pre-join (smaller equipment needed), but testing lateral stiffness would be a lot harder on an unjoined blank.

-Deflection will be tested with a standard gauge, fixed in the test right and properly calibrated.

I'm thinking 16" span would work well on a setup to test lateral and longitudinal stiffness of joined blanks. Any thoughts?

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

 Other top deflection testing set-ups -

David Hurd - Left-Brained Lutherie - shows a defelection test bed in his article "Static Measurement of Young's Modulus Parallel to the Grain". Includes some math. I think I saw this on his website.

John Calkin shows-off a beautiful defelection testing machine built for Huss & Dalton in a recent(?) American Lutherie article.

Roger Siminoff has a pretty elaborate deflection testing procedure in his paperback The Luthier's Handbook.

Hope this helps.

 

Here's a sample of the way I keep my data and what I check.
My weights a little heavy in this sample. I think about half of it is about right.
As you can see I just set up a standard spread sheet and track my data that way.

_________________
Jim Watts
http://jameswattsguitars.com

Sorry for all the white space
Don't know how that happened, well I have a clue.
Oh I do the same thing for braces, but I only check the deflection along the length.

_________________
Jim Watts
http://jameswattsguitars.com

I have been gluing on the tops to rim first. Once this is dry, I test deflection at the bridge and tap tone of the top with a tuner. I can adjust the deflection and taptone by shaving the braces or by adding mass. Of course I flex things to get an intuitive feel before Starting the process of gluing on braces.

JIm, I assume you have the formulas embedded in that spreadsheet, right?   Can you display that?

[QUOTE=azimmer1] I only deflect longitudinally. [/QUOTE]

Andy, why only longitudinally?

_________________
Mike

"The Dude abides. I don't know about you but I take comfort in that. It's good knowin' he's out there. The Dude. Takin' 'er easy for all us sinners. Shoosh." The Stranger

[QUOTE=SniderMike] [QUOTE=azimmer1] I only deflect longitudinally. [/QUOTE]

Andy, why only longitudinally?

[/QUOTE]

That's a good point. To get the full picture, we should also test perpindicular to the grain. The deflection and response of a guitar top would definitely depend on this value as much as that of the value parallel to the grain.

_________________
http://www.PeakeGuitars.com

Jim my friend I really like your record keeping and spread sheet - any chance of sharing the spread sheet (without your data or course)?

Regarding standardizing dimensions and the brand of beans - HGF, GAL, and ASIA would never be the same with everyone having to wear a respirator.......