Acoustic archtop non-traditional design and build

[Nick Kitchener]

After researching around 85-98 sounds like an option. Former being more focused/tighter bass and the latter more open bass/balanced.

I suspect this will actually depend heavily on how the top behaves.

[Alan Carruth]

The whole point of an archtop, as I see it, is that you get stiffness from the arch shape and graduation. The bracing on the top, in my experience, is mostly there to replace the stiffness lost when you cut the sound holes. No bracing is needed on the back. Adding a bunch of braces based on flat top construction is a 'belt and suspenders' approach that is unnecessary, IMO. The only archtops that use a lot of braces are the ones with formed plywood tops, and they're not noted for bass response. Or sound....

If you're contemplating notably high tension, then you'd start by using thicker graduations, and raise the arch height to suit.

A friend of mine saw one of Freddie Green's guitars, a 19" wide Stromberg, at a vintage show. It was set up the way Freddie used it: a .07" low E string 1/4" off the 12th fret. When he managed to actually finger a chord and strum it, everybody in the ball room jumped. He didn't notice any structural issues...

[Nick Kitchener]

Thank you - food for thought and it makes sense.

I think (amongst tackling the mould, cradles etc) I think it's now time to really nail down the profile of the plates, then follow the wood.

[Nick Kitchener]

I've decided to make an armature for testing the top. It's a piece of wood that has a pivot on one end (bolt) and 600mm along has a weight (over the tail of the piece under test. The bridge point is then at 350mm. Naturally this magnifies the force so only 350/600 of the down force is needed in weight - so between 5-14Kg gives the option to test ~20-53lbs.

I'm currently look at parabolic arches and bending moments. I recognise the complexity of a 3D surface so the idea is to use the contraption described above to test as the carve and thicknessing progresses. The armature can then show if the top depresses (I can also use the indicator gauge for this at the same time at points across the top).

[Alan Carruth]

That's a fairly common test rig among archtop mandolin makers.

[Nick Kitchener]

I do have a couple of random questions I feel I need to ask:

1. If a parabolic load pushes the arch outwards, does it make sense to have the top with a small underside recess that interlocks with the side lining? (or is the glue able to cope with the sheer stress?) That way the down pressure of the lower arch is transferred out towards but encounters the band of the lining, spreading the load. I can see bracing supporting both the down pressure and the side push out. However I get the concept of using a higher arch to spread the load - in a guitar that would mean that the steepest part of the arch (ie closest side/recurve to the bridge) would carry the most down force into the guitar. A notch would then create a compression joint between the top and the lining. Possibly a pain to break if a repair is required.

2. Shellac and titebond/hide glue - I've read that the alcohol can cause titebond's joints to weaken/break. Longer term the cold titebond liquid hide glue is not recommended - for the same reason joint failure is likely longer term. My thinking here is - for the bond of the plates to the side this could be a problem no? (am I over analysing?). Is it worth the learning curve of hot hide glue over titebond original for what is likely to be one off guitar?

3. Finish - I'm thinking waxed shellac (darker buttons) will offer a better resistance for sweat but reduce the clarity of the finish, vs, using dewaxed (clear/blonde flakes) then furniture/car wax over the top to improve sweat resistance with a better clarity of finish. I'm not trying to hide the grain whilst at the same time not have the finish immediately tarnish (thinking a debuff on string change). I've used danish oil before because it's easy to fix when iterating through modifications.

The 7 string pickup is, as you'd expect, either a compromise with a larger/deeper pickup (like the SD Jazz size etc) or custom. I emailed Armstrong's UK dealer and they suggested contacting Armstrong for a custom build. An alternative is go down the DIY build with that too. However I feel that's a question for another day.

[Alan Carruth]

I'm not sure what you mean by a 'parabolic' arch. No recurve? The curtate cycloid arch has the recurve built in, so if you do it right it minimizes the 'spread' problem and also avoids sinking at the edge.

Again, the arch height is related to the thickness of the top. Decide how thick you need the top to be, and scale the height from there. Violins usually use a top thickness of ~3mm or so, and the arch rises ~12 mm from the upper surface at the low point of the edge (just inside the purfling) to the base of the bridge on the centerline. I've simply used a linear scaling rule, so a 4mm thick top would see about 15mm rise, and so on. That's an average thickness... There's no penalty for making the top thicker at a given arch height, except the added weight, of course.

With too much arch height on a thin top the edge buckles in odd ways when you look at the 'free' plate modes. Instead of a single 'closed' ring-type mode you can end up with two that have openings, either at the corners (shoulders and hips) or at the ends and the waist. The timbre gets 'tinny' on the assembled instrument.

I've never had problems with shellac and hide glue. I have avoided liquid hide glues myself. I do use liquid fish glue, but not for structural joints. I've never heard of anybody having issues with alcohol and regular Titebond yellow glue; the only thing I've heard of that breaks that down is acetic acid. For what you're doing I'd go with regular Titebond.

I did try laminating tailpieces using Titebond, and they all failed eventually, but that's under much higher stress than you're contemplating.

With shellac my experience is that the more refined it is, the less durable. At one point I got some un-processed seedlac that as not too dark (I don't remember if it was 'kusmi' or 'elemi'). I used the 'teabag' method of dissolving it: put the shellac you want to use in a cloth bag (old T-short stock is good) and suspend it at the top of a jar of alcohol. The wax particles of the seedlac were too big to get through the weave, and all the dirt, bark, and bug parts also stayed in the bag. If you don't squeeze it you end up with clear shellac. If you learn to French polish you'll be able to touch up your shellac finishes fairly quickly.

Most of the pickups I've seen have been fairly shallow in themselves: it's the mounting hardware that can add depth.

It's easy to get hung up thinking about this stuff for sure. I usually find that when I get started most of the imagined difficulties disappear.

[Nick Kitchener]

Thank you alan,

On the curate cycoid, I can see two conceptual approaches:

* axial - the c.c curves across the guitar and a second “height” curve axially from tail to neck. This seems to be the tried and tested approach

* polar - the c.c curves could also be across the guitar through the bridge point, rotating at the bridge point.

Just seems if the bridge energy and down pressure is inverse radius squared, that a polar form would mirror the power/mass curve.
The stiffness would create a ellipse.

Just interested if anyone has considered this concept?

[Nick Kitchener]

I ache..

So that's the mould outside 90% done. Just need some final finishing, some edge straightening. I also need to glue them rather than screw them together. Next will be the inside jacks and the cradles, both of which need the finished mould to make a proper outline to mirror.

To the left is the pipe I'll use for bending. I've already tested it with a flame to see if it will rust/mark wood and that seems fine.

[Nick Kitchener]

Both sides are now bent. I will make a small mould adjustment for the acute tail bend - that will be smoothed a little, which isn't a surprise.

I had four starting grain breaks which I just added some superglue and after that they've been fine.

I'll sort out the tail and neck blocks next, along with the tail bend. After that it's time to join the two halves, then get onto the lining and side struts. I'll be happy at that point that it's less likely to split (although the wood has been well behaved so far).

[Alan Carruth]

I have had good luck in reinforcing sides with paper. You can glue it onto the outside of the bend with almost any sort of glue. Obviously, with something like Titebond, you wait for the glue to dry before bending the wood. Paper is amazingly strong, and does a good job of keeping the sides from splitting out. Of course, you'll need to scrape it off afterward....

[Nick Kitchener]

That's a great tip. So far I think it's been ok.

The lining is 4mm so it could be even more fun.

[Nick Kitchener]

The next stage will be slightly different to tradition.

Next and tail block prep but with the through spar that needs to be incorporated with some joint support to prevent the load applied to the tail breaking the joints or blocks. I know the body provides this traditionally with the bracing but I want the spar to be taking the majority of the load.

I'll then bring the sides together and start on the linings and side structs.

I'm considering putting small corner blocks into the tail bends (the acute bends) to provide some additional support for some protective perfling in that area. I'll be tempted to stand the guitar on the tail on that back edge so that perfling and the wood behind could do with supporting some knocks. I know that's not traditional treatment of guitars but 'incase'.

[Nick Kitchener]

Just got the next few bits - including the carbonfibre rods, the machine heads, etc.

As the weather here is sub-zero at the moment, I've been considering the next step of the neckblock/tailblock but before I did I thought I'd think through how the spar and neck would connect together (I hear dove tail but hear me out!).

The traditional approach would be to create the box of the body, adding a neckblock, tail block, closing up the box, then routing a V dovetail joint for the neck to create the required neck angle. The string tension is taken by the body, the neckblock and neck joint and the neck. So the neck joint being a dove tail works nicely and the length off the dove tail spreads the load of the force along it.

(laminate is hard maple, wenge, walnut, purple heart, walnut, wenge, hard maple)

I have a neck block made of a laminated offcut (see above) from the though neck of the solid body. I'll use the mid sized piece I can extend the contact surface of the dove tail and then this neck's length ways offcut then becomes the spar between neckblock and tail block. I was thinking about the forces involved and the dove tail cutting through the laminations, weakening the neck joint with the additional tension.
So going back to fundamentals - I was thinking the idea originally was a neck through style with a hollow body attached.
If the neck block is actually the neck, and spar connects into the neck block itself, it creates a through neck (although slightly angled) and the joint can be deep without a dovetail. This strengthens the joint, permits vibration through the joint between the two sections (both are laminated).

Adding the tailblock then means you have a neck, spar and tail. Adding a small tail edge temporarily for the trapezium string, the entire thing can be strung to test tension (even the headstock with the tuners added) over a temporary tall bridge block.

Picture = 1000 words (not entirely to scale but indicative):

This joint then keeps the full width of the laminate and doesn't have a dove tail problem with the dove tail cutting across the laminate joints.

Now you can see that I can make the tail block intially not bonded to the spar and it shouldn't cause a problem. However the neck block in the body will be a U shape that then the neck sits in. This rethink of the body, means I can close the box by doing the top and sides with the U neck block and the tail block, route and add the perfling/binding and then notch the top as needed for the neck joint (naturally attempting to keep it as small as possible and flush with the neck.

[Nick Kitchener]

Slow progress but deliberately so. I've made one of the more difficult cuts - 23" (60cm) with a Japanese hand saw keeping within 1mm of the line on both sides of the block. This means I now have the spar and neck blocks.

A mock up of the spar concept:

The fingerboard should be be an inch to the right but the idea is just to see what it will look like. I think another optimisation is to shift the spar right and have a long scarfe joint provide ample contact area.

I'm also heavily reading into getting the arch structurally sound - the cube rule seems logical and mathematically correct.

I've also tried chaldini lines with tea leaves, but using thin balsa ply and weighting down on top off the outine (anchored edge will be closer to the final expectation). This works - you can see mode 3, mode 2 (dipole) and I think mode 1 (monopole) but the little decrepit 4" driver isn't able to move the board enough (50W amp and signal generator driving). It's interesting but more of a test of what's needed to make it work - I may need a larger full range driver (like a 6 or 8"). It may be that sand works better than tealeaves.

Next steps:
1. sides final bend of the back
2. prep neck and tail blocks
3. bend the linings - two layers of 4mm thick and glue to the sides
4. prep some side braces and glue those

I think at that stage there's not much I can do - I really need the arch height/bridge height with any deflection to design in the neck angle and for that it's looking at sorting the top plate:

1. Top plate carrier
2. Prep and glue top plate
3. Create load test too and initial load test with no profiling (this shouldn't move at all)
4. Create initial arch profiles
5. Test the arch (not hollowed).
6. Start with a thick profile, then slowly modify following the wood behaviour rather than specific set measurements.

These are the wedges - they're not AAA quality as I'm on a budget and I've prioritised their sound rather than stunning looks - after all this will be my guitar.

Spruce top plate:

Maple back plate:

This has a few occlusions but the plate is bigger than the spruce (both in terms of length, length and maximum height), it looks like an ex-cello plate.


I'll post up the profiles once I've done them for critique.

[Nick Kitchener]

I wanted to have a good look at the back occlusion, so I've tided up the back and glued it. It still needs work but for now it shows me more than enough:

I've marked in red circles the key defects and a proposed outline on the underside. Although this will cause a few issues I feel that it should be fine as long as the tools are sharp but if needed I can put small bracing to stabilise after the carve.
Note the circle top right is on the opposite side of the board (a small hole).

The plate just as it is now, before any carving, rings approximately E4 (open high E string).

The front plate is clean, there's no issues with the wedges and any of the marks have disappeared with a plane.

[Alan Carruth]

The only real 'defect' I see in that is the bark inclusion, which falls outside the outline anyway. The rest looks like 'normal' figure.

[Nick Kitchener]

The only real 'defect' I see in that is the bark inclusion, which falls outside the outline anyway. The rest looks like 'normal' figure.

Agreed.

I prepped the front wedge. Quite impressed myself I managed to get zero light through vs the straight edge and with the joint dry fitted. Glued up and it came together really well. I came back 3 hours later to find it stuck to the ply former thanks to some squeeze out creep past the masking tape (my fault for not putting something that's glue proof down. Finally managed to separate the two (I'll not make you wince with tales of needed screw clamps applying pressure and a wooden block and mallet, it finally moved when I was about to give up to cut the ply wood to pieces), it's back in clamps and drying but no damage and the joint is really good - it already sounds like a tom tom drum.

Next on the list are the two chassis and contraption to show movement when apply bridge downforce - probably next week some time.

[Peter Wilcox]

I glue tops and backs on wax paper.

[Nick Kitchener]

I glue tops and backs on wax paper.

I’ve used baking paper before. Waxed paper is probably a little more resistive to disintegrating.