Official Luthiers Forum!

Hi luthiers,

I started doing a soundboard deflection test for the first time at the beginning of my current built (4th) and I'm wondering how to compare those results with a different guitar shape. My current built is a dreadnought with a redwood top and the next one is a custom jumbo shape with a sitka spruce top, so a slightly bigger instrument. I don't think the type of wood is a factor since what we're looking for is a deflection under a known load regardless of the type of wood (is it so?) but the shape of the guitar might. However, the waist of the jumbo guitar is narrower than the DN so I wonder if the total area isn't similar and if that's what I should take into account.

My deflection board is a simple mdf board with two 1/4 inch slats 18 inches apart. I load it with a 1.8 kg weigh and the sweet spot is when the soundboard touches the mdf board (1/4 inch deflection). I brought that redwood top at 3 mm to get that deflection (2.8 after sanding). I'm not sure those number mean anything since i just started collecting data and I'm kind of going blind.

Anyway, bringing the soundboard at an optimal thickness is still a bit of a mystery. I know I'll get better at it with experience but it's still a bit frustrating not knowing exactly what it should be.

Any advices?

Thanks

I have been doing deflection measurements on my tops for quite a while. My experience has been that my optimal deflection with spruce tops carried over to mahogany and cedar tops.

It helps to build enough with one wood species that you get a feel for the optimal numbers with your deflection measurement and build techniques.

_________________
It's not what you don't know that hurts you, it's what you do know that's wrong.

I wouldn’t think the slight size difference between instruments would require much variance in your deflection testing.

With deflection testing, the thickness is a byproduct. I record thickness before working any top, but only because it’s required to get the density which is how I choose different tops for different projects.

1.8kg is just short of 4lbs, which is about 20-25% short of the 5lb weight I use, depending on which way you do the math, and your deflection number is 25% less stiff than I use for a dread and larger with the heavier weight. I don’t know if that will help, but tracking your deflection and keeping notes is very good practice.

Sharing data on deflection is tricky because there are so many other variables that effect a guitar that one really needs to build their own database from their own work for it to make sense, which can only come from building, observing, and recording.

I will say though that your top is considerably less stiff than what I do for a guitar that size, but…that might not mean anything because of the other aforementioned variables.

That helps a lot thanks! This first guitar I started keeping track of deflection is not completed yet, I'm in the finishing process now. But based on what you say, I can expect that top to be too flexible so too bassy/boomy. I wish I could have played before starting the next one to figure it out. Of course the top thickness is not the only factor and the bracing will influence the top flexibility too.

Only building a few will give you a map for next time. Just because that top is more flexible than what I would do, doesn’t mean it can’t or won’t work for what you’re going to do, and we may have very different opinions on what we want to hear out of a guitar. Let us know how it goes!

Here’s my deflection rig. I do it after the rosette is installed and soundhole open. I have a standard spot for weight application and measurement. I use one lb. preweight and zero the dial and add one lb. Use an average of several measurements.

For me around 0.045 to 0.050” seems to work well with this system and my build style.

IMG_0179 by Terence Kennedy, on Flickr

IMG_0180 by Terence Kennedy, on Flickr

_________________
It's not what you don't know that hurts you, it's what you do know that's wrong.

I don’t tend to deflection that the final sound board. Instead I have specific sizes that I cut my boards to I then thickness them, join them, thickness again, install a rosette, thickness them again, at each point I record how much they deflect both along the grain with a 2kg weight and cross grain with a 1kg weight. I also measure the tonal quality of the board, its frequency response using FFT . This measurements help to determine what the final thickness should be.

I don't have much experience with using deflection to determine top thickness. I've been using a vibration test to find the Young's modulus of the material, and going from there. You can find the Young's modulus from a deflection test as well, and IMO that would be the way to go. For one thing, there is a certain amount of literature out on how to use the information, and, being a 'standard' format, it's easy to compare results with other makers.

Every measurement you make includes pitfalls. How you set up the test, how you do the loading, and so on, can all make a difference. For example, a 'point' load such as Terrence shows can produce a different deflection than a 'line' load (a pipe or sash weight), where the weight goes all the way across the plate. You can get different results by flipping the plate over sometimes, and the time you take to make the measurement matters. This is not to say that there are no potential issues with the vibration test, but simply to point out that you need to be aware of the problems that can arise.

As an example; some years ago I tested a Sitka top that looked very good visulay, but the the test said that it was far less stiff along the grain than it 'should' have been. The node lines across the top were not perpendicular to the center line, and the 'long grain' bending included a 'cross grain' component, which lowered the apparent stiffness. Trimming the plate to a slightly narrower size shifted the cross grain bending pitch, and gave a much more 'normal' long-grain result. The guitar I subsequently made fom it turned out to be outstanding.

Trevor Gore talks about deflection measurements in his books, as does David Hurd. Hurd focusses on it more, and his book is less expensive (if you can find it: I seem to have misplaced my copy!).

My standard practice is to measure each side several times, find the average of measurements for each side, then average the averages. I don’t use a preload or tare weight, my weight is a longer than average whiskey tube filled with 5lbs of pennies. Full panel, rosette in but sound hole not cut out.

Yes, Trevor shows how to determine the top panel's stiffness (Young's modulus, E) by tapping the piece to get three frequencies from it (using a microphone into a computer). This is more reliable than a deflection measurement, but requires the panel to be rectangular (squared up and flat). From these, you can calculate three E's: long grain, cross grain, and shear.

With these plus the panel's length, width, existing thickness, weight, and the guitar body's dimensions (length, width of lower bout) you can calculate target thickness.

So basically, measure three frequencies, three dimensions, and one weight, get a thickness number. Put the panel through the drum sander until your micrometer reads that number.

You can use my spreadsheets to do it, if you want.

Greg

I use a signal generator to drive the plate and look at the Chladni patterns. This will tell you how 'pure' the bending wave pitches are. That was the problem with the example I gave; there was a mix of crosswise and lengthwise bending bending in the plate initially, so the pitch was determined in part by the lower crosswise stiffness. Without seeing the patterns you might not catch that. As with any test, there can be pitfalls, and you have to understand the limits of the test.

All of this is orders of magnitude easier than it was when I started doing it 40+ years ago. We had to work up hardware signal generators.... I've been recommending a free Android app; 'Luther Lab', that a friend of mine wrote. It has a signal generator module that is easy to use, as well as a frequency analyzer for tap tests. You can take pictures of the modes on your 'phone or tablet, label them with the pitch, and save them in a 'project' folder. There's lots of other stuff, too.