toolpath strategy for complex shape bookshelf

[alex pavel]

What is the size of the bookcase you are trying to make? The picture of the completed piece seems small although there is nothing to show scale.

I am pretty sure that the pictured piece is not veneered. I have done a lot of veneering and veneering of curved panels over the decades and I don’t think that you could veneer it.

It's not a picture, it's a 3D model. I know the image is misleading regarding size, because the wood texture on the model is not the correct scale. The size is roughly 450x2070x2600 mm or 17,7 x 81,5 x 102,4 inches.

[alex pavel]

The piece dimensions are around: 450x2100x2600.

Are these numbers correct? They seem much too big.

18" deep??

7 ft x 8-1/2 ft??

Yes, 6,8 ft x 8,5 and 17,7 inch on the deepest part, because the thickness (deepness) is variable. It should be a bookcase. Do you think it's hard to build it so big? I know that weight is a problem and I calculated about 160kg or 353 pounds if it's solid Oak.

[ger21]

Here's how I would approach it.
I would cut it in layers of 35-50mm thick.
Each "layer" would actually be made up of two thinner glued up layers, with the joints staggered. Cut the wood so that the grain follows the shape, so that there is minimal end grain. This will both look much better, but also make the individual layers more stable. It will also be much stronger.

Dowels or "dominoes" would be helpful at "T" joints, but placement would be critical to avoid cutting into and exposing them.

This would be a huge amount of work, but it's the best way for a successful outcome. Imo, of course.

[tomgardiner]

No doubt you know this will be a challenging project to say the least. Your strategy to cut this project in slices will have a waste factor on the order of 75 percent (you will be essentially cutting out the cabinet from a solid block of wood). I hope you have a fireplace to at least get some heat from the waste.
I would suggest taking a small representative section of the cabinet and machine the layers to see how it reacts when cutting and gluing (a 500mm x 500mm section with a complex intersection) .
There are many challenges to the design. You will have areas that will be thin wall sections of short grain which will be very weak and prone to breaking. Wood also expands and contracts with humidity change. Expect the overall dimensions of the cabinet to alter by a few centimeters in height seasonally. But I think your biggest concern will be sanding the finished piece.
Use a cheap hardwood like poplar or ash on a test section to work out the machining and assembly strategy.
If you haven't seen Wendell Castle's work, Google him. Have a look at what he was doing in the late 60's and 70's.

[Wayne Locke]

You could make it as a stacked lamination, however, the planning, construction and cleanup -sanding would be enormous. You would need to plan the joinery between the vertical and horizontal pieces so that the grain was going the appropriate direction for strength. You could not just glue up a wide piece and start cutting. If cutting from 8/4, 2' not sure of mm, you would need probably 10 slices. There shouldn't be that much movement with a good, properly dried hardwood but each or most slices might require a pass on the jointer which would require a bit of hand planning at places for the next slice to fit well. The back of the attached chair was two slices but each side required maybe a 1/32' jointer pass to get a good fit.

Whatever level your woodworking, Aspire and cmc skills are, this project will stretch all of them. Oh and sanding skills too.

[rscrawford]

I'm not sure how I'd cut a project like that. Making it out of solid wood in slices, you are going to burn through A LOT of wood, probably more expense in wood than you'll be able to sell the completed project for.

I estimate 1000bf of rough oak, or approximately $4500 worth of oak in my neck of the woods.

Would be more feasible to design it such that you could carve each shelf as a two sided cut (top and bottom of the shelf), then carve the uprights as two sided flat cuts (right and left side of uprights), designing them so you can cut each part out of 12/4 or 16/4 oak (might need a little less curvature to the design). Then join them (the joints will be on curved surfaces, so hard to glue up). If you use jigs and a finger joint bit, the glue-up might be easier.

[alex pavel]

Here's how I would approach it.
I would cut it in layers of 35-50mm thick.
Each "layer" would actually be made up of two thinner glued up layers, with the joints staggered. Cut the wood so that the grain follows the shape, so that there is minimal end grain. This will both look much better, but also make the individual layers more stable. It will also be much stronger.

Dowels or "dominoes" would be helpful at "T" joints, but placement would be critical to avoid cutting into and exposing them.

This would be a huge amount of work, but it's the best way for a successful outcome. Imo, of course.

Thanks Gerry! Great tips.
So what you're saying is double the number of layers? Why would i do that? Maybe i misunderstand what your'se saying.
Wouldn't glueing be enough? Do i also need joints to keep the layers together?

[alex pavel]

No doubt you know this will be a challenging project to say the least. Your strategy to cut this project in slices will have a waste factor on the order of 75 percent (you will be essentially cutting out the cabinet from a solid block of wood). I hope you have a fireplace to at least get some heat from the waste.
I would suggest taking a small representative section of the cabinet and machine the layers to see how it reacts when cutting and gluing (a 500mm x 500mm section with a complex intersection) .
There are many challenges to the design. You will have areas that will be thin wall sections of short grain which will be very weak and prone to breaking. Wood also expands and contracts with humidity change. Expect the overall dimensions of the cabinet to alter by a few centimeters in height seasonally. But I think your biggest concern will be sanding the finished piece.
Use a cheap hardwood like poplar or ash on a test section to work out the machining and assembly strategy.
If you haven't seen Wendell Castle's work, Google him. Have a look at what he was doing in the late 60's and 70's.

Indeed the waste factor is so high and this is one of the biggest problems. I don't see any way to reduce the waste other than bending the wood, but that's too hard for me to do. Yes that's a great idea, i will start with a small test section first. I didn't know about Wendell Castle. Impressive work and so long ago. I've seen him using an industrial machine in one of the pictures.

[alex pavel]

You could make it as a stacked lamination, however, the planning, construction and cleanup -sanding would be enormous. You would need to plan the joinery between the vertical and horizontal pieces so that the grain was going the appropriate direction for strength. You could not just glue up a wide piece and start cutting. If cutting from 8/4, 2' not sure of mm, you would need probably 10 slices. There shouldn't be that much movement with a good, properly dried hardwood but each or most slices might require a pass on the jointer which would require a bit of hand planning at places for the next slice to fit well. The back of the attached chair was two slices but each side required maybe a 1/32' jointer pass to get a good fit.

Whatever level your woodworking, Aspire and cmc skills are, this project will stretch all of them. Oh and sanding skills too.

OhMyView2a copy.jpg

OhMy3_2.jpg

I could make the slices thicker so that i have fewer layers and minimize the movement of the parts, maybe? Yes, i'm aware that i need to use the jointer to make the wood slices flat.
My levels on these areas are very low so after your feedback i am considering a redesign of the piece to make it easier to produce. I plan to mass produce it in the future so it needs to be built faster and easier. I do get help from someone who builds furniture but they don't have experience with projects like this either.

[alex pavel]

I'm not sure how I'd cut a project like that. Making it out of solid wood in slices, you are going to burn through A LOT of wood, probably more expense in wood than you'll be able to sell the completed project for.

I estimate 1000bf of rough oak, or approximately $4500 worth of oak in my neck of the woods.

Would be more feasible to design it such that you could carve each shelf as a two sided cut (top and bottom of the shelf), then carve the uprights as two sided flat cuts (right and left side of uprights), designing them so you can cut each part out of 12/4 or 16/4 oak (might need a little less curvature to the design). Then join them (the joints will be on curved surfaces, so hard to glue up). If you use jigs and a finger joint bit, the glue-up might be easier.

The wood here is maybe half that price, but still there will be so much wood wasted indeed. What you're saying is cut each shelf in 4? top-right, top-left, bottom-right, bottom-left? Because of it's irregular shape it would be like a puzzle to put everything back together. It seems that if i want to minimise the wood waste, the solutions increase the amount of hand working and organisation. Another idea is to use the wasted wood. Maybe make some stools with it. I guess i'll try a redesign. I'll keep this post updated with my further progress. Thanks for your help!

[Adrian]

Have you thought about doing it in high quality plywood instead?

[Tailmaker]

If I had to build this (I would try this only for the challenge, not for profit), I would do the following (assuming a solid model CAD system):
- "drill" dowel holes into the 3-d model every 6 inches or so.
- slice the resulting 3d-model into 3/4" - 1.5" thick layers
- break each layer into mostly straight segments (individual 3d-bodies) that butt up with each other and contain each at least 2 dowel holes.
- nest all segments onto a plane (size of your board or sheet blank) and combine back to a single model
- import to Vcarve/Aspire and make tool paths for the dowel holes and the 3d outside shape (probably only finish cut required)
- drill the dowel holes and use them for holding screws into the spoilboard
- probably a good idea to carve some sequential numbers into each piece to not get confused later
- cut the 3d-shape
- glue up with dowel alignment and clamp, probably best layer by layer with PVA glue
- sand, sand, sand

[tomgardiner]

I would also suggest breaking the whole into components. Here you will have to lean on your Fusion skills to create joints between components. I envisage tapered socket joints where the bottom of the wedge like joint would have a flat. The face of the finished piece would show the joints as sort of mitres. I wouldn't good for dowels necessarily as when assembling the joint must go together in a single plane of movement which will prove to be difficult (not to mention the difficulty of drilling off angle holes in the end grain.)
This will greatly reduce the waste. It will also allow you to do a great deal of the sanding before assembly.
Castle's shop would create foam models to perfect his design then digitize the model. The wire frame was sliced to create patterns for accurate bandsaw templates for each layer. Of late he was using a robotic arm to carve the form after glue up. His staff still spent tens and sometimes hundreds of hours to carve and sand to final form.

[ger21]

Here's how I would approach it.
I would cut it in layers of 35-50mm thick.
Each "layer" would actually be made up of two thinner glued up layers, with the joints staggered. Cut the wood so that the grain follows the shape, so that there is minimal end grain. This will both look much better, but also make the individual layers more stable. It will also be much stronger.

Dowels or "dominoes" would be helpful at "T" joints, but placement would be critical to avoid cutting into and exposing them.

This would be a huge amount of work, but it's the best way for a successful outcome. Imo, of course.

Thanks Gerry! Great tips.
So what you're saying is double the number of layers? Why would i do that? Maybe i misunderstand what your'se saying.
Wouldn't glueing be enough? Do i also need joints to keep the layers together?

Yes. Each "layer" to be machined would consist of two glued up layers. See attached image. You cut the parts close to the finished size, to minimize waste. You stagger the joints between the two glued up layers, to make it more stable, and much stronger.

This is more work, but uses a lot less wood, and is much stronger, and more stable. I would have to charge about $7,000 to $10,000 to build something like this at the size you mentioned.

[alex pavel]

Have you thought about doing it in high quality plywood instead?

Would there be any advantage? It's much more expensive and the layers are too thin.