<!doctype html public "-//w3c//dtd html 4.0 transitional//en"> <html> Delwin D Fandrich wrote: <blockquote TYPE=CITE>  The understanding of soundboard function has moved a bit beyond the purely empirical--even beyond 90% empirical--though certainly not to the stage of pure science either. Still, techniques such as modal analysis have enabled the study of soundboard function at a level not even dreamed of even thirty or forty years ago. It is my hope that before I pass from the scene that this understanding will be still some closer to scientific and much less reliant on the empirical. Del</blockquote> Been reading and weeding through and find much that is interesting, but this comment caught my eye and I must admit is confusing. I wrote about modal analysis a year and a half back and you came out and said it was basically useless in designing soundboards, since the conditions for measurement are different then the conditions for full strung, and if done full strung well the deed is already done so to speak... further you pointed out then that you meant that there was no reliable way of forcasting said changes.   From the fall of 1999  where we were in a disscussion about impedance matching, and the usefullness of modal analysis came up I submit the following three posts. <blockquote>- ----- Original Message ----- From: Richard Brekne <richardb@c2i.net> To: <pianotech@ptg.org> Sent: Friday, August 13, 1999 3:12 PM Subject: Re: Impedance Matching   > > Hmmm.. In the section of the Wogram article entitled "Influence of string > tension" He seems to be saying that the condition of an uloaded soundboard is > not a problem. He says that the basic impedance curve and sound radiation > curves remain basically the same and that only the lowest resonances are > effected and that the effect is predictable enough to take into consideration. > Is this outdated ?? - ------------------------------------------------------ I don't know what the basis for this assumption is, but my own -- admitidly limited -- tests have indicated otherwise.  Soundboard characteristics do change as it is loaded.  Partly because of the increase in stiffness within the panel, but also because the string plane affects the mobility of the entire assembly. One of the problems I have with most of the testing I have seen on the piano soundboard is that it was done on an unloaded board and is, therefore, largely meaningless.     > The article seems to point to some significant degree of potential for > re-distributing resonance points for the different nodes by introducing new > stiffness and / or mass to different areas of the soundboard. Course one would > have to gain lots of experience to know just how, why and where to introduce > such changes on an existing board, but it seems like enough information can be > obtained by the modal analysys to accomplish this. Or is this just a sugar > coated red herring ?? Opinions ?? I consider modal analysis of the working soundboard to be a good first step. You know, like, "one small step for man...." Del</blockquote> Followed by the following notes from Ron N     <blockquote>Date: Mon, 16 Aug 1999 20:01:48 -0500 (CDT) From: Ron Nossaman <nossaman@SOUTHWIND.NET> Subject: RE: Impedance Matching >Del, >I'd second the notion that soundboard characteristics change with load. >For SURE! > >While I'd agree that a modal analysis is meaningless without the full system >(strings), it can be very valuable calibrating FEA models.  It's very simple >to add/remove components of a finite element model and is invaluable for >checking the accuracy of parts of a very complicated model (like a piano). > >I know, "so what are you waiting for"? >I'll get back to it ;-) > >doug richards >San Jose, CA   Well, sure, soundboard characteristics change under load, but how much? My point is that we, or at least I, make design decisions based on how I think the board will respond in use, not how it sounds lying on the floor. I have no way to determine how the board reacts without the string load. I can't hear it without the strings, so I am, in effect, factoring the difference into the design. How well it works in use is determined by how good my initial design was, how accurate my judged factoring was, and the characteristics of the materials being used being what I thought they were. If we could compute an unloaded impedance gradient of the assembly, across the scale, with the physical information we *do* have, and working backward from the way it *sounds* in use, compared to what the computed gradient says, we should get both a roughly quantifiable cause and effect relationship (which we are already working with, without the detailed figures), and some idea of the difference between the loaded, and unloaded states. Again, I believe the differences between the unloaded state of the assembly, and the loaded state, can be quantified to a practical degree without taking a FEA model to the sub atomic level on every single piano that's worked on. The impedance loads, taken from the scale and bearing schedules should be relatively simply applicable to the computed static impedance model of the assembly, wouldn't you think? It ought to be proportional, shouldn't it? How close to an absolute dead center impedance match/mismatch is necessary before we are beyond the resolution (granularity?) of the medium with which we are working... wood? What the heck IS the ideal match between soundboard assembly and string plane, given high impedance mountings and terminations? That's what I was getting at. I suppose it's still a chicken/egg sort of thing though. Someone has to do it the hard way and make it work before it can be simplified to usability. I sure wish I could pound some of the missing math receptors into this old brain, I could certainly use them.  Ron N</blockquote> And then again from Stephen Birkett:   <blockquote>Date: Sat, 14 Aug 1999 00:12:48 +0200 From: Richard Brekne <richardb@c2i.net> Subject: Re: Impedance Matching Stephen Birkett wrote:     <blockquote>> > I have my opinions about modal analysis (mentioned here once a > while back)...that, at best, it can tell you that the soundboard is > working as you already know that it is supposed to work. As a practical > tool for design I have my doubts it will ever give anything, primarily > because it treats the sb unloaded and in isolation from the interactions > with the rest of the piano.</blockquote> ></blockquote> -- Richard Brekne RPT, N.P.T.F. Bergen, Norway <a href="mailto:rbrekne@broadpark.no">mailto:rbrekne@broadpark.no</a>  </html>