Official Luthiers Forum!

This is one of those questions that I want to ask to confirm I’m doing something right before I do it again, and potentially do it wrong.

When I use a radius disk to sand down my rims and blocks / Kerfing and such, where should the center of the dish be in relation to the guitar.

I’ve been putting it in the “center” of the body, but is this right?
Should it be center of sound hole?
Should it be more in the lower bout?

Why am I overthinking this.
Maybe too much early coffee

B


Sent from my iPhone using Tapatalk Pro

It doesn't matter because it's part of a sphere and all the same. You can send in a circular motion, side to side, back and forth, etc

_________________
Pat

Ok so help me think through this...

The edges are lower
The middle is higher (dome)

The waist is always the lowest point so it is the last area to get hit in the process.
If the dish wasn’t in the right place it seems that this geometry would be thrown off


Sent from my iPhone using Tapatalk Pro

when i do a radiused top i'll center on the bridge. i have a fixture which holds the body in place and the dish revolves on a bearing around a movable and inclinable post. i'll move and tip the post forward or backwards as needed to create a shallow slope to aid in setting neck angle.

The waist is always the highest point so it is the last area to get hit with the dish.

Assuming you were starting from a flat rim. Which is what I do on a top rim. The back rim is usually profiled first, so there are variables involved

The radius dish is domed, assuming you're not using a cylindrical shape. So it is oriented exactly the same no matter which part of the dish you are using.

_________________
Pat

"The waist is always the lowest point so it is the last area to get hit in the process"
If the disc is a sphere, it will be the last place to make contact no matter what part of the sphere it is on. Draw a half circle, and then place a short straight edge whose ends are touching any two points along it. The perpendicular bisector length from the straight edge to the circle will be the same.


My radius dish is a homemade thing - a thin piece of plywood mounted on 3/4 inch plywood, tacked down in the middle and propped up along the edge. It is more of a spline curve, so I try to center the guitar on the disc. Sometimes I use a Solera, and it makes a "hump" that is even a bit more wonky, but a flat rim will work for it. And then there is the cylindrical arching I use on some other instruments....
I guess what I'm trying to say is that it isn't critical how you arch the soundboard (or even if you arch it) as long as you have worked out the other elements (neck geometry, brace shaping, etc.) that are affected by it.
Some builders use the spherical radius dish and then intentionally "flatten" the rim of the upper bout.
When using a powered radius dish I mostly check to make sure both sides measure the same at the same points. I try not to build "wedge guitars" unintentionally.

For the cause of clarity.
I’m also using a method where I radius the thing and then flatten the upper bout.
Always followed the way John Hall demonstrates in his videos


Sent from my iPhone using Tapatalk Pro

I center mine about an inch behind the waist-towards the tailblock..

FWIW I have a large platen with an aluminum shaft that fits through the hole I drilled in the dish. The shaft is theoretically where the X braces cross when the rims are properly loaded. I guess I am somewhere between Brad and Arie and there is substantial slop so it doesn't bind.

As long as it's a true radiused dish then it doesn't matter what the position is. What does matter, a lot, is that you don't sand more on one side or end than the other. I put pencil marks on the tops of the linings and the blocks so I know when I can stop sanding.

_________________
Steve Smith
"Music is what feelings sound like"

I profile my sides prior to bending, the last area to get hit used to be the front edge of the heel block. Now I profile that as well.

When I was wrapping my head around a true radius block and whether location would matter, I realized what has been stated, it doesn't matter.
I imagined myself inside of a 15' radius sphere, and what would happen if I "sanded" a body on the inside of a sphere, it doesn't matter where I am inside the sphere, its all the same.

Anywhere you choose to sand in the dish seems fine to me.

How I do it.
On the top I mark the sides and sand away until I have a continuous rim to attach to. Then I clamp the braced top down on the top and lay a straight edge firm against the top from the neckblock to the soundhole. Then I measure the clearance from the straight edge to the saddle position on the top. I'm looking for proper clearance over my bridge once I add a fretboard w/frets and bridge. I like something like 1/64". Then I take a flat sanding block made from melamine (2'x1') and sand the area from the soundhole to the neckblock until I have a smooth transition and the clearance I need. (I usually end up with a slight drop off of around 1/32")

For the back, since I have profiled so the body thickness goes from the tailblock to about 1" less at the neckblock (see sketch) I sand the radius into the back in two stages on the disk. First the lower bout and then I start pushing the upper bout into the dish making a nice transition and ending up with the upper bout firm against the dish.

_________________
Joe Beaver
Maker of Sawdust

Are you saying that when you're done, the radius dish won't touch the whole perimeter of the back Edge at once?

_________________
Pat

Pat, Yes, that's correct. The dish is used to shape the back but not as a unit. I do it in two stages. Actually when the top is done it doesn't fit that well into the dish either since I use a none radiused top under the fingerboard extension.

I'm not saying my way is better but it is what works for me.

_________________
Joe Beaver
Maker of Sawdust

We radius the back of the rim with a 15' dish...the waist area initially comes down, then the end blocks make contact. We use some pencil or chalk witness marks on the block to show where material is being removed, and measure at the waist side-to-side to keep things even. The General brand layout Pencil is available in most Michael's and other stores carrying art supplies, and works well on lighter woods with nice, heavy, visible lines, while standard white chalk or the big sidewalk art sticks do the job for rosewoods and ebonies (colored chalks use dyes that make the material more durable...not what is wanted on a guitar...white only!).

General's Layout Pencil: https://www.amazon.com/Generals-Layout-Pencil-Graphite-555/dp/B002Y2KVGA

Unlike Mr. Beaver's approach, we finish both back and top in the 15' and 28' dishes, respectively, with the adjustment for the fretboard extension area angle of the body done by bracing with 28' radius X, tone, fingers, and bridge plate glued up in the 28' radius dish and a 60' radius UTB glued up on a flat caul (along with the UTG if used). Both the top and the back of the rim fit perfectly in the dish when done (this approach works for Size 5 up through the largest acoustic bodies we build).

For those cutting the relief angle into the top rim (versus building it into the top), the additional step of routing or sanding the upper bout is necessary and accomplishes the same result as our 60' UTB, etc. Multiple paths to the same result.

_________________
For the times they are a changin'

- Bob Dylan

" what does matter a lot is that you don't sand more on one side or end than the other "

That is exactly why I have tried to center my dish over what I want to be my high spot. You are really sanding an arc and it only stays in the same plane as the sphere if it has the same center line and radius.

Probably overkill but I put a lot of things into space when I had my business.
There are lots of ways to work around it. I try to make it as simple as I can as long I am going to make a fixture

Tim

We sand the rim mounted in the outside mold with dish firmly mounted to a surface and the movement a reciprocating one. The weight of the mold and clamps does most of the work, and the ability to control where the cutting action occurs with a bit of newsprint makes the process a precise one.

We could use a shop-sized radius dish and work the rim in one small corner of the main room...other than the weight of the mold and clamps favoring the 'downhill' side (reverse orientation and repeat to balance), there is no difference in the geometric sense. Or we could build a full sphere of the correct radius and pick any point on the inner surface of the sphere...the radius sanded will not vary.

To understand why, imagine walking on the inside of a large sphere directly between two points...the path taken - if the shortest between the points - is a great circle, and every great circle path is a section of a sphere that contains a diameter of that sphere. If any two unique points on the surface of the sphere will generate a great circle - including the antipodals, which generate an infinite number of great circles - then every set of unique points on every sphere or section of a sphere (e.g., a radius dish) will do so as well. Thus, any point on a radius dish may be selected as a 'center point', with the practical limitation that the guitar shape is best contained within the boundary of the dish for most consistent sanding action (although carrying the rim off the dish 2-3 inches does not do any apparent harm). The two points - the selected center and every other point on the dish - will always describe a section of a sphere containing the diameter of that sphere.

Which is to say, it does not matter what portion of the dish is used, as every portion will generate the same shape in use.

_________________
For the times they are a changin'

- Bob Dylan

Hey Woodie, no disrespect, but that's not the way I learned it or made and measured things on a CMM using Geometric Tolerancing.

Regardless I think what you are trying to say and what you don't acknowledge is that every point on your great circle originates from the same center point. You can have different points all around your great circle.. BUT..The minute you change the center point within your given sphere you have created a new sphere and it is no longer on the same plane as your original. The distance from your new center point to the edge of the original sphere will be different lengths OR your sides will be different heights.

Reality is that we are sanding roughly 1/2" on the perimeter of an odd shape. The tool cuts a spherical shape. If you move either the workpiece or tool, in a simple linear motion relative to each other, the contact points change resulting in sides that will be uneven in width. The shape of the spherical radius will be the same because that is determined by the tool, but the workpiece will vary accordingly. If that matters. Some people are still using planes to put the pre glue angle on and some are using a linear radius.

A great circle is defined by ANY two points selected on a spherical surface - there is no need for a center point, although I can envision how tooling design might force the issue where you require rotation around a fixed point. Think about moving the work piece versus the tool, and substitute linear motion versus 'driving the bus' - there's no need to define a center point because there is no center of rotation or plane of rotation to accommodate.

Sure - fix the dish in plane, rotate it for cutting action, and fix the rim...shifting the rim results in an error because the rim and tool are locked to a common vertical axis. Move away from the production line tooling mentality , and move the rim versus the tooling (or rotate the dish and allow the rim to float), and any path across the dish always results in the same radius being milled into the rim because any point on the dish is joined to every other point on the dish surface by a great circle path as defined by the radius of the sphere from which the dish is drawn.

_________________
For the times they are a changin'

- Bob Dylan

You can wax poetic all you want about geometry , however-wood moves so what you start out with might not be exactly what you end up with......

There is a difference between a minor arc and a great sphere. Any 2 points on the surface of a sphere can be used to construct an unique great circle, although that does not necessarily mean the same great circle we are working with. Sounds like you may be referring more to the great circle used in navigation than engineering. They are a little different. The great circle in engineering is constructed by a plane that always travels through the center point and the antipodals, which are the ends of the constructed diameter. In navigation it is used to determine the shortest distance between any 2 points on the surface of a sphere. You can't go in a straight line because it's a mass, although I think they it once in the movies.

The only radius I really want right is the one on the top plate, which is why I do mine the way I do. If yours works for you, that's good. Way I'm doing it doesn't take long to drive the bus and I know everyone will arrive safe.

The navigational version of a great circle (a great circle sailing) has to deal with the fact that the Earth is an oblate spheroid; I borrowed the boss's copy of Bowditch and quickly gave up making sense of the corrections to be applied due to our tired, old planet's midriff budge problem. We probably don't want to get into great elliptic and geodesic sailing, but they are apparently a thing. I prefer my water viewed from a well-equipped, sandy beach or frozen and bobbing in a well-made drink, but Bowditch (the boss refers to it as 'THE Bowditch'...shades of older Gibsons) is fascinating reading, if a little dry.

A great circle in spherical geometry (and I suspect in the descriptive geometry sense we are employing here) has a much tidier definition, which results in a path of shortest length between any two unique points on the surface of a sphere or section thereof possessing the same radius as the sphere. I don't believe that point is in dispute, so much as the practical consideration of how that spherical surface might be used to shape the edges of the rim.

The antipodal points only come into play when we lock into what seems to me to be a least time/production line mindset that a) we must designate a fixed axis through the centroid of the tool and normal to the local surface at that designated position (in this case, the radius dish), b) we intend to rotate the tool or the work about this axis to mill a matching radius on the work, c) the work remains locked to the same axis through a designated center point, and d) we must minimize the role of skill in the outcome.

If we dispense with the notion that we need to lock the tooling to the work along a common axis, things get simpler in a small shop...the tool gets clamped to a flat surface, the work can be both translated and rotated on the surface of the tool, and the taper of the body ceases to be a factor in ensuring an adequately accurate radius on both back and top edges of the rim. The two figures below borrowed from the drafts in the builders guide illustrate the shift in paradigm.





As to accuracy, the surface of any radius dish will depart from a perfect spherical section by some amount...MDF moves, machining cells are not infinitely precise nor accurate, the rim deflects (particularly if not held by an outside mold and spreaders), the movement of the craftsman is unique in every iteration, and all the while, shop RH and temperature are varying. With that said, none of these cumulative or offsetting sources of error compel us to surrender to the notion that we should not at least select a target value and some reasonable bound for those errors in our processes. It's not difficult to imagine an aggregate error of 1/4" or so in the milled radius...but for the top, which is arguably where we desire the greatest accuracy, that error still represents under 1/10 of a percent variation from a 28' spec value. As Mr. Goodman suggested, the variation due to wood movement quickly dwarfs any error in our tooling or processes, so we handle that by shooting for a target value for radius at a designated RH, knowing that the instrument will both function and survive in a range above or below that set point.

_________________
For the times they are a changin'

- Bob Dylan

So, it does not matter where the radius dish is centered from a geometrical standpoint. For practical purposes if the rim is stationary and you are driving the bus, it makes sense for it to be centered to try to even the pressure that you are applying.

Having the dish mounted on a shaft, stabilizing its rotational axcess but still being able to pivot is an upgrade I've always wanted to try as it does take some effort to not only drive the bus but to keep the dish centered over the workpiece. (Although it's probably more of a CrossFit work out then stationary machine) I typically use a combination of a clamped dish with reciprocating motion of the rim/mold and driving the bus varying my orientation on both trying to average out my errors.

_________________
Pat

It does not matter where on the dish you center, and it does not matter if you spin the dish or rub it back and forth and side to side. What matters is that you hit your mark on the 2 end blocks and that both sides are the same. If you have those points, then the shape has to be correct.