Thanks for your thoughts BH. I don't see a problem gluing up a multi layer ring. Each ring would be butt glued separately, and then combined in one glue up in the vacuum bag at the end. When I made the original ring the segments were joined as straight sections. I cut the ring after glue up using a router on a large trammel, I would pin the layers as I added them to the stack. Whatever little misalignment there would be would be inconsequential when I then routed the ring.

John

I have a couple thoughts about this.
1- is that piece of glass fit tight to the ring?
If it is seems like this would restrict seasonal wood movement of the ring.

2- I have had some bad experiences building doors out of Sapele. It seems to have a lot of inconstant movement. More so than what the shrinkage calculator here shows. Never had any issues building windows with it but doors with larger parts and panels I have experienced movement and cracking.

Nice table, risky construction and given its outside horizontal even with the cover I think you could expect some movement.

I wouldn’t worry about it either.

Hi Joe. Thanks for your thoughts. There is a 1/8" gap all around between the glass top and outer ring so I don't think there's any reason to believe that's the source of the problem. I've used Sapele on quite a few projects now and I agree that sometimes it can be very squirrely. Usually, however, I find those boards when ripping or jointing them. It's like they have internal tension in them and every time you cut them they spring a different way. Beyond that I haven't found any problems with unexpected movement in use, but I'm sure you've used a ton of it for every pound I have so if you've seen it then I know it must be true.

You mentioned "risky construction". Since many of the joints opened up that must be true. So how would you build it if someone asked for one?

Thanks.
John

Yes, it could get crowded with the three remnants of the butt joint and the two splines - .225 if you make them all equal. However, even if your splines were only .120 or so, they would do the job. It is the surface area 9 times greater than the original that will hold it.

But B H Davis' solution with layered rings is a good one, tho I doubt the TBII on the butted miters. I'd run 5 ply. But if Bernie made it, it would look like window casing, eh?

You mentioned that if you could hold the wood - prevent any gaps - then the rim would warp or do something as it moved the stress to another part of the table. I think that is faulty thinking - the weakest point will show stress if there is any stress to the table. What is going to do? Move to the area between the joints? Move to the toe? How will it get there?

One must rigidly think only along the lines that define the materials not adding properties we wish it had, or work with vague notions of inaccurate properties of wood. The stress will not move to another area.

I see no problem with the basic design and build - I think you did a fine job in the mechanical design. It shows you understand things correctly. Were I to pass by it in your shop in process, I would make mention of generous splines, but that is all. I would not consider it risky behavior.

Only in hindsight would I reexamine every assumption I made earlier. Seeing those gaps, feeling my stomach sink, I'd be back in the shop, making longer splines, a narrower rim, and for good measure, I would add a rim to the rim. Same thickness, about 1" wide or so, maybe a v joint where it meets the segments. The rim parts - 12 of them - would have a simple butted joint halfway between the segment joints. You could even make the rim sit on a continuous tongue routed out of the segments for additional glue surface area.

But, and this is a big butt, so to speak, I would bet lunch at a modest restaurant of your choosing, that if you doubled the splines, and made them 3" total length, that the thing would stay together. 10 years, outside.

And five bonus points for starting off one of the better technical discussions. A lot of good thought in this thread.

John,
I’ve had some good Sapele but have had more problems with it at than any other species. Even the lowly Khaya. Sipo is a better choice if you can get it.

How would I build it? Probably the same amount of segments you did only with 1 1/2” To 2” tenons on one end and a slot on the other. Just because I am set up for that type construction. I like that you used small width segments, that is good construction. I would probably use west epoxy just because it’s horizontal. I have never had issues with TB3 and use that for everything vertical.
I think loose tenons or brick layed as BH recommends would be fine also. Last few years I have been using end grain finger joints for all my door and window radius work but that might not be the best here.

Honestly I would never build such a fine piece for outdoor use. I’m reminded of the time I built a nice solid teak outdoor golf storage cabinet for one of my well to do customers so he could practice on his ranch. I thought I had all the expansion contraction figured out. Unfortunately the ranch hands placed it where sprinklers hit half of it. It did not fail but some very strange movement and open joints resulted from that!

Okay deep spline proponents. What happens when the 3" wide tenons gain moisture? Won't they drive the joint apart if the tenon is seated fully in the groove? I'm not picking a fight - an honest question.
I build rings for tent roofs. Each segment is joined to another with two 10mm x 30 mm x 50 mm dominoes. I'm starting to think that the cross grained tenons are not as good in this case as tenons running with the grain.

Tom, I think Sipo Dominos would be a good way to join these. Easier to assemble than the slot and tenon I was thinking about.

It sure sounds to me like this thread, given how little material there is in the ring, has the makings of a build challenge. If you think more splines/glue area will hold, make one and lean it outside under the roof overhang on the sunny side of the shop. If you think bricking would work, make one and same exposure.

I would guess on close inspection a single season would speak volumes

Thanks for the additional input, everyone.

I'm a little surprised no one proposed to remake the ring like a wagon wheel, with a steel band shrink fit over a ring with unglued joints. It's one of the ideas I thought about when I first saw the cracks, discounted, then came back to recently. It would be interesting to pursue if for no other reason than I've never done it.

I like the idea of the build challenge, but I don't see how to do it w/o building each idea full scale. And if that's the case I might as well just build what I think best and see how it works. It won't take long to see the initial results, just one Winter. Nothing bad seems to happen during the other three seasons. It's the trip from Fall to Spring that does the damage.

John

A reasonably full size mockup is only about 10lf of material. Not cheap but not unreasonable to prove the point.

John,

Your table looks very nice. Sorry to hear about the expected failures. It would be useful to provide a close up pic of the joint failure.

Rather than doing two tenons I would stick with the single tenon for aesthetics. Go deeper. 1.5 or 2” is easy on the table saw.

T-88 is a good epoxy. Not one I would choose. West System is the one by which all other’s are judged. It’s decent and available.

The mistake people make with epoxy is machining & clamping too tight. Typically we make mort/tenons .002” loose for pva wood glue. Epoxy needs really thick glue joints. The joints should be visible 1/32”(.031).

Properly done epoxy joints are ugly. All surfaces(metals, composites and wood) should all be sanded with 80 grit. Table saw kerf marks are a good thing. Never take something off of the jointer or planer.

West System will soak into end grain and fills it up. Adding filler does not make it stronger. The cotton microfibres soak up the resin and keep it leeching out. Precoat. Wait 10 minutes and glue it together.

We have laminated insane amounts of teak & Mahogany with West System. Never once used filler.

I believe that your project could have worked.

Pixels, not cellulose and lignin, but for fun tossed in a quick 12 segment ring likely much smaller than the original table (40" outer dia segmented, 38"+ round). Sapele on Shrinkulator - .040" radial on a 5.5" width with 5%MC change. Keeping the heels tight leaves .0259" and takes less than a tenth of a degree off the bevel. Taking all that error out on a single joint leaves a 1 1/8" gap.

View higher quality, full size image (1632 X 1056)

I've seen full circles fail as Mark shows in his CAD drawing, and in multiple smaller gaps around the circle.

Picture a simple MSG interior door or window 45 degree casing corner. If it shrinks it opens up on the inside of the corner. If it expands it opens up on the outside. The stronger the wood species, wider the moulding and greater the change in seasonal moisture content the harder it becomes to prevent this from happening. Take the same concept to a full circle and you are making the separation forces even stronger.

I feel strongly that the only way to pretty much assure you won't be making this a third time is by using the brick layered method. The wood would have to split through both layers along the grain for that to fail.

BH Davis

Make it anyway you want and coat the entire thing in epoxy, a few coats thick. No moisture, no movement.

Here is a photo of a joint that has barely opened up and one of the two really bad ones. The ring is 50-3/8" OD and the ID is 4" smaller.

John

View higher quality, full size image (666 X 375)

View higher quality, full size image (666 X 375)

So at 50" thats some back of the napkin increase of like 24% on the numbers in the drawing. Depending on how many joints are really bad it seems your well below the math though any movement and the forces associated there would seem to present a problem.

It looks like you have nailed the failure math Mark. The 1/8" gap in the original posts seemed to be crazy big for simple expansion/contraction, but it comes down to weakest link opening up. The rest are gaps due to mitre angle change.
I'm with BH here, brick laid construction. Add suspenders of a UV stable epoxy sealer .

You guys are all way smarter than me with this but I would say a combination of several solutions would be the best. Brick or more tenons, perhaps increase segments, and while I hate the thought of epoxy encapsulation (most building a table like this would not be pleased with the encapsulated finish) it would definitely slow the moisture though it will still happen eventually with time. Perhaps the very best is the OP's idea of a heat shrunk wagon wheel arrangement (stainless) though it has issues of its own (scorching).

My only liability statement would still be to build a small "table garage"/vestibule and simply roll the table out when needed. ;-) Completely impractical but...

I agree with David though, nice to see some chatter here.

Thanks all. Yes, nice to get some constructive discussion. At this point I'm thinking either a two or three layer brick laid ring with an epoxy sealer, but not encased in layers of epoxy.

One more thought. Similar to the metal hoop, what about putting a dado in the outside of the ring and then winding carbon fiber cord saturated in epoxy into it? What say you composite guys?

Other than my carbon fiber road bicycle I don't have experience with that material. However there is more flex in my carbon fiber bike than an alloy bike. My initial reaction to the idea is that while it might be strong enough to prevent the outside diameter of the circle from increasing it might not be rigid enough to keep the circle from going out of round.

BH Davis

If I were in damage control the last thing id be doing is contemplating the introduction of an unknown variable.

I would not recommend adding the carbon fibre (even if it when a weave which would look cool) because your photos show gaps opened on the toe of the mitre indicating shrinkage being the culprit. Your carbon band would be wrinkled or delaminated.
One poster recommended the possibility of more segments being better. I don't think it would as Marks cad drawing shows. When the diameter grows with wood expansion the circumference would have to enlarge by 3.14 times that length. However the wood doesn't expand appreciably in length. The number of segments actually would make things marginally worse mathematically at least because the grain becomes more perfectly perpendicular to the radius.

My error disregard my last comment. Expansion opened the gaps.

I reasoned the following some time ago, and times like this it is good to revisit. Best for open miters and breadboards.

Imagine 2 parallel immovable steel beams 12” apart. Rip a piece of 3/4” x 2’ long commercially dried white oak to just fit between the two beams. It needs to fit across the width of the board. That oak is at 8%. Now reduce the moisture to 6”. The oak will now have a gap - will be less than 12” and fall out of the beams, due to its hygroscopic nature. Now hold the oak in place, between the beams, and add moisture to 12%. The board will swell to the 12” fit, and be tight in the beams. As we add moisture, the board will not increase in width due to the restraint. However, all the wood fibers are trying to swell, but can’t due to mechanical restraint.

Next, we will decrease to the 8% MC that held the panel at 12” where we started. The panel will be less than 12” since all the fibers have been crushed.

Now let’s glue our panel to those beams. We will assume the glue will easily hold the oak. So the 12” 8% MC panel is glued in place, and we will reduce moisture. The panel will eventually crack, often with a sudden, loud crack. All the fibers were being stretched to the breaking point, and one point eventually gave way.

Back to the 12” panel and the beams. Glue this one in, at 8%, then add moisture. The fibers will all compress and will take the abuse. Then reduce moisture to 8% and the panel will crack as it “shrinks”.

This thought experiment (apologies to Dr Einstein) has helped me understand the basics of wood movement.

And, Tom G, I noticed a incorrect statement. You said the 3” wide tenons could swell in the grooves and force the joint open.
However, the grain in the tenon or the spline is glued to the segments as long grain to long grain, or nearly so. There is no appreciable movement that would degrade the joint. There is no movement or expansion in the length of wood.

Ahh. I misunderstood the orientation and thought I was seeing end grain in the photo. Now everyone can disregard everything I say :) I'm used to it

I feel (not assume) most anyone reading pretty much either understands on paper or has suffered their own misery with regards to radial and tangential wood movement. At its core its pretty basic stuff. Anyone who purchases a dry compressed sponge and tosses it in the sink understands. The leap to longitudinal is a short one as well.

To me the nuances come in all the tricky algebra and trig that come in when a situation like this thread pops up and the radial/tangential suddenly turns into a change in the miter angle, which when multiplied out over several segments equals a sad-face event.

Im still interested in the comments that say the fury that the compounded condition will exact on a joint or joints can be confined with glue in any long term warranty-able sitiuation.

I personally dont know how it can. Which leaves me with the agreement to "celebrate" the joint (tip of the hat to old 'nahm).. or engineer some other non-wood solution into the process (wagon wheel, kevlar, alien tech)

The wagon wheel has worked through the ages because it creates so much compression it crushes the wood fibers trying to expand. But of course through seasons and use this wet/expand/crush, dry/shrink/get lose, wash rinse repeat lands in a make a new one situation.

Im going to build the brick and a deeper or perhaps more tenon option and put them in the torture test. Out behind my shop, wet, damp, cool, then to the front, sun hot, dry.

I cant fathom any glue line resisting 1-1/8" of movement across all the joints with the exterior swings.

Mark, I can imagine a glue holding the joints together, whether spline, M&T, or brick laid construction. What that glue is I'm not sure, but I'm very sure the wood will expand/contract as it wants regardless of what glue is used and that means either the wood is going to crack somewhere or the ring will not remain flat. Somehow that stress is going to be relieved.
While the total change in expansion/contraction remains the same for a segmented ring of any given geometry it's clear the amount that occurs in any one joint is reduced the more segments the ring is made of, So if you use enough segments the stress will eventually be low enough for some glue to hold the segments together. I used 12 segments; perhaps 24 would have survived. .
A brick laid ring undergoes the same change in expansion/contraction but the stress is distributed over an infinitely larger glue surface area. I'm not convinced the toe of each joint will remain tight, however. Logic says they open or the wood cracks or the ring bows.

The test pieces you plan to make should provide some solid answers and I'm looking forward to how they behave. Thanks very much for doing this.

John

I’m one of the resident WW boatbuilder/woodworker guys. I have done thousands of hours of composite boatbuilding(carbon/glass/epoxy/foam&balsa core)

The easiest way to pimp your ride with carbon fibre would be to route a dado on the underside of the wood ring. Bond in a tow(yarn) of unidirectional carbon/epoxy.

It is a really bad idea. The carbon will not move. Large chunks of wood needs to move a bit in width & thickness. Also you are making an imbalanced panel.

You could use your lovely table and skin both sides of it with carbon fabric. Essentially using the wood as a core. Kinda defeats the purpose of building a nice wood table.

If you made the table out of wood veneers you could add a layer of uni carbon, which would work and make it have a nice looking jet black inlay.

Your 2nd bad picture is what an epoxy butt joint should look like. The first one looks like a pva joint. Like I wrote before. Perfect epoxy glue joints are ugly.

Putting 2 coats of West System clear with UV(to minimise the moisture changes) plus a couple of coats of high quality marine varnish(to protect the epoxy) applied yearly forever may keep your table looking good.

Send Santa an annual fruitcake for good luck and you might have a chance.

Thanks for your thoughts Adam. So if I wanted to make it as a veneered composite what would be the best core material for exterior use? I guess I could make a Sapele plywood ring with all the layers running at 90° but that's a lot of work. Alternatively I could veneer some core material, but which? I doubt exterior grade MDF is up to the task, even with a hardwood insert on the OD and ID, but that's just a guess on my part, no actual experience with it. Would balsa wrapped in or layered with fiberglass, Kevlar, or carbon fiber be suitable?

John

Veneer was the incorrect word. Imagine the wheel of a ship. It’s a ring. We make them the same way you would make a curved jamb. 3/16” laminated layers. We’ve added carbon laminate layers in between teak. Makes a rather stout wheel.

Kevlar is only really used for impact resistance like a kayak or canoe. Glass is structural and used for waterproofing layers. Carbon is stiff. Very stiff. Glass has the same tensile strength, but carbon is 50 times stiffer. Plus it looks cool.

We’ve built many tables for yachts. Weight is a serious concern(more weight=less speed) We typically build them with end-grain balsa core. It’s available 1/2,3/4,1”. It comes to us skinned on both sides with 1 thin layer of fibreglass. It’s used for structural bulkheads in boats. The problem is if you buy regular balsa off the shelf it’s kerf cut, so it can bend around shapes. The core by itself is weak. It needs decent strength skins for its strength.

Point being veneering a sheet of mdf/ply with thin wood is a table. The wood might as well as be paint(non structural) You are now asking about making a composite panel akin to a hollow core door. You need to use decent strength identical skins to stiffen the light weight high compression core.

Another option is VG WRC(vertical grain western red cedar)

We build sailboat rudders using cedar as the core and skin it with glass and a bit of carbon. They also build entire 50’ sailboats out of plywood like layers of cedar. Then glass it inside and out.

Cedar is very stable & strong when mixed with epoxy & glass.

I’ve been working at the yurt last few days and thinking of this thread looking at the center ring. It’s quite nice and made by a local shop with CNC. In over a year and joints are tight as ever.
Poor picture but it’s bricklayed in 2 layers.
Not outside but exposed to radical temperature changes, sunlight and humidity changes at 11000 ft.

Am I right in thinking there would not be much movement of a ring with a lot of segments? Only in the width of opposite segments?

John, have you noticed any change in the 1/8” revel of your glass?

View higher quality, full size image (1280 X 960)

Adam, if I understand you correctly, the sailboat wheels are laminated with the layers wrapped around the circumference, yes? If so, that's one of the first ideas I though about in how to solve the problem. What's the cross section of a boat wheel? My ring is 4".

Joe, that's a pretty similar application for the yurt ring, of course it's much smaller in diameter and hopefully never gets wet, but still, it's the same concept. Yeah, lots of segments divides the total mismatch in outer diameter as the wood expands and shrinks into a smaller amount that each individual joint has to accommodate. I picked 12 segments for the ring I made because it looked good, but if 24 would solve the problem then it would look better after 3 years. Brick laid might get me back to only needing 12 segments. I don't know.

Adam, tell me again why putting a layer of carbon fiber between a brick laid construction is a bad idea. Or perhaps a layered construction of 3/16" wood with carbon fiber between each layer.

I'm not married to Sapele either. The ring could be made from mahogany, etc. and still look good with the rest of the table.

Thanks all,

John

Joe,
All an exercise in information and overthinking at this point, and I have no real data, but would assume at that elevation and your location things would be relatively dry with regards to wild humidity swings so to speak?

I would imagine just having a roof over it and being up at the peak would help a bunch but then you'd have the off angle sun exposure throughout the day/seasons hitting just the inner edge of the ring. Place sitting empty/cold then toasty fire, etc.. Would seem that ring for inside exposure gets a good beating but no direct water.

My only thought with more segments was to spread the movement out over more joints thinking it would reduce the movement to something whatever glue used could resist. But that depends theoretically on the movement being equally spread over all the joints which may not be the case given the material variability, any gluing/clamping inconsistency, etc.. The old "weakest link in the chain".

Broken record, you all have way more data but pretty interesting exercise none the less.

I make 4 yurt rings a year from Eastern white pine. Their section is 6" x 6" four layers of 1 1/2" cnc cut bricks. I route tapered mortises into each half ring then glue the pair together to accept the rafters. While the rings don't get direct exposure to the elements, they get baked in the summer heat and are subject to daily condensation from interior moisture.
Once I began to buy pine well in advance so I could get the MC to 8% we have had no joint issues. The rafters enter the ring on an angle and wedge in both dimensions (a PIA to get right) so they inflict a torque on the plies but I have seen rings up for four years now without any damage.
The glue is type 2 pva. Each layer is glued with two large dominoes per joint and the rings are glued into two plies the same day to maintain flat faces.
The bricked assembly has proven very stable.

I like brick laid construction. We use it for curved door and window frames. No spring back to try to anticipate. I can see it as a solution for the table in question. In fact,
I can see brick laid work in thickness or in width.

Wagon wheels for rapid continental expansion are built from green wood since there was no time for drying. These wheels would shrink and dr6 on the fly, and serve well. Once dried out, it was customary to let a wagon stand in shallow water to tighten up the wheels again.

The rim controlled/limited expansion and took the wear. The hub, spokes and felloes all expanded and shrunk within the metal rim.

In that sense, a “wagon wheel” approach to building would not give better results. In fact, John's 12 segments mimic the felloes of a wheel.

Tom,
That sounds like a very high end ring!
This one is only 2 layers of 2” flat sawn fir. I think it’s about 4 1/2 foot diameter.
I like the idea of pocketing the rafters also.
We have a local yurt company here that has been around for a while. I forgot about the heat in those. It’s like a solar oven in the summer.

I think brick layed would be the strongest but not sure I would like the look on a table like this. I’m not a fan of splined construction. I think I would still slot and tenon if building this.

Just give my suggestion of the TB II pre-sized joints a try on a couple scrap sticks. That will show you the strength of that joint for temporary layer joint construction.

I can assure you that you don't need any splines or tenons during the glue up. And as I said the face to face glue surface will negate the need for anything other than butt joints once completed.

BH Davis

Very interesting forum to read, but I think I would just go out and buy a plastic table at this point.

Btw, your table looks beautiful so I hope you get it figured out.

Have a wonderful day everyone.

Dave

It sounds like you are speaking from experience?

BH, agree no joinery needed for brick construction. I usually just butt and pinch dog the segments together for this.
Slot and tenon- I was referring to segments only for building this.

Joe:
The yurt company is Yurta.ca. The rings are for their 21' yurts. The rings are complex to make and I'm sure I'm not charging enough for them but I make up for it in the other wood components for the smaller yurts.
I use the Dominoes for alignment during glue up. All ten segments are glued at the same time. A single band clamp pulls the ring together.

I don't think any high quality hand made furniture should be kept outside, and I'm skeptical that a construction method which could prevent all those outdoor problems with solid wood even exists.
If you make joints strong enough not to crack, then probably the wood itself will give first.

In this case I would cut the ring out of decent quality plywood (probably two overlapping layers if you can't find sheets big enough to make it out of one piece) and cover it with veneer (maybe 1/8" thick) for that kind of solid wood look and soak it in whichever finish/conditioner that would help it survive outdoors as long as possible.

Would West System 655 G/flex be a good adhesive to use in a glue-up like this. It has a tensile strength of 32.7% ? I have used it when I thought wood movement would be an issue.

West System G/flex Epoxies