Official Luthiers Forum!
http://www-.luthiersforum.com/forum/

Interpreting frequency response curve
http://www-.luthiersforum.com/forum/viewtopic.php?f=10101&t=56793		 Page 1 of 1
Author:  James Orr [ Thu Sep 19, 2024 12:47 am ]
Post subject:  Interpreting frequency response curve
Hi all.

I'm in the process of trying out Guiliano Nicoletti's approach to measuring tonewood properties as described in Appendix D of Mastering the Sound of the Acoustic Guitar. It's essentially identical to Trevor Gore's approach wherein you measure dimension, mass, and frequency response of a sample of wood, plug the results into the supplied spreadsheet, and get a suggested target thickness.

I have the RTA in REW setup as described in Appendix A and captured around 15 seconds worth of samples of the cross bending mode, but I'm not sure how to interpret the results. Is the peak amplitude of the wave (99 hz) what I'm plugging into the spreadsheet? If this is accurate, does this hold true for all three modes?

Image
Author:  Alan Carruth [ Thu Sep 19, 2024 9:18 am ]
Post subject:  Re: Interpreting frequency response curve
I'd say that the 99 Hz peak is the number you want, but, of course, without all of the information it's hard to be sure.

I have not seen Nicoletti's book. IIRC, Gore records a tap spectrum. I've been looking at the actual Chladni patterns on half top blanks for years, and find the mode shape information useful. Your spectrum is showing other peaks in the output, and I have to wonder about the dip around 88 Hz. A look at the mode shapes will indicate whether there are other modes coupling with the fundamental crosswise bending mode, which would mix in other things, such as the lengthwise Young's modulus, or a combined shear modulus.

It's possible to fool yourself with this stuff. I had a top once that gave a disappointingly low long-E value, but the lengthwise node lines ran diagonally across the top, which was suspicious. I trimmed a bit of wood off the width and re-did the test. I got much better mode shapes, and more reasonable E values. The top ended up making a very nice guitar.
Author:  Kbore [ Thu Sep 19, 2024 9:43 am ]
Post subject:  Re: Interpreting frequency response curve
James Orr wrote:
Hi all.

Is the peak amplitude of the wave (99 hz) what I'm plugging into the spreadsheet? If this is accurate, does this hold true for all three modes?

Image



First, I have the book and am by no means an expert. My reply is to support discussion of the question, I haven't even done this. According to the book: In the graph illustration on p. 141, the accompanying text says:

" The first peak of the curve identifies the FREQUENCY.... to be entered"

so you enter the frequency not the amplitude. Your first peak is 80 hz. On p.143 he references Trevor Gore's recommendation of 75 hz as target for the first peak, apparently an indicator of the vibrational stiffness, as I interpret it. Of course the timber being measured must respect the overall dimension, the calculated holding point and the tap/ impact point (p.140).

Kudos for going for it, more than I have done to date.
Author:  James Orr [ Thu Sep 19, 2024 10:48 am ]
Post subject:  Re: Interpreting frequency response curve
Thanks, Karl! You're a champion!

This is a great book, but I've struggled with this section and think it could benefit from better organization.
Author:  Kbore [ Thu Sep 19, 2024 12:38 pm ]
Post subject:  Re: Interpreting frequency response curve
Giuliano is all over social media, and posts regularly on The Looth Group.
He seems to be approachable, hit him up on Facebook with questions......
Author:  Kbore [ Thu Sep 19, 2024 12:47 pm ]
Post subject:  Re: Interpreting frequency response curve
Alan Carruth wrote:
I'd say that the 99 Hz peak is the number you want, but, of course, without all of the information it's hard to be sure.

I have not seen Nicoletti's book. IIRC, Gore records a tap spectrum. I've been looking at the actual Chladni patterns on half top blanks for years, and find the mode shape information useful. Your spectrum is showing other peaks in the output, and I have to wonder about the dip around 88 Hz. A look at the mode shapes will indicate whether there are other modes coupling with the fundamental crosswise bending mode, which would mix in other things, such as the lengthwise Young's modulus, or a combined shear modulus.

It's possible to fool yourself with this stuff. I had a top once that gave a disappointingly low long-E value, but the lengthwise node lines ran diagonally across the top, which was suspicious. I trimmed a bit of wood off the width and re-did the test. I got much better mode shapes, and more reasonable E values. The top ended up making a very nice guitar.


Alan,
Bluetooth generator and Bluetooth amp (and smallish 50w broadband speaker) to do Chladni patterns are sitting in boxes. I to make room to actually do it in my small shop. In what stages do you do patterns (individual plates, joined plates, braced top, finished guitar? Is there any one build stage more useful that the others? I'm gonna have to just make room.....
Author:  Alan Carruth [ Thu Sep 19, 2024 5:42 pm ]
Post subject:  Re: Interpreting frequency response curve
I find the fundamental lengthwise and cross wise bending mode frequencies and band widths of top half blanks and use these to derive the associated Young's moduli and damping factors. These, and particularly the lengthwise modulus, are used to determine the top thickness, in a method that's a bit simpler than Gore's, but perhaps not as exact. Once I have the top braced I use the mode frequencies, and especially the shapes, to 'tune' the bracing stiffness to work with the top (or that's the intent!). Once the guitar is together I get 'impulse spectrum' measurements of it, and also look at the assembled modes of the top and back, and take the pitch of the 'main air' and 'neck' resonances. I get a fair amount of use out of my signal generator and the recording and analysis software on my computer.
Author:  James Orr [ Thu Sep 19, 2024 9:12 pm ]
Post subject:  Re: Interpreting frequency response curve
Alan Carruth wrote:
Once I have the top braced I use the mode frequencies, and especially the shapes, to 'tune' the bracing stiffness to work with the top (or that's the intent!).


Al, this might be grossly simplifying it, but are finding those initial chladni patterns as a baseline, and then working the bracing until you get the chladni patterns you’re after?

Is this done free from the rim, or attached to the rim?


Sent from my iPhone using Tapatalk
Author:  Alan Carruth [ Fri Sep 20, 2024 11:53 am ]
Post subject:  Re: Interpreting frequency response curve
These are 'free' plate modes.

Every top is a bit different, so while there's a sort of 'ideal' set of modes that seems to work well you don't always get that. The basic idea is to get the top to work as well as possible, and avoid things that have been associated with problems in the past. For the most part it's a matter of getting modes that are well defined, and with smooth curves in the node lines rather than sharp corners or openings. The pitches of the modes do convey some information about the over all relationship of stiffness to mass of the top, so we try to avoid going too low in pitch, but the exact pitch of any mode doesn't have much predictive value. Everything changes a lot when you glue the top and back to the sides, and there are just too many variables to take into account.
Page 1 of 1 All times are UTC - 5 hours
Powered by phpBB® Forum Software © phpBB Group
http://www.phpbb.com/