floor medallion construction

[al ladd]

I'm building a 44" diameter 7 sided floor medallion for a very unusual house. The medallion will replace one in very poor condition, (and not well-made to begin with). I want mine to better endure the test of time. The medallion serves as a hatch for a crawl space used in theater performances, so it needs to be an integral single piece that's removable on rare occasions.
The vendor that specializes in this (Czar Floors) seems like they do 5/16" thick CNC cut jigsaw puzzles essentially, glued to a substrate for total thickness of 3/4" to integrate easily with wood flooring.

I'm wondering about glue and movement issues, and so ideal thickness of the hardwood, and glue type. Mine will be constructed of 7 pie shaped pieces, with the pockets CNC cut into each pie slice, and the male members cnc cut to match, the whole thing assembled and vaccuum pressed onto a Garnica ply backer. My CNC isn't big enough in either table size, or travel, to do the whole thing at once. The background wood is hard maple, and the pie slices are about 19" wide at the periphery, so there's a lot of potential force on them here in New England.

Seems like there's a difficult choice on wood thickness--thicker and more chance of bad movement problems, --thinner and short life for future sanding.....Polyurethane glue?
Rough preview attached--though the maple grain will run from periphery to center in each pie slice.507872

[Kevin Jenness]

Interesting project. I would be wary of thick solid wood overlays. The shop I used to work at did a 48" diameter circular countertop with random 1/4" maple pieces glued to a ply substrate which developed some significant problems. If the finish is kept up thinner face material will work out better over the long run imo. I would suggest 1/8" with the same thickness backer laid up on a 1/2" substrate.

Are you planning to do v-inlays or conventional style?

[Clint Baxter]

I put together some six sided pieces similar to your design last winter and each of them had separations in the center during the more humid summer months. I'm located in the upper plains, not that far from the Canadian border and the humidity is quite dry during the winter. My maple pieces were slightly more than an inch thick, had the grain run from the perimeter to the center, and they were not glued to any type of backer. Fastening to a backer will definitely help, but I'd advise you to keep your maple thickness relatively thin. I don't know how thick you're able to get away with, but the thinner it is, the better your chances that the joints do not separate. Kevin's suggestion above sounds about what I'd try if doing it again.

Good luck on the project and I hope to see your end result posted here in the future.

Clint

[Jim Morgan]

In segmented turning, the grain runs tangent to the arc rather than perpendicular to it; you might be better off like that here. Pie-shaped pieces with grain running radially will move more at the periphery where they are wider than at the center where they are narrowest. This will make the assembly try to assume a conical shape, but if it's glued down it can't do that and instead will crack.

I agree with keeping the top layer no more than about ⅛" - that's enough to provide plenty of wear.

[Bradley Gray]

You mentioned that the hatch was removable. Is there a sub structure?

[Jim Becker]

I think that 5/16" or 1/4" thick inlays will likely not present a wood movement issue for this application. I'd probably do the 5/16" in this case if the assembly will be put in place and then the whole floor leveled by sanding prior to finishing. If you will be pre-finishing to match the existing floor height, I'd be comfortable with 1/4". Hopefully, there will not be a huge amount of sanding required over time, honestly.

[Edward Weber]

I think that 5/16" or 1/4" thick inlays will likely not present a wood movement issue for this application. I'd probably do the 5/16" in this case if the assembly will be put in place and then the whole floor leveled by sanding prior to finishing. If you will be pre-finishing to match the existing floor height, I'd be comfortable with 1/4". Hopefully, there will not be a huge amount of sanding required over time, honestly.
+1
Typically, 1/4" inch is about the max where you don't need to worry, anything over 1/4" is when I have to consider movement. That being said, I don't think 5/16" is a deal breaker. Laminating to a good solid substrate is the key. I would also align the grain to flow around the medallion.

[Tom M King]

If this is not the typical 30 foot rule theater construction:

https://www.youtube.com/watch?v=nTaQC7lZHTI

[Richard Coers]

https://sawmillcreek.org/showthread.php?78265-When-is-a-veneer-too-thick
https://sawmillcreek.org/showthread.php?81314-Making-Your-Own-Veneers
https://sawmillcreek.org/showthread.php?247759-Veneering-over-stave-core
https://sawmillcreek.org/showthread.php?245977-Veneer
https://sawmillcreek.org/showthread.php?241041-Wood-Veneer-on-Wood-Substrate-Part-2
https://sawmillcreek.org/showthread.php?152189-How-to-edge-glue-3-16-quot-solid-wood-veneer
https://sawmillcreek.org/archive/index.php/t-81314.html
https://sawmillcreek.org/showthread.php?178546-Veneering-Resawing-questions
https://sawmillcreek.org/showthread.php?43697-Thick-Veneer-Question
https://sawmillcreek.org/archive/index.php/t-20747.html
https://sawmillcreek.org/showthread.php?68736-Thick-Veneering-and-Wood-Movement

[al ladd]

Thanks for these replies. There doesn't seem to be a consensus. Respected opinions recommending 1/4-5/16", with most, including most opinions on the many posts Richard linked to suggesting 1/8" max is safest. I think given that the house is for sale, and so I'm more worried about splitting this winter than sand-through 20 years from now, Ill go thinner.

My plan is 5/32" to start, glued to 1/8" bb so I can pocket to the same depth. Then the pie pieces will glue to 3/4" Garnica with hdf faces, and I'm thinking of balancing with 1/4" bb on the bottom. There is no super structure supporting it anywhere but the edges, so either a web of 2x4's or 3/4" ply to stiffen it so it can be danced on.

I'm wondering if anyone has experience with Unbond800 vs Gorilla glue in the sort of application. I guess the key quality is protection against glue creep.

Here's the setting (worth looking at !)and the original I'm replacing, when it still looked decent. (My recommendation was to replicate the original, but owners kids decided they had a better idea...)
https://www.trulia.com/p/ma/colrain/68-van-nuys-rd-colrain-ma-01340--2000404425?mid=11#lil-mediaTab

[Kevin Jenness]

Seems like you could use a thicker substrate for stiffness rather than adding framing after the fact. Assuming the medallion's edges sit on the subfloor they would be rabbeted to get the face flush w/ the surrounding floor. Could be a torsion box to minimize weight. How will the piece be lifted for removal?

I would suggest using the same or similar material with a similar grain layout on the bottom for balance. Plywood on the bottom won't balance the pull of thick veneer in a piecut pattern.

I was interested to see the variation of opinions on maximum recommended face material thickness for your project. My own experience leads me toward the conservative position favoring thin veneers, but others must have had success with thicker, perhaps more stable species.

[John TenEyck]

I would use Unibond 800 over GG. Both are rigid and waterproof, but Unibond 800 has a very long open time, doesn't require high pressure, and is far easier to clean up.

As for thickness, I'm in the group that says no more than 1/8". If your inlays are full depth then there's no reason. If they are V-bit carved then bad things will happen the first time the floor gets sanded regardless of how thick it is. With 1/8" thick veneer you can run the grain of the maple pie segments as you want. If they are 1/4" thick I'd be a LOT more comfortable if they went parallel with the perimeter.

John

[Christopher Charles]

I'd also tend toward 1/8" as a max thickness for the veneer, 1/2" (or thicker) ply with a rabbet around the perimeter to fit flush, and would consider an inlayed element in the center to minimize issues getting all the points to come together perfectly and future-proof from separation in a visually obvious spot. Cool looking house/project!

Best,
Chris

[Edward Weber]

Just some food for thought
https://flooring-experts.com/hardwood-floor-thickness/

[Tom Bender]

44" diameter might be a bit springy for a dancer in a theater situation. Adding a couple of ribs to the bottom might be wise.

[al ladd]

Work finally underway. I'd already cut my maple to size for pieces going radially, which also gives nice pattern with curl, so I'm going against the grain of advice from trusted voices here,---but I guess I'm unsure of how important that is. I am getting bowing, gluing the .15" veneer to 1/8" bb with gorilla glue. I dried the veneer some before adhering it, getting it 1% below emc in my late summer New England shop, and I think (though i built a fire last night to keep the shop air drier) they're taking on some moisture. The triangles are nearly equilateral (51,64,64) so I'd have that action going on anyway i think. Would have been nice to balance the tri's with a backer--but sooo much work.

My router is a weird one, i built largely for joinery work. I use a small rotary vane vacuum pump for my prime hold down, and gasketing out something as large as these triangles, especially as they're a bit bowed, is not easy, but I got through it. I was disappointed that my seams were just a little open at the ends, and so I took a light jointer pass on each side, and had to modify the angles a few thousandths to get the last piece snugly in there. So I might have to pay a little extra attention to the male pieces that span a seam--make them a bit narrower in the lateral direction.


Making the flowers will be the fun part. I'll inlay the half or so of the roughly 100 male pieces that don't span a seam, and then thickness sand while i still can (16" open ended wide belt) before final glue up. Brazilian tulipwood, west Indian satinwood, satine, even some pink ivorywood, lots of figured koa, a really tasty mix.

This all would be a whole lot easier if i had a big CNC router! Cutting a 44" diameter 7 sided regular polygon wasn't a piece of cake on my shop made sliding table. I had to raise the blade into the cuts to have enough of the table engaged to get a straight stroke.
508244508245508246

[Kevin Jenness]

Would have been nice to balance the tri's with a backer--but sooo much work.


Essential work, I believe. The pieces will keep trying to move unless you keep the shop at a constant relative humidity. A backing veneer would make the assembly more stable

I like how you used the gasketing with your gridwork and in combination with the t-slot hold-down. Holding the work in place is half the battle.

The piece looks great so far, quite a challenge.

[al ladd]

"Essential work, I believe. The pieces will keep trying to move unless you keep the shop at a constant relative humidity."

Well, Kevin called that right....but i got it done, and now have questions about how to complete the bottom:

I have an unusual veneer balance issue with the 44" diameter 7 sided floor medallion I'm constructing. It will serve as a hatch over an unheated, often damp basement space, with no support except at its edges. Since the bottom surface will be exposed to air I'm guessing often in the 60-80 % humidity range year round, normal recommendations to treat top and bottom of panel the same don't seem like they'd apply to keep the panel flat.


The top veneer, mosty northern maple (plainsawn) is about 1/8" thick. It was laminated in pie shaped sections to 1/8" baltic birch so I could machine 1/8" deep pockets for inlay. Those triangle pieces, 18" wide at the base (the width of the veneer, as the grain runs radially) were like hygrometers, and as soon as the shop air started to rise above about 50% they started warping significantly. Before gluing the veneer down, I dried it to about 1% below my un-dehumified shop's EMC for this time of year, verifying the MC with a pinless meter by comparing it to my maple workbench. My reasoning was this would be about the midpoint of the air it will be exposed to in use. Then I had a struggle holding the pieces flat for CNC work as they warped. I wished I had balanced the panels, just to ease their construction. I finally realized that i simply need to bring the shop air moisture down further than I had thought, and once I built a fire and used my minisplit to bring down shop moisture to about 48 they flattened pretty well. I kept it there for laminating with Gorilla glue to 3/4" Garnica HDF plywood.


After seeing how sensitive the pieces were, I did the math, and realized that a 1% change in wood moisture (which we'd expect from about a 5% change in air moisture)results in about a full 1/16" of movement across the 18" of maple. So with a likely difference of 20% or more between the moisture levels of the top of the medallion versus its bottom, relying on balancing the lamination on both sides won't work to keep it flat. (Right?) So I"m thinking i need a strong structure to resist the warping stress, and I don't know how strong, but i should err on the side of too strong. The hatch will be removed by pushing it up from the bottom very infrequently (it was previously used for a cool effect in theater performances, and the house is up for sale, and I can't think of another use--the basement it leads to via a crawl-through tunnel has stair access), but the current owner/seller is adamant that its current functionality be retained.


So, i guess a torsion box is what i need, and the engineering for this is something I have no experience with. The torsion box will need to fit in a 35" diameter circle, so some of the periphery of the hatch will have to be only .8" thick. The opening the hatch covers is a 35" diameter circle, so the plywood to the outside of this will have to be reduced from its current thickness of 1" (3/4" ply + the )
1/4" veneer open sandwich.

So, something like a 16-24 sided polygon with a web-work of frame members inside? Maybe make the ribs from 1/2" pine 4-6" wide, for light weight and substantial glue surface? Maybe 1/4" BB for the bottom skin?
Photos of top awaiting finish.


508438508439508440.

[Kevin Jenness]

Nicely done, it came out just like the picture. Getting all the points to line up is not easy.

Making a thicker substructure will tend to keep the overall medallion flat, and a torsion box can keep the weight down relative to framing lumber. I would suggest a box 2 1/2" to 3" thick with 1/4" skins and 1/2" ribs 5" o.c. I still would recommend a balanced panel, with at least four 1/8" veneer pie-cut pieces on the bottom. Equal plies on either side of the centerline is just a good rule to follow in most cases unless the panel is fastened to a structure that can't move, and the rh difference in your case suggests that it will want to move. A highly moisture resistant finish would be a good idea, such as epoxy overlaid with varnish.

[al ladd]

Thanks Kevin--can you recommend an epoxy finish? I can see where limiting the extremes the top veneer will experience would greatly reduce the chance of problems.

[Kevin Jenness]

Thanks Kevin--can you recommend an epoxy finish? I can see where limiting the extremes the top veneer will experience would greatly reduce the chance of problems.

Two or three coats of West System using 207 clear hardener sanded back to smooth followed by a standard floor finish should work. You could omit the sanding and topcoat on the underside as long as you get an equivalent thickness of epoxy. As always, a sample is a good idea to evaluate how the epoxy will color the inlays.

Rather than make a many-sided polygonal torsion box I would just use 6 or 8 wide pieces for the outside frame and bandsaw it circular.

[al ladd]

I finished with 4 coats Boero Challenger UV two componant polyacrylic varnish. It's a high end boat finish, and 2 part polys have similar moisture content change resisting qualities to epoxy. I didn't have the thinner for it, and so I used foam brushes, and it dried very fast, and so I couldn't get perfect coats. Between coats i cut back high spots with a light touch with a scraper . In place I see it's too high a gloss for the surrounding satin water based floor finish, (not very well done)which I expected. I tried a thorough wet sanding and 0000steel wooling to see if I could create a more satin finish, but there are still glossy spots, and I think i have areas of overlapping coats making a uniform gloss impossible.


I'm hoping I can now simply brush on a coat of suitable floor finish for finish coat, preferably an oil based poly with a good brush and call it done.... Do I need to try one more coat of the 2 part first? Suggestions for the finish coat?

I'm not looking for a rubbed out look, just a uniform(ish) satin finish.


508900508901