Top and back radius

[James Merideth]

I am a first time builder considering a steel string build. I have a number of books on the subject and have a question regarding the accomplishment of the front and back radius. The Kinkead book suggests preparing a separate board with concave surface for each piece and gives contours for various cross sections of the two boards. Other sources suggest using a radiused sanding board (available commercially or build your own) to prepare radiused braces and then conform the front and back to these braces using vacuum clamping or perhaps a board with cork shim around the edge of the front or back board and clamps or go-bar press. The two processes will give somewhat different results i would think, since the shaped molds from Kinkead's book are not perfect constant radiuses. I'm inclined to go with the radiused sanding board to prepare the braces and then vacuum press with the flat side of the braces on a flat surface and the bag pressing the back or sound board onto the radiused braces. Any advice from you experienced builders?

[Bob Gramann]

No matter what method you use, what you want in the end are perfect joints between the braces and the plates and between the plates and the rim. Any gaps in the glue joint are invitations to failure later. I use spherical dishes, one commercial, one self-made. I recommend the commercial. Amortize them over a few guitars and they don't cost much. And, you save a lot of time and miss a whole lot of dust that you would otherwise experience in the making. The dish not only helps you shape and glue the braces but also helps you to prepare the rim surfaces for gluing. I have vacuum tools but I use gobars for gluing my braces (I use the vacuum for gluing bridges, holding things while I'm working, and occasional weird stuff). And, for what it's worth, I use a plane to initially shape the brace radius and finalize it in the dish.

While building my first guitar, I used clamps to glue the plates to the braces, one brace at a time. That worked fine, but graduating to dishes was a big improvement.

[James Merideth]

Thanks Bob. Any thoughts on best radiuses. Kinkead uses about 19' on the back and 28' or so on the front. I've seen other recommendations for similar numbers, but maybe a bit less on the front, maybe 25'. Also, I assume you mean by "preparing the rim surfaces for glueing" that you use the sanding boards to sand the edges of the sides after the kerfings are in place.

[Rodger Knox]

One thing you need to consider in conjunction with the top radius is the neck angle, and how the flat fingerboard fits onto the curved top.
There's a number of ways to flatten the top where the fingerboard sits on top of it. It's a minor thing, but it does have to be worked out, and it's easier in the planning stage.
There's also several ways to set the neck angle, the important thing is the string height above the top at the bridge, and that needs to be about 1/2".
You can draw all this to scale and get a pretty good idea of how it fits together, but you have to remember the guitar will distort some under string tension. The bridge will rotate forwards, the top will pull up, and the neck will bow. All this adds up to less than 1/8" depending on how stiff everything is, but your tolerance for string height at the bridge should be plus or minus 1/16", so you have to allow some for distortion.

[Bob Gramann]

I use 25' for the front and 15' for the back. A 25' radius usually makes the fingerboard hit the top just right when the neck angle is set for a half inch string clearance off the top at the bridge. Yes, I mean using the dish to sand the kerfed linings and blocks prior to gluing. I also sand the rims before I glue the linings to them. It makes the whole job easier for me.

[Mark Fogleman]

Like Bob, I use radius dishes, top is 25" and back is 18". They were made like this but bigger: http://forum.ukuleleunderground.com/sho ... adius-Dish and covered it with 80 grit sandpaper. The braces are contoured in the dish and get glued to the plates with Go-Bars forcing the plate to conform to the radius. Next process using the dish is to contour the rims to accept the domed plates. John Hall at Blues Creek calls this driving the bus because he inverts the sandpapered dish over the rim (in the mold) and turns the dish back and forth.
Excellent information here by Jon Sevy: http://jsevy.com/luthierie/workboards/A ... oards.html
and here by Liutaio Mottola: http://www.liutaiomottola.com/formulae/sag.htm (be sure to convert all distances to inches or MMs when completing the worksheet)

[James Merideth]

Thanks for the input guys.

[Ed Haney]

I use radius dome dishes. 15 ft. for the back and 25 ft. for the top. The braces' bottom's are shaped to match those and I use the a go bar deck to glue the braces while the plates are in the domes. I then use the domes to radius the rims (sides & linings). This seems to be a very common approach.

The interesting thing is, and I've never seen this mentioned anywhere, is that there is no way that the rims and plates could match "perfectly". Why? Because after taking the braced plates out of the domes, there is ALWAYS spring back since the plates want to return to their original flat state. They can't go flat again because of the braces, but they become less than the radius of the domes. Also, if the back is ladder braced it tends to spring back more into a cylinder shape than a dome. That means that both the braced back and the braced top have less radius than the domes, and in reality they have different shapes. Yet the rims have the same radius and shape as the domes. Of course, regardless, I glue on the plates (as most other people are doing) and everything seems to work fine. But my point is, there is no "perfect" method for "perfect" joints that I have seen described by anybody when using a radius dish.

[James Merideth]

Ed,
Thanks for your thoughts. That all makes sense to me. I do a lot of woodworking, mostly furniture, and nothing so delicate as a guitar, but I surely know that perfect doesn't often happen and close enough is often pretty good. I looked at the sites that Mark furnished and on one of them, the guy actually talked about making a "dish" that was actually only curved in one plane, more cylindrical rather than spherical, and used that for backs, which, I guess, is just accepting what you are saying typically happens with the back anyway. Might be interesting to try to figure out how much spring back you actually get on the sound board and the make a dish of that radius to use for sanding the sides before fitting the front on, but I've never seen anywhere where anyone seems to think that level of finesse is needed to get the job done. Maybe pressing the fronts and backs on to the sides, where not a perfect fit, actually adds some additional tension to the construction and actually helps the final result musically. Way beyond my level to even think about that in any case.
Thanks again,
Jim

[Bob Gramann]

When you glue the sides to the plates using the dishes to support the plates (one at a time, please) it pushes the plates back to the shape they were in when you glued the braces to them in the dishes. If you sanded the rim surfaces in the dishes and then use the dish is a gluing caul, you're getting a joint as good as can be made. I clamp it dry and check for gaps with a flashlight inside the guitar pointing out. I don't glue until I get everything fit just right. After the glue dries, when I take the clamps off, the guitar body sits in the dish like it fits perfectly. There's no springback at that point.

[Ed Haney]

Bob,
I understand how to glue one plate to the rim while the plate is in the dish. But how do you glue the second plate? If you glue the top on first, then you could put a couple of go-bars in the sound hole to hold down a part of the back while gluing, but its hard to see how that could hold the entire back down flush in the dome.

Also, when gluing the first plate with go-bars (I assume several) inside the rim holding the plate down into the dish, what clamping/caul set-up do you use to clamp the rim to the plate?

Do you have any concern that the top and back will be in tension putiing pressure on the sides as it tries to springback?

Thanks for your input.

[Todd Stock]

I build with 15' back radius and 28' top radius, except for the UTB, which is radiused to 60' and is glued to the top on a flat caul...works for anything from a Size 5 up to an ABG for correct fretboard extension fit.

As to glue-up...rim in mold goes top edge down on 28' dish, back is glued on (with hot hide) using a 6mm textured-one-side fiberboard caul, then after 30 minutes of drying time, the bars are removed, 15' dish swapped for the 28', glue squeeze out cleaned up inside, and the top goes on with same caul/go-bar deck system. Other folks use other systems that all seem to work pretty well, but this one works well and minimizes the investment in clamps (still have too many, but maybe 1/3 of what I would need to close box with cam clamps).



Caul goes rough side up to minimize tendency for things to move around as glue tacks...also remember that the line of action on a go-bar is the line drawn between the contact surfaces of the tips...keep things vertical and nothing will slide around...even when using Tite-Slip....Slip...Slip...Bond

[Bob Gramann]

It doesn't have to be hard when everything fits. These are way old photos but they're what I had. I no longer use the foam (it was to take up irregularities in the dish--I have better dishes). The screws on the mold were for rubber bands. I no longer use those either. The plates are glued on with hot hide glue. As you see, the rims are stiff enough that clamping is needed only at a few key points. If I have an area where there is a small gap that can't be closed with a clamp, I use masking tape on the other side of the plate to bring the plate up to the rim.

[Ed Haney]

Bob,
My method is basically like Todd's above, except I have used Titebond so far (I'm using hot hide glue in other joints and improving my skill with it at this time.)

In your pictures I do not see any go-bars or other clamps or weights to hold the top's surface down into the dome as the gluing takes place. Without something to force the top (and back) down flully into the dome their springback will change the angle out at their outter edge (i.e. they will have less radius). I don't think that 4 cam clamps (or even more) pushing the small thin surface area of the rim down can force the top itself (or back) down fully into the dome. Therefore, the top's edge angle is slightly less than the rim's edge angle (your 25' dome angle from sanding would be true on the rim edge).

Please don't take anything I've said as being harsh or combative. I believe what you are doing works fine is as good or better than what I (and many others) are doing.

My point was just one of interest. Namely, I don't see how, using the domes and methods we are using, the plates' shapes and rim's shape are a true match up. The plates will have less angle than the rim because of their springback. Nevertheless, there is enough contact surface area that it still works fine for all of us using this method. When I put a 25 ft radius caul (same 25 ft as my dome) on my top and compare it to the final radius of the top, the top's radius is noticably less than the 25 ft caul. Have you done this? It seems to me that your top's radius will be less than 25 ft too (and back's less than 15 ft).

Thanks for posting the pictures and taking the time to communicate. I do appreciate it and I do not intend any criticism.

[Bob Gramann]

Clamp yours up dry and see what you get.

[Todd Stock]

Clamp pressure propagates in roughly a 45 degree (from line of application, so 90 degrees total) cone from point of application - Bob's clamps provide pretty decent overlaps on the upper bout, and assuming each clamp provides a realistic 200 lbs each of clamping pressure, that's 800 lbs on a glue line that has about 22 square inches of area, or 36 psi....fine for TB, as even 10 PSI on a well fitted joint is enough. The 70+ go bars on the body shown should give about 560 lbs, so the pressure is less, but a little more consistent at the glue line.

Weight of Bob's mold and the rim creates some clamping pressure as well, but I can see the argument for adding a few more clamps in the area where there is no direct effect from the clamps, but not sure it matters ... the neck and tail block get clamped pretty well either way, and we cut away a lot of the glued surface on the plate and rim otherwise, then purf and bind...not sure it really matters all that much provided the fit between the rim and the plates is good, and with a dish, it usually is decent.

[Todd Stock]

Ed:

My plates tend to stay put at about plate radius, which may be a slight tendency to over-radius the braces or just good luck. I radius the bridge plate and tend to let the plate rest in the dish for a day or so, but have not seen much in the way of a tendency to de-dome unless it's related to stress relief during carving (which means the top or back brace gets stripped out and a new one glued in.

[Bob Gramann]

And with hot hide glue, you really only need enough pressure to make sure that joints touch everywhere with no gaps. The heel and tail block areas require the most pressure because there is more deformation of the plates there as they bend to match the longitudinal radius. The guitar in the picture was numbered somewhere in the 40's and I had been doing it that way for quite a while before the picture. I finished 99 a couple of weeks ago. I'm pretty happy with the method.

After I glue braces, there's always some springback when I remove the back from the gobars because there's nothing in the braces to hold the curve along the length. The top pretty much just sits there when I remove the gobars. I do radius the bridge plate. Sometimes I leave a flat in the middle of the upper transverse brace, sometimes I just sand the gluing surface for the underside of the fingerboard flat. At that location, it doesn't take much sanding to make a flat surface the size of a fingerboard tongue and my bracing underneath at that location is pretty stout.

[Alan Carruth]

Once the sides are profiled to the top and back radii, I fit the plates and use a trimmer bit to take them down the the correct shape with the sides in the outside mold. I remove the sides from the mold at that point and use spool clamps to glue on the top and back. If you ensure that the plates line up all around when you glue things up then the sides will be in the same shape as they were in the mold; perpendicular all around. This works pretty well. Spool clamps are cheap: you can make up a box in an afternoon from thread rod, dowel, cork and wing nuts.