Official Luthiers Forum!

Hey all,

I am currently pondering Top bracing ... and I need to get my head around the characteristic sounds of various schemes...
My specific plan is quickly knocking up a couple test guitars with different top bracing schemes and then modifying them to see what happens..... Yes, A Science project.

One Idea I have is starting out way too light (on the Lower X) and adding finger braces around the rims to re-stiffen the top.
Maybe fool around with Ladder bracing, fool around with different X's (Center stiff vs Center loose), tapering vs scalloping, or even something like Taylor's Asymmetrical X.... or checking the concept of "Section Modulus" -- whether the Statics book is approximately right with shapes vs strength.... or Carbon Fiber capped braces..

Unfortunately, Doing modifications on that sort of thing is pretty challenging through the soundhole -- so it will need some sort of "Trap Door"

I want some recommendations for what you do when you are building "Science project" guitars.
Like you have some ideas and want to try them out quickly....

What sorts of stuff get left off? In this case, I am thinking minimal or no Rosette, no bindings, back strip, or tail graft at all.
What sorts of wood get used to keep costs lower? Ebay 2nds neck? Maybe use cheap Mahogany or Maple for back and sides for easy working.... Would it be too terribly heretical to use Masonite or Birch aircraft plywood back and sides -- because it is just so cheap and available

Or... is this the place where a "Sacrificial" $50.00 Pawn Shop beater can be handy.....
Personally, I feel absolutely no remorse when cutting up a Miscellaneous Asian plywood Beater....

Thanks

John

this is where FREE wood comes into play!!

table tops, church pews, bar counters, fire surrounds, coffee tables, and wood from dumpsters.

that'll be about 90% of everything I've ever built then!!

_________________
My soundclick xx luthier blog xx luthier soundclick

John, It may be handy to make one with a removable back if you want to play with or change the top braces. Set the back into a bezel inlet along the rim and use a bunch of 4-40 screws to secure the back on so it still works like it was glued. I have made major changes to top bracing after it was glued to the rim but the back was still off.
There is likely other ways to do this but this was one I was thinking of.

Sounds like an interesting project. I hope you keep us in the loop!

_________________
Wade
Nashua, NH
http://www.wadefx.com

Why not put some of the smaller braces on the outside?

Some sort of objective measurement of sound is a nice reality check in cases like this. It's easy to hear what you think you'll hear, but a microphone and computer with an FFT program are harder to scam.

Leave off all the inlays, binding and finish. I built the 'corker' with a removable neck, and have swapped that onto another box for further experiments.

NOTHING takes longer than science, except art. As (iirc) Gabriel Wienrich pointed out, the problem with science is that you're always doing something you're not good at. By the time you get to where you can do the experiment well, you've got all the data, and another question that needs to be answered. Remember, too, that you don't know what you don't know: you might get results that make no sense because of some little thing you don't understand. This is the stuff that takes time to figure out, and then, of course, you find out somebody else did that years ago, and 'everybody knows that'. My last three week project took three years.....

I've heard of people glueing on the back with a layer of heavy kraft paper between (brown paper grocery bag paper). It makes a weak joint that can be separated cleanly with a chisel and redone. I haven't done it, but sounds reasonable.

Al,
One thing I feel like I really need is to find white papers on stuff that has already been done.
Unfortunately, besides the GAL Red Books -- that sort of info seems pretty hard to get ahold of.
I have found plenty of abstracts.... but for some reason the actual papers may not be available.

Like... most of the "Luthier" Symposia and Conventions don't have any real record of proceedings...
and the CAS Journal and papers seem to be at a college Library up north somewhere, but not accessable otherwise.

Where can you get this stuff?

I actually got my "Minimal Guitar Test Bed" ideas when poking through your website, but I didn't know the Corker had a totally removable neck -- which is a great idea I will definitely pursue.

I am a believer in "Feynman's Dictum" and his "Cargo Cult Science" Corollary -- so I really want to build a couple different designs I can run "Head to Head" rather than 1 bed that gets generations of modifications.

The part that is vexing me is not knowing enough about what to look for other than "Responsive" and "Sounds good to me"
and as you well know "When a man does not know which port he is sailing for, No wind is the right one."

Thanks

John

[quote="martinedwards"]this is where FREE wood comes into play!!

table tops, church pews, bar counters, fire surrounds, coffee tables, and wood from dumpsters.

I agree! I have found some nice wood while walking my dog from scraps and old furnurture that neighbors have thrown out in their trash or were just looking to get rid of for my wood working projects.

Just thinking out load......but for a test bed to test soundboard bracing do you really need teh fret board extension? You could end the fretboard where teh neck joins the body to make neck removal easier. As long as you are comparing different soundboards on that same body (whether you had a fretboard extension or not) seems it would be an apples-to-apples comparison.

Also seems like you could run oversized lining on the outside of the rim where you could bolt the top around the perimeter where it could easily be removed. Use a paper gasket cut from cardboard or thin gasket material and it would seal off the sound chamber. If needed, you could make a flange that fits around the perimeter that bolts down so the top has a tight fit around the inner lining. The nice part about this setup is you could switch tops back and fourth for recording or to get other folks opinion.

Shane at High Mountain Tone woods had A grade Lutz tops at a good price and you could probably get a better deal if you buy in bulk. They may not be perfectly quarter sawn but it's a fair comparison when compared with each other.

Seems on guitar frame built where you could switch out the tops would give you the best comparison without a different neck, back, or sides influencing the tone.

_________________
Formerly known as Adaboy.......

It is really difficult to build two instruments that are 'the same'. Even when you match the wood and construction as exactly as you can they can still end up sounding different. I'm going to try that experiment again at some point, to see if I've really figured out what I did 'wrong' the last time, but until I can make a real matched pair, I'm not going to take any results from that sort of thing by itself too seriously.

It's far better, IMO, to use one instrument that can be modified in some way. That was the idea behind the 'corker': the only difference between 'all closed' and 'one port location open' was the weight of two corks; hardly enough in itself to alter the tone of the instrument noticably (and, of course, one could always stick that much weight on if one was really worried about it). Yes, a side full of corks is not the same as a whole side, but that's not the point. The issue is whether a side with twenty corks is enough different from one with eighteen to bother about that aspect.

This brings up another issue: the effect of ellapsed time on perception. Most people don't really remember sounds as well as they thingk they do, it seems. It also seems as though other sounds in between tests can mess things up. So, if you really want to compare the sound of something in two different configurations, you need to be able to switch back and forth between them more or less silently, and quickly. Again, the corker was designed to be able to do that: there is a little noise from manipulating the corks, but not much, and they could be switched quickly. If one had to stop and re-tune, for example, I'd think that would be a deal-breaker.

All of this makes it awfully hard to do direct comparisons of the percieved sound of different brace configurations. I did an 'almost matched' pair assesment of 'regular' and 'double-X' bracing a few years ago, and the results were interesting, but far from conclusive. If they were not consistent with other stuff that I've seen, I would not have put much weight on that one test.

Obviously, if you can get some objective data to back up subjective impressions it really helps with both of those problems. Again, I did that in the bracing experiment, and it's one reason I trust the playing tests to some extent. Objective data by itself is not convincing, though: it's so hard to say what it 'means'. You can talk all you want about spectral differences: if people can't hear them, they don't matter, and vice versa.

Maybe I'm being too fussy on this stuff, but with all the myth out there, I think it pays to be careful and at least try to settle things as much as possible. This makes it hard to design good experiments. On the other hand, a bad experiment still takes a lot of time to run, and you are less likely to learn anything from it, so it just seems better to try to do it well.

On the 'other' other hand (shades of FDR and economists), there is no 'perfect' experiment. You can nit-pick anything enough to call it into doubt, but we can't let that stop us from trying. If nothing else, a poorly set up experiment can tell you waht NOT to do, which is actually valuable information.

Yes, it's hard to find the articles. Have you got Jansson's book: "Acoustics for Violin and Guitar Makers"? It's available for download in .pdf free, and has a lot of good info, as well as references to many good articles.

It's sad that the Catgut Journal is so hard to find. A friend of mine offered to scan all of them into digital form when the VSA took it over, but they seemed not to be interested. Part of the reason it was abandonded in the first place was that the editor felt (I think wrongly) that the small circulation, and the fact that some 'pop' science outlets didn't refer to it when they could have, meant that it was irrelevant. Now, of course, there's no place, aside from the 'big' acoustics journals, that you can go to publish a peer-reviewed article on guitar acoustics. I'm still ticked off.

I guess as the engineer working in a large production plant.. I think:
"The Same" is a great thing to aspire to.... but They are never going to be *the same*...
All you need to do is measure to 1 more decimal place and they are different again.

Maybe part of the trouble is that these guitars are built to respond to subtle changes...
specifically to be responsive enough so that differences come through.....

This would create a problem that subtle "Intentional" changes blend into the weeds too easily....
and we end up wondering "Was it the change, or was it something I wasn't paying attention to?"

We have a couple choices.....
Large sample sizes
Large changes.
and Making the rest of the instrument fairly unresponsive

If we were running the Taylor plant.... we could easily get the large sample sizes.
Knock out 50 of each change Today and measure.....
We could quickly get ahold of the "Average" effects of a change.
and that is kinda what Bob Taylor did...

But when I am mucking around in the weeds with "Small" sample sizes... of 1...
Changes can't be subtle... and probably... the rest of the instrument shouldn't
be so responsive.

This way, what I am fooling around with only has to be similar enough....
so that unintentional differences in the rest of my "system" are driven down into the mud...
so that the Change could potentially be what drives a difference or no difference....

I guess this would have some repercussions that I am maybe not ready for yet....
like using thin Masonite or Aircraft birch plywood for soundboards when testing out bracing schemes....
or maybe using heavy, stiff Masonite bodies when testing out spruce soundboard changes...

But... as you said....
It is science, right?

Thanks

John

There are several variables that make the scientific testing of an artistic device difficult. First there is the complexity of the instrument itself. There are so many construction and design variables: body size and shape, plate thickness, glue type, finish type and thickness, bracing pattern, brace size and shape, etc., etc., etc.. If Al can't make two that sound the same then it would be quite hard to do experiments that could actually be replicated even if you were to find out what you think made the difference.

The next variable that inhibits scientific testing is that the instrument is used for so many different types of music. What might sound great as a bracing pattern for bluegrass might not work well for jazz. Discussions on the OLF about ladder bracing vs. x-bracing were limited by the fact that the proponents liked different kinds of sounds and different kinds of music.

The next variable that might affect results is the test guitar itself. If you make a guitar that is unresponsive in areas that are not part of your test you won't find what variables make a great instrument because all of the components work to make the sound. Some have advised using techniques and materials to make the test guitar easier to take apart and adjust, but those will have an affect on the sound and won't tell you if your experiments will have the same result on an actual instrument.

The player is another variable that will affect the results. You need someone to test drive that can get the subtle differences to become apparent but that person's playing could vary from test to test.

You have to know what the purpose of the testing is. Do you want to do it so that you will find something that no one else has discovered (unlikely) or are you doing this so that you can find what bracing makes the best sound for the music you like to play or hear? As you said, there probably isn't much literature out there on guitar experiments but there are guitars.

I would suggest that you play as many guitars as you can old and new, experimental and traditional, and find the one that you like best. Try to replicate that instrument in the most basic way, no rosette, bolt on neck, minimal finish, etc. Then have someone whose playing strikes an emotional chord with you play it. You could have a lot of people rate the sound on some scale you devise, but I would think that 1-100 might be good. Then make your modifications without changing too much and do it as quickly as possible. Then have the same player play it and the same people rate it. That would probably give you the best information you are looking for.

John How (and others I'm sure) sent a guitar on a road trip around the country to get responses from different players. He got a lot of feedback from some very talented people. It might not have been scientific but it was definitely helpful.

Ultimately you need to determine why you build and who you are building for, but I don't think there is any substitute for building a lot of guitars so that you gain a working knowledge of the variables involved. Good Luck.

_________________
Ken Franklin
clumsy yet persistent
https://www.kenfranklinukulele.com

Oh, for a few weeks in the Taylor or Martin plant! You could select out top, back and brace wood with properties within specific ranges, have them build the guitars with their well-known QC methods, and then see what you got. It would be fairly easy to control for, say, cross grain stiffness in the tops, and find out what difference that one variable really made.

Ah well, dream on.....