Official Luthiers Forum!

I've been looking at that a bit but for different reasons. I used to do a simple deflection test of the finished top by placing a weight at the bridge area and measuring how much the top sinks. I found out pretty quick that it doesn't really correlate to what happens under string tension.

After the guitars were finished, I went back and checked deflection with no tension and then strung up to pitch and checked again, a couple inches in front and a couple behind the bridge. Two tops that deflected the same with the downward test deflected very differently with the string tension test. One guitar moved mostly by rising behind the bridge, the other sank some in front and rose some in back. The effect on saddle height would have been enough to notice. Since I'm using an adjustable neck, that wasn't what I was looking at though.

Point is that I think you have to simulate string tension to really know. Drilling out the two outside bridge pin holes and bolting on a temporary bridge and stringing it up was the best idea I could come up with for doing that. Let me know if you come up with a better solution.

_________________
http://www.chassonguitars.com

How are you measuring the tops rise ? I use a straight edge across the top an measure off the edge. I do use a 28 foot radius and this usually has me about a 1/16 to 1/8 inch off the edge before stringing . I will see this rise about 1/16 from that. Is this about what you see ? This is with a 5/16 X brace.

_________________
John Hall
blues creek guitars
Authorized CF Martin Repair
Co President of ASIA
You Don't know what you don't know until you know it

Question; Do you'all track top rise after stringing-Redwood?

For OM size body, I will have approx 3/16" arch prior to stringing at edge of soundboard. Tension will lift approx. 1/16" or slightly more behind the bridge for a total 1/4" at edge of soundboard.

I will check parlor tonight, but it seems less (reflected light test).

Wow, Todd, you are a control freak, aren't you!

_________________
Waddy

Photobucket Build Album Library

Sound Clips of most of my guitars

Hi Todd, I guess the inevitable question from a newbie, is... What would you normally expect from your data, and what would you consider a reasonable amount? Next up would be, what RH are you building at and are you stringing up in the same shop... Its just very interesting. I have a build that did not seem to move too much at all for about 3 -4 months after stringing up (probably did but did not notice) despite being relatively lightly braced, yet during the recent snow we had and the reslutant fall in RH I have seen the top flatten out noticeably... despite the RH still being around 50-55% ? yet built at a constant 40-45% ....confusing data? (A period that also saw some figureboard shrinkage on an expensive 'factory/boutique' instrument)

Needless to say, I may need a humidifier at some point and and keeping a close eye on things... but any advice greatly appreciated..

Todd, a while back I was playing around with Finite Element Analysis on one of my guitar tops and I pretty well convinced myself that the bridge doesn't really "pull up" much.
I was noticing in one of my FEA runs that the bridge didn't really move, just a few thousandths or so. I thought this can't be right, we all know that the bridge pulls up and we take that into account when setting the neck angle. So I made myself a little jig that rest on the sides of the guitar and holds a dial indicator as an attemp to veryify my deflection model, other places also in addition to the bridge.

I use this jig by slacking the strings all the way loose, zero out the indicator, return the guitar to pitch and take my reading.
I found that the bridge really does stay pretty stationary as far as "pull up" goes. It moved just a few thosandths like the model predicted.
All that said, things definately are moving around under string tension, just not the way I always thought they did. And doing what we do to set the neck angle is still right (obviously), it's just that more of the guitar is in the game than I thought anyway.

Here's a shot of the deflections on one of my runs. As a quick topo map guide the darker the blue, the less the deflection and as it approches red, the greater the deflection. For all practical purposes, the darker blues are basically zero deflection. As a technical note: the top is constrained around the perimeter and the load is applied to the saddle.



I thought you might find this interesting, and you may already have an indicator jig like I show, but if you don't, you can learn a lot about tops with one.

Jim

_________________
Jim Watts
http://jameswattsguitars.com

Well I'm interested. So what's moving to account for the rest of the "deflection" that people are seeing?

John

Well, I'm interested too.

Jim, are you using static loads? Is there any intent (can your model support?) to model the effects of a vibrating string?

_________________
Steve Smith
"Music is what feelings sound like"

Very cool run!

As a mechanical engineer that uses ANSYS a bit, I have a few questions about the boundary conditions you set up for you finite element model.
1) are braces applied to the underside of the model?
2) how did you account for the anisotropic nature of wood? as well as grain orientation for the braces and or graphs.
3) Does your loading scheme create a contact surface with the saddle? thus appling a load down through the top or is it only pulling on the bridge?

Also i wouldn't just assume its correct because you measured a top and it was a close match to the model... I've learned from experence that two wrongs can make (what you think) is a right. (this is in reference to using it to gleam information about other guitars from the model)
Not trying to nit pick or any thing, i just love discussing these types of problems with others =) I learn so much that way!

I don't do FEA, but I've done some measurements of deflection at the saddle with a jig very similar to Jim's, and got the same results. I've assumed the changes that happen to neck projection when the guitar is strung up must be mostly due to neck rotation. I would assume the back is flattening out a little, and the sides distorting some to allow this, but I haven't measured that.

Frank, I'm pretty sure you need a better humidity gauge. When the action suddenly drops on one guitar, and the frets start sticking out on another, things have gotten way too dry regardless of what your gauge says. Get some more water in the air quickly while you shop for a new gauge.

It's just a static model, I used Cosmos to do this. I have done dynamic modal analysis in the past to do “thought experiments” on bracing patterns, but found it hard to correlate the data to the actual instrument.

This model is an assembly of different parts, and as such I can tailor the properties of each part in the run.
Boundry conditions: The perimeter of the top is fixed representing the top being glued to the rims. Each piece is bonded to its mating part in the run.

braces: the top is braced.

Load: Strings are bonded to the bridge and at the saddle contact point. A load in tension is given to each string. I used the numbers from Daddario's website.

Wood properties: I measured the wood properties in two axis of the particular top and braces used in this build and used that data in the Cosmos run. Cosmos allows for anisotropic materials.

I measured the actual instrument with the indicator set up shown above and the readings were with 10% of my model. I thought that was pretty good considering it’s a bunch of wood glued together. I think if you use an indicator on one of your own guitars in the way I described above you’ll find a similar pattern.

If you think about it for a minute it makes sense, the saddle is the pivot point being pulled up from behind and pushed down in front from the pressure on the saddle. I think the whole sound box and neck are rotating towards one another accounting for the pull up that we see.

_________________
Jim Watts
http://jameswattsguitars.com

I missed Erics post prior to me posting mine. It's nice to see some more empirical data backing up the general pattern of my model. Thanks for posting that Eric.

_________________
Jim Watts
http://jameswattsguitars.com

I got similiar results with my model, which has less refined elements than Jim's model. It wouldnt' take much tweaking of geometry and materials to get extremely close to the same results.

One difference with my results is that it shows a pull at upper fingerbraces. I could see the same pull on a guitar that I had handy at the moment. (although perhaps a bit less than the model shows.)

Displacement data.



What the model looks like.



Thanks for posting your results, Jim. It saved me a lot of time trying to figure out if I should believe what my model was showing.

John

Nice John,
what were your loading conditions?

_________________
Jim Watts
http://jameswattsguitars.com

I'm using 180lbs for total string tension.

The way I see it:

The string tension applies forces at two places on the top. The string anchor point and the saddle.

1) At the anchor point there's an upward force and also a force parallel to the top towards the neck. I've been using something like a 27 degree break angle (angle of string to top) to resolve the forces.

2) At the top of the saddle there's a force parallel to the top and towards the neck (the obvious string pull force) but much of that is balanced by the string tension from the anchor point to the saddle.

The string tension from the anchor point to the saddle also puts a downward force on the bridge.

I haven't dialed in neck angle as I don't think that I have things precise enough for that to matter much.

Here's a picture of how these forces resolve when they are applied to my model.




I've played around with my model some more and it doesn't take much tweaking to get under .004 max displacement and when I did that I got rid of a lot of the rise in the forward finger braces.

John

Cool, thanks John.

_________________
Jim Watts
http://jameswattsguitars.com