Official Luthiers Forum!

This is my first post, so hello everyone!

I have been reading the forum for a while, and I have to say it's an amazing source of information.

I want to start deflection testing my tops, because it seems to me like the most realistic way to get a consistent start on a new top. Can anyone give me some pointers on how to set up for this? How much mass should I use? Where should I place it? I have an idea in my head, but I'm sure I'm leaving something out. What kind of range of numbers do you guys usually shoot for? Are there known optimal deflection numbers for certain builds or is that all a trade secret?

Thanks a tun
Andy

Hi Andy,

Welcome aboard. Don't have too much to add yet because I just started taking deflection measurements myself but haven't calculated E or I yet. The book "Left Brain Luthiery" by David Hurd has all of the formulas (as does my engineering mechanics text) and they're not too complicated. Basically you want to pre-load the top with a weight, then add additional weight and measure the difference. With that plus the second moment (I or http://en.wikipedia.org/wiki/Second_moment_of_area) you can predict the deflection of a top and make adjustments depending on the wood you're using.

Hopefully Al Carruth with chime in as IIRC he's got a rather novel way of getting the various E's (i.e. the long grain and cross grain E) by tapping + dimensions. He also has written before that it's really the long grain vs. cross grain stiffness ratio that decides whether a piece will sound good in a particular guitar.

Really I just chimed in because I too am planning on re-educating myself on this topic and because I see you live in Waterford which is not too far from me. In fact I met a luthier once in Kenetic systems, it wasn't you was it?

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

Hi Andy,
Welcome to the forum. There are quite a few posts in the archives that i would highly recommend. If you use the search facility and look under "deflection testing for tops",you will find loads of info. I have printed a number of these out and read them semi-regularly.
GregE authored one such post on the 11th of July last year. There are a number of pics of different methods and also some ballpark figures to get started with.

Regards
Craig.

There are a number of good ways of testing a top, and which one you choose to use will depend on how much time and money you want to spend, and what you think is important. As always, there are lots of opinions one way or the other about every aspect.

I don't use static deflection tests. In part it's because I already had the equipment to do dynamic tests, and in part it's because there is information you can't get from a static test. It's NOT because one or the other type is more accurate when properly done.

The simplest type of static test is to just measure the deflection of a piece with a known thickness under a known load across a given span. If you always use the same mass, thickness, and span you can use the deflection info directly without having to calculate a darn thing. All you need is a few data points of tops that worked, and you can figure out if the one you've got in your hand now will or won't. You do need to be careful to do the test the same way every time, but that's fairly easy.

If you start with a rectangular piece, such as a half a top, you can easily find the equations that will tell you the Young's modulus (E) from the dimensions and mass used. This gives a more direct way to compare things, and allows you to tap into other people's data bases. If I tell you I had a top that had a lengthwise E value of 10000 MegaPascals (metric measurement) and used it to make a classical guitar at a thickness of 2.8mm, you could scale the thickness of any top you have from that, assuming you're construction is more or less like mine.

The thing you don't get from static tests is a measure of the 'Q-value': how fast vibration energy is dissipated within the wood. This relates to the 'long ring' we listen for in tap testing. With the proper computer software you can actually find the Q of the wood from a tap test, although there are pitfalls. I use a 'bandwidth' method that relies on driving the top with my signal generator and reading the amplitude as the frequency is changed around resonance. Knowing the mass and dimensions of the sample, I can use the frequency data to calculate the E values as well. This method avoids some problems inherent in staic deflection tests, and runs into some others: there's no free lunch.

You can drive yourself nuts trying to do a 'perfect' test. There are all sorts of things that can mess up any test, and reduce the accuracy and usefulness of the data. What's most important, IMO is to always do whatever you do the same way, and to be aware that the outcome will be somewhere in range of 'reality'. I figure I'm within 10% of the 'real' stiffness of the woods I measure, so if two pieces are much closer than that, they're 'the same' for all intents and purposes. The actual difference in thickness you'd get would be a few thou one way or the other, and you can lose that much cleaning up before finish. The idea here is not to be 'lab perfect', but to get information that you can use in the shop without blowing the budget.

Thanks everyone

Wow I knew it was going to be more complicated. I did some searching, but I must not have got back far enough in the archives to find anything useful. I'll look again for GreagE's post.

Mr. Birko, thanks for the warm welcome. That guy you were thinking of wasn't me. I work down at Wilson Fine Violins in Birmingham. If any of you guys are ever in the area feel free to drop by.

Alan, you have given me a lot to think about. What software do you use? I fancy myself to be computer savvy and I have a simple function generator that I use in tuning plates on my other instruments. Your method appeals to my EE background. My inner geek is dying to know more about the math you are using, but I don't have a huge budget for new gadgets.

You guys certainly have given me a lot to think about.