Acoustic archtop non-traditional design and build

[Carl Dickinson]

I use 4" wide packing tape down the center of my go-bar deck base. Renew it with a new layer occasionally.

[Nick Kitchener]

An update. After getting hit for a week with this flu/cold and then slicing my hand open and various other pass-times:
1. The mould is smoother and now fits together.
2. The neck and tail blocks have been cut, then planed so the side edges are completely flat.
3. The sides and blocks are glued.
4. The solid linings are cut and shaped to be very close in size, they're also scraped to remove any contaminants and smooth the faces. The sides where the linings will go are scraped.
5. Two of the linings are bent to shape - some crazy fool decided solid linings would be fun not thinking of the really gnarly tail and neck bends.. supergluing fractures works well.
6. I've glued one of the linings in, this takes 4 g clamps and 48 pegs for one lining. This clamping hell, plus the spring of both side+lining, means I'm giving the glue over 12+ hours for each lining to ensure it's bonded.

Dry test fit.

I'll be double lining and bracing, so this could take a while!

[Nick Kitchener]

The first four braces are in and next it's the vertical side braces.

I know there's quite a discussion between cloth and solid wood. Also I did note a conversation that Alan had with a museum restorer about not leaving the braces simply butted up against the linings.
I've taken that into account - shaped and interlocked them accordingly:

I still have some more to go but once they're done, I will flush cut the tops then install the second set of linings over the top. Joys of bending thick wood!

[Barry Daniels]

One of the mistakes that new builders usually make is overbuilding resulting in a non-resonant body. My motto is "Build Light".

[Nick Kitchener]

Strength in geometry!

The linings are about 4mm, there are two layers giving 8mm total, and I will scrape them back a bit, i’ll also round the inner edge (neat and tidy!).

The neck block is quite chunky due to the non traditional neck joint planned.

January should see the top being carved and more interesting progress. I’ll also update on the profile for that once the xmas chaos has subsided.
This is my first archtop so it’s a slow thoughtful progress, i got some of the impatience out with my first solid state build!

[Dave Rickard]

One of the mistakes that new builders usually make is overbuilding resulting in a non-resonant body. My motto is "Build Light".

Yep

[Nick Kitchener]

Update - I've finished gluing in the linings today. With the holiday break and the local temperatures -5 below freezing it's been a case of running an oil heater whilst working outside to warm the piece and keep the glue above 10degC. It's been brought into the house for curing and that's paid off.

It still needs to be cleaned up but it survives being outside of the mould and goes back in for safe keeping. Also it's got a drum sound just on the sides.

So think a quick clean up tomorrow and I'll also clean up the top and take some final measurements before I start working on it.

[Nick Kitchener]

A little clean up, I've had to adjust the colour balance as the phone hates the garage light.

The sides are straight, but it's neither symmetrical or perfectly flush again the mould - that's fine, it's my first (and probably only) acoustic. It's got its own tap tone too which surprised me (although at this stage completely useless observation).

I've not cleaned the outside or fully flattened for the plate surfaces yet, that will wait until closer to the glue up.

Next up is starting on the top.

[Nick Kitchener]

I've started to make the deflection tester for the top:

The idea is that the parts are clamped down to the bench, weight is placed at the end of the tube, and downforce is exerted via a dummy bridge. The dial then measures any change of the top.

The same dial should then allow me to thickness the top against the defined profile. I'll draw out the profiles before hand and post them up here for comment.

[Nick Kitchener]

I've been having a bit of a think - before I charge headlong into the whole carve-test-repeat process.

I figured some simple deviation calculations would be prudent, just as a finger in the air. I'm interested in if the design could be done without bracing on the top.

This has some basic assumptions, not specific to the wood sat here, but the modulus of elasticity (MoE) and the difference of spanning across the grain as the guitar would be doing.

There's good news and bad news testing with both 20Kg and 10Kg.

A simple beam is realistically going to have to be over 10mm deep to reduce the deviation to around 1-2mm (closer to 1.1 so age creep is managed).

I managed to get Gemini to give me what it thinks will be rough curate circloid calculations for 10Kg load and max of 2mm deviation - it's looking at basically 31mm curve height down to 15mm base height with a flate base (so that's 31mm of thickness!). I don't think that's going to really resonate much! This is before adding a bridge on top!

The obvious point here is to add some bracing and reduce the thickness.

[Alan Carruth]

Based on my experience this sounds 'way out of line. Is this a 3-D shell model?

[Nick Kitchener]

Based on my experience this sounds 'way out of line. Is this a 3-D shell model?

That was an experiment with the AI. It felt unreliable. My own guess would be something like 15mm and possibly 8mm.

I'll create some profile curves based on something like that and see. In the end there's a shape that would fit the guitar, if that shape is not able to support without bracing then I'd add bracing as a remedial action.

[Alan Carruth]

On the current builds of a 16" steel string archtop I'm using an outside arch that rises to 19mm above the flat bench top. The edge is 3.5mm thick where it levels out. If it wants to be thinner I'd remove material from the inside. The last such guitar I made had a top that was uniform in thickness at 3.5mm, with 'X' bracing 9.5mm tall at the high point.

My most recent steel string at 17" long had a top arch 21mm above the bench level. Somehow I neglected to note the top thickness in the center, but recall it as ~4.5mm, tapering to ~3 at the edge above the waist. It has 'parallel' bracing, again less well documented than it should have been, but probably on the order of 12-13mm tall at the bridge. The long arch on that one was flatter in the middle than my usual, since it was built as an 'homage' to Gibson; an attempt to make an 'acoustic copy' of a particular Lloyd L-5.

Neither of these showed anywhere near as much as 2mm of sinking on the top when strung. Both have been in use for some time with no issues.

[Nick Kitchener]

Thank you. The day job is the tech industry so it was more of an explanation, just the research alone was enough to trigger some reality check questions based on research. Look back on my second post I can see I'd not been precise in describing this. The task for this Sunday is to calculate and finalise the contours, then this week I can start.

The original design discussed had a 12mm dome height with X bracing, and I was guessing at around 5-6 mm thickness under the bridge and go from there. I thought - well could I make a non-braced guitar. I felt the 12mm is a little too shallow for that, so adding some arch height would help.

However now I'm guessing (perhaps a little more following experience of others) with more towards 15mm, to keep it on the more mellow side of the sound, start at 8 mm thickness and then tune from there with a bridge around 5mm, with around 3-4mm at the recurve. I'll then use X bracing.

[Alan Carruth]

'Way back when, Frederick Saunders wrote a paper in the IEEE Journal (did I get the right number of E's?) on 'The Violon as an Electrical Circuit'. In one of the footnotes he pointed out that the thickness of the top should scale with the arch height to preserve dynamic similarity. I've tried using a high arch/thin plate to get high stiffness for weight, but it doesn't work acoustically: the sound is 'harsh'. I suspect that the plate 'breaks up' into smaller vibrating areas too soon. A top that is too thick is simply unnecessary weight. It turns out that the 'right' thickness for a 12mm arch height is 3mm or so: it's what they use on violins (more or less) and also works for 16"wide archtop Classical guitars.

[Jim McConkey]

Yes, correct number of E's! (used to be a member), but for what its worth, they publish many journals with different specialties.

[Nick Kitchener]

I'll continue with the programme this afternoon and make it available for anyone to use if they can operate it - the idea is to have the outline in a csv, have some parameters, run the python in google colab notebook (ie this is free to use and cloud based), and then it will output a top with the profile for the top and underside (creating a thickness profile). This shouldn't take long. I don't want to turn this into a tech thread, so I may put that in a separate threat elsewhere.

[Nick Kitchener]

Getting there..

Maths and AI is interesting.. in that it will decide that the slope of a function can simply be mirrored in the incorrect axis and still be valid! Some hand coding and were slowly getting there:

The idea is to create a top of profiles, then create a second function to add thickness...