Official Luthiers Forum!

This is part of thread viewtopic.php?f=10101&t=41876.

Hi all… I thought I'd post the process I am going through to tune the top and back. As a newbie, I don’t know that I have ever seen the process posted and I know this is the sort of thing that’s really interesting to someone who has never seen it done. At least it would have been for me had I ever seen one.

A few notes about my top:

The top is LS redwood.
It is thicknessed at a gradient - 3.2 mm at the upper bout down to 2.8mm at the bottom end shooting for 3mm at the bridge.

While I don’t have a copy of the actual stiffness/density ratios, what I do know is that when we charted this top against other tops Al had done, this top was considerably off the average line in terms of stiffness.

This is the initial check of Chladni patterns. Top row is the top and bottom row is the back. The b/s is ziricote.


Where there are multiple values, those are the resulting frequency drop as I shaved the various braces.
The goal for the top was to get the Frequency to drop for the image where the pattern was a ring in the lower soundboard (357 hz initially). My attempts to get there follow…

This was my initial pattern where the goal was to have a complete ring in the lower bout. It started off at 357 hz.


For the most part when I shaved the braces, (at least in the beginning) I would just take one pass off the braces.

With the goal to transform that initial pattern into a circle by closing the open bottom and upper sides, I took a bit off the lower legs of the main X braces.



The resulting change started to bring the bottom of the ring northwards.
I did not log the change in Hz at that point.


That change seemed to bring the ring up form the bottom, I went ahead and shaved another layer off. I did not measure how much the thickness was reduced that this point. The result seemed to again bring the lower portion of the circle upwards. The frequency dropped to 350.



The decision was to see if reducing the lower legs of the secondary X would tend to close the upper corners of the circle (if circles had corners!!)


The result was another drop in the frequency but I didn’t really see a change in the pattern.


Time to try reducing the upper legs (tone bars?) of the secondary X to see what effect that would have.


Hmmmm... i think i have to post this now. It's not allowing me to add new images.
No error, just doesn't add the new image to the list.

Must be a limit per post. Has to start a new reply...



Missing a photo here but I think it was because I saw no effect from that change. For no better reason than to see what happens, I took more off the 2ndary X lower legs. Reduced them to 4.8 mm measured at the center of the leg from the X to the end.


The result was pretty!
The frequency drops were pretty slow in coming. I was now down to 336 hz.
We started to suspect that the top was much stiffer than expected.



I started to take less pictures of the bracing changes but documented a bit clearer what I was doing.

Because in the earlier pictures the top of the circle was more closed, I tried to reduce the main X legs a bit to see if I could bring that top of the ring back.

I could start to see the top lines start to curve toward each other so it seems that reducing those braces was having the desired effect. Went a bit further on those ~ 7 mm @ center. Not much of a change in pattern but another 5 Hz drop.



I wanted to keep the two sets of braces in balance (ha!) so I reduced the Lower X legs a bit too. Went from 4.8mm to 4.0mm. Must have been too much as it blew open the sides.



The hope was that by reducing the long legs again it would bring the circle back. Fat chance!
Reduced them from 7mm to 6.2mm and while it did not seem to close the sides, it did bring the top down.



Reduced the long legs further. Brace near the bridge is now 10.1mm and the center was down top 5.9mm. Result was the top of the circle was back but the sides were still wide open.



Next I tried reducing the secondary braces near the X. AT this point it was 7.6 mm and I brought it down to 6.6. This left the slope of that brace less sharp. I seemed to close the sides a tiny bit. Freq. was now down to 321.



So… I again hacked away at the long X braces. Brought them down to 9.6 at the bridge and 5.6 at the center. Top of circle came down slightly and sides seem to close a tiny bit more. Freq only dropped 1 to 320. At this point, I am guessing if I reduce the braces to negative numbers I might get that darn Freq. down to the goal.

At the limit... Next page coming...

looks like a fun project. When the top is glued on to the body and the bridge is glued on, those values will change. The bridge really is a major brace. Were your target values targets for a free plate? I have played around not so much tuning my guitars but recording the tap tones using VA at different times in process; hoping to use the information in the future. I will be interesting to see how your guitar progresses.

_________________
http://www.Harvestmoonguitars.com

It’s hard to read my own writing but I think the long braces were again reduced to 9.2 from 9.6 and down to 5.2 at the center down from 5.6. NO change in pattern but another miniscule freq drop to 317.



Reduced the main legs again to 9 and 5.2. Side again seem to start opening. *sigh*



And more… down to 8.9 and 4.9. Top of the circle does seem to want to be starting to move towards the bottom of the lips (no longer looks like a circle!)



Reduced the upper portion of the secondary X and the result was Sad. Less of a circle but still looked decently active in the middle. It became pretty obvious that it was not going to get down to the 186 hz range. The decision was made to stop there before the braces were completely shaved off.



Final brace configuration… pending review next week.



What did I learn?

That tuning the bracing can have a definite affect on the patterns observed.

That the goal is not to try to force the top to do something it doesn’t want to, but rather to try to do no harm to the basic pattern of the raw unbraced top. I don’t have an image of that with me, but it was an almost closed elongated circle that was open on one end. The last result is not too unlike the original freeplate before the braces were added.

I also learned that it is not critical to power up the speaker to the point of where everyone tools start jumping poff the walls... Sorry guys!

It will be interesting to see where we go from here and how this top will sound deviating form the norm.

Rob

Rob,

Can you talk a bit about why a circle is what you and Al were going for? I'm just learning about Chladni patterns and tuning the tops via brace manipulation. Before joining this forum, I was under the impression that I should shave my bracing to make it look like a pre-war Martin and that would be it.

Also, can you talk about your choice to use the double X bracing? Specifically what benefit you feel it has over traditional tone bars and how those benefits will help you get to the type of tone you're chasing?

Thanks again. Wonderful thread.


Sent from my iPhone using Tapatalk

John... i don't recall having a goal for the free plate. What you are asking is interesting. Might one thin the top to a point were a spcific frequency is reached? Makes sense.

So much of this is new to me i don't recall whether we actually discussed a frequency range for the free plate or not. Could have easily escaped my notice at that point as just another bit of info that i had no internal reference on, so it didn't register at the time.

I think something like VA would also be interesting to use. At this point it's more than my head can handle. I may play with something when i am building at home and have time to do a bunch of testing. Al's shop is a couple of hours away from me, so i tend to want to just keep moving forward. That's not to imply that I won't do things over if there's a real problem. Don't want to give the impression that i'm not going for the best instrument i can build. I just may not be taking the time to understand all the details that i might if i had better access.

Justin asked: "Can you talk a bit about why a circle is what you and Al were going for?"

I don't have a good answer.

I don't know that there's a reason for picking that particular pattern. I have seen a number of threads here that all shoot for that pattern at a specific frequency (~ 186 hz) so it may be just that is a common one to work with. Certainly there's several patterns at various frequencies one could focus on.

Maybe it's one that's easiest to work with? It certainly seemed to my untrained eye that there were very definate responses to the bracing changes i made. More complex patterns could well be just too difficult to work with?

Justin said: "Before joining this forum, I was under the impression that I should shave my bracing to make it look like a pre-war Martin and that would be it."

I think the goal is that as YOU build, you want to be able to replicate a sound that YOU like. So the patterns you chose to work towards would help you become more consistent. Whether that happens to match a pre-war martin would be up to your goals.

Justin asked: "Also, can you talk about your choice to use the double X bracing? Specifically what benefit you feel it has over traditional tone bars and how those benefits will help you get to the type of tone you're chasing?"

First my preferences. I primarily finger pick. I like a decent amount of base and i love sustain. I want to work more with alternative tunings.
I also should say that I primarily looked to my choice in wood as affecting the sound and that the goal for bracing is to inhibit the sound of the wood as little as possible. I am sure you can alter/shape the response of the guitar by bad bracing. I am not quite as certain that you can improve on it.

There's a ton of bracing patterns out there and some that when i look at them i wonder how the heck they even work yet they are on some amazing guitars. So i think the pattern one uses is less important than one might be led to believe except in that if it helps to define YOUR sound then it's the bracing pattern to use. AL uses the double X in all his SS guitars. He's done some blind tests at the various lutier meetings and finds for fingerstyle that most people prefer the double X sound in spite of the bad press it had gotten in some guitars (I think it was Guild that had some dogs if you believe the press). So i wasn't stuck on any bracing pattern and since Al likes it, i have no issue going with it.

Glad you're finding the thread interesting!

Rob

Thanks for the answers Rob. I think I might try the double X on my next guitar, which I think will be a parlor.


Sent from my iPhone using Tapatalk

Rob's redwood top tested out with the long grain Young's modulus at 9800 megaPascals, and cross grain at 804 mPa, so the stiffness ratio was 12.2:1. The density was 513 kg/meter^3. The long grain stiffness was lower than you'd expect for a normal softwood that density range, and a bit low even for redwood: softwood from the base of big trees tends to have higher density, as it builds up the latewood lines to support the weight above it. The Q values were 204 along the grain, and 59 across. The long-grain damping was a bit lower than many redwood tops, and those, in turn, have much lower damping than most softwoods.The stiffness ratio was good for the relatively wide Small Jumbo pattern.

In the business, when we talk about a 'free' top, that means 'not glued down around the edges': it has nothing to do with whether it's braced or not. What we've been looking at are pictures of a 'free' top. The resonant modes of an unbraced 'free' top are surprisingly similar in shape to those of a braced top. After tracking a bunch of tops through the process years ago, Mark Blanchard decided hat the bracing can't do a lot to 'improve' a top that is not working the way you want it to. You can, of course, mess up a good top by bracing it badly. Thus the braces, when done right, mostly add stiffness without adding much mass. A god bracing job is one that does that without spoiling what the top 'wants' to do.

Why the 'ring+' mode? The short answer is that it seems to work. That's not very 'scientific' in some ways, but there you go.

I started to concentrate on the shape of the ring+ mode some years ago after trying a 'matched pair' experiment with Classical guitars. Despite my best efforts to control everything I thought was important, they ended up sounding noticeably different (although they were quite similar). The only thing I could find that was not 'the same' within pretty close tolerances was the shape of the ring+ modes on the tops, and the one that had the 'better' mode shape was the one everybody preferred.

There are a couple of ideas about why this might work, but going from there to any sort of 'proof' would be very difficult. The mode shapes are indicators of the mass and stiffness distribution in the top. Although that must have some effect on the way it vibrates when it's assembled it's mathematically difficult to get from the 'free' to the 'bound' plates, and there are lots of extraneous variables that must also come into play. It might be possible to do some meaningful correlation studies if you could build a few hundred similar instruments with good quality control, and do a lot of tests along the way. No hand maker would be able to approach a valid sample in a lifetime's production, and the factories that might be able to have no particular reason to be interested.

As for Rob's using the double -X pattern: he's doing it because that's what I do, and I do it because I like the way it sounds, and find it easier to 'tune' than standard bracing with more or less parallel tone bars.

Alan,

I appreciate your response as well as the honesty. You said that the ring mode seems to work is the perfect answer to the question. I'm sure there are some folks here who are more experienced than I am, and would prefer a very detailed answer as to why the ring shape produces the type of sound you are going for, but since I'm just starting building I'm perfectly happy to take your word for it.

Your answer as to why you like the double X also makes sense to me. Since there aren't a set of tone bars slanting across the lower bout in one direction, but rather two identical sets evenly spaced, I'm sure that would have an effect on the tuning process.

Thanks again! I'm learning a lot here!