Official Luthiers Forum!

I use a lot of the information and procedures from the Trevor Gore with Gerard Gilet 2 volume set Contemporary Acoustic Guitar Design and Build also available at stewmac and they stewmac offers half price on books and plans if you are a StewMax customer. One purchase more than pays for the membership!

I get a lot of questions about using tap frequency measurement to determine the target thickness for a guitar top or back. This procedure is based on the fact that there is a known relationship of the young modulus for a rectangular bit of wood and the frequency that the plates vibrates. You can calculate the young's modulus knowing the dimensions and the mass of a rectangular plate and the frequencies that it rings. It is better explained in the above book. There is an equation in the book (4.5-7) that allows one to use that information to establish the target thickness for the plate.
Here is a video of me doing the test on a port orford top plate. This is NOT a tutorial and really requires the information from the BOOK



Post video I use a defection test to verify the cross grain as I was suspicious that the cross grain frequency was 2 x the Long grain. The deflection test gave me very close to same answer.

Other information for the video:
From Trevor Gores Site Technical note on collecting spectrographic data
Software used: visual analyser download, Strobosoft



My current project provides an example where I switched tops. The video shows an example of measuring a great tap sounding port orford plate. It was very stiff and my hands on lead me to believe it would make a good top for a nylon string guitar I am making.

Referring to the image below one can see that the top target of 2.03mm is relatively thin for a classical. This thickness calculation normalizes each plate tested mostly for stiffness but also takes the mass of the plate into consideration. But the density, 464 kg/m^3, will result in a top that is about 20 grams heavier than I normally use in my nylon string instruments. Maybe ok for a steel string, but I go through a bunch of effort to remove a few grams from my classical bridges, maybe this wood is not the best choice.



So I went through my stack and pulled out a WRC cedar top I have had for 6 years or so. Very tight grain lines across the entire plate and perfectly quartered. It also had a wonderful tape tone that rang clearly over 5 seconds. The tuning app used it in the testing heard the ring for over seven seconds. So I measured all of it attributes as shown in my video.



As one could see that this plate was not near as stiff and resulted in a thicker top, but with this woods significantly lower density of 331 kg/m^3 even with the thicker top it will weigh more than 20 grams less.

While the method uses the plates tap tones in combination with it physical measurements, I have seen alternatives where the plates while thicknessed are normalized for mass or using deflection test normalized for stiffness. I like the method I got from the Gore method as I am able to make determinations early in the process. If the two alternative methods were used with the Port Orford, in the case of normalizing for mass I would have ended up with a very thin and floppy plate; normalizing for stiffness I would have ended up with a heavy plate. Working with wood one knows these different methods would be OK as in a given species of wood there is a good correlation between density and stiffness.

I never used Port Orford before so it was hard to know how to compare it to wood I have experience with. From simple measurements I would know that it was very dense. I could also feel that it was very stiff, but was the measured stiffness enough to make up for the above normal density. With the work of making sure the plate was flat and rectangular I was able to make that determination very early.

_________________
http://www.Harvestmoonguitars.com

It's pretty cool data!

Hi John,
I’ve been using Trevor’s method for determining plate thickness since the book first came out. It just made a lot of sense to me as a way to target comparable sonic performance from one top to the next. Note I said “comparable”, not “optimal”.

My setup and routine are similar to yours, but I’ve streamlined data collection and calculations on a single row in a spreadsheet for each plate. The calculations for my entire top inventory are on a single Excel sheet. Adding new tops is as easy as inserting a new row. Normally (but not always) there are 2 plates per set, so two predictions of target thickness. Fields for data input are coloured, and calculations for the remaining cells are automatic. When a top is allocated, I record the guitar ID and shade the rows to indicate the top as no longer available.

I looked at VA, but have settled on Strobosoft for everything, as it has a pretty sophisticated “sampling” algorithm for taps and has graphic display of frequency response. I use a “Snowflake” USB microphone, because the reviews were excellent and the price was right. For a mallet, I have “standardized” with a drafting eraser stuck on the end of a long screwdriver. It has just the right heft and leaves no mark whatsoever on the wood.

About once a year, I prep my new tops (surface sanding and cutting to rectangular shape) and take the required measurements to update my spreadsheet. I find it takes one or two tops to regain my dexterity to get good tap responses. After that, I can whip through top sets fairly quickly. The method can tolerate a wide range of test plate dimensions (unlike deflection methods), but every now and then I have tops with defects outside the template, and for those the method is less useful. Backs can also be assessed, but irregular shapes and defects are even more common, so I find it less useful.

My only complaint: you need to do this testing in a dead quiet setting. Can’t have music playing in the background and certainly no air filtration!


Sent from my iPad using Tapatalk