Wood warping after carving

[ExperiaUSA]

Hi everyone.

I purchased a 3D model of a Byzantine cross for carving. I've carved it once before, out of Mahogany, and it came out great. This time, I carved it out of cherry. I glued up three strips of cherry wood that were approximately 8" x 36", giving me a blank of 24" x 36". I ended up carving the cross against the grain, length-wise (spanning all three boards).

Though I did not plane the blank prior to carving, it was pretty darned flat. After carving, and removal from the material (it was tabbed), the cross was really warped. I'm guessing that this is from "stress relief" within the wood itself causing this? Anyway, it is what it is.

I CAN push the two ends flat, and was considering cutting out a plywood cross that is offset 1/2" in from the carving, and gluing it. I'm worried that the stress on the cross could cause it to break at the seams.

Does anyone have a way to try to un-warp something like this without breaking the glue seams?

Thanks.

[ExperiaUSA]

I thought I uploaded images. Not sure what happened.

[litzluth]

The lesson here (which doesn't solve your immediate problem) is to try to use fully dried slow growth quartersawn wood for carvings. But even that doesn't always prevent stress relief when sections are carved. To further minimize movement, one could also create sections where the grain opposes from piece to piece.

As you can see from the pictures, most of the warpage is in the middle piece, which is a flatsawn section of rapid growth wood. The warpage is exactly as expected, tangent to the annual rings.

To re-bend the wood, you will have to plasticize it. That means a combination of moisture and heat. How much heat before your glue lines fail depends on the glue you used. Titebond/PVA is going to fail somewhere in the neighborhood of 150-155 degrees. What I would recommend is that you brush on a plasticizer (on the back) that is more effective than water. I like SuperSoft Veneer softener. https://www.veneersupplies.com/products ... ioner.html. Based on the pictures, you may only need to do the center section. Hit it with a heat gun and then apply the SuperSoft liberally. You can use a cheap Harbor Freight handheld laser temperature gauge to check how hot you are getting the surface. Keep the heat gun moving. I wouldn't go above 140F. Then clamp it in a fixture that will provide a flat surface and pray (should be easy given the subject matter of the carving). Leave it there till fully dry, maybe even up to a week. Adding a backing of plywood after that process sounds like a good idea.

[ger21]

Always run the grain along the longest dimension, as warping along the grain will be much less, if at all.

If possible, use thicker material to start with, and before carving, remove the same amount of material from the back first, then flip over, and carve. When finished, place the part on sticks so that air can get to both sides. Finish asap, both sides.
If you lay it on the table, the side exposed will gain or lose moisture, and warp.

[ExperiaUSA]

Wow. Thanks for the helpful suggestions! As you can tell, I'm a little out of my element here.

I'll now look into how to determine old growth from rapid growth. The three pieces that are glued together were all from one (roughly 12' long) board.

Also, noted on the direction-of-grain suggestion. In the first one I cut out of Mahogany, I did cut it lengthwise as opposed to against the grain. I cut this one against the grain to maximize my yield, not taking this warping into account.

The issue I have with cutting both sides is that I wanted to maximize the thickness. The thicker the wood, the greater depth I can make the model. At the standard board thickness (1.1") the thinnest part of the cross is just under 1/4". If I cut any off the back, I'd have to flatten the model. If I could have found a 2" thick piece of wood, I would do that. How can I get/make thicker material this size (roughly 22" x 14"). If I had a planer, I suppose I could cut a board into 2" strips, plain the flat sides and glue that up. But I don't.

[TReischl]

How can I get/make thicker material this size (roughly 22" x 14"). If I had a planer, I suppose I could cut a board into 2" strips, plain the flat sides and glue that up. But I don't.

That size and thickness of wood is very difficult to find. If you do find it the cost will be much greater if it has been dried. That is because it takes a lot longer to dry thicker material.

If you are going to do a lot of this sort of thing I would recommend a drum sander rather than a planer. I got rid of my planer at least 5 years ago in favor of the drum sander. All of the sages say that a woodworker must have a tablesaw, jointer and planer. For doing this stuff I find a tablesaw and a drum sander much more useful. A band saw helps for really thick stuff. Anyway, what have I got against a planer?

1. Snipe
2. Nicked blades which are not cheap to replace
3. The inability to thickness shorter stock, I can easily run a 6 inch long piece through the sander.
4. Tearout in figured wood

Yup, if I had to thickness a lot of board feet and time was a consideration a planer is the way to go.

I have done quite a few thick models, up to a little over 3 inches in thickness. I also like to work with pine. So after cutting big box 2X material to length I flatten one side on the CNC (that is why I also sold the jointer quite a few years ago) using a large bit 1.5 inches diameter with a 75% step over. I then flip it over and flatten the other side. Pass both sides through the drum sander after that. That all goes surprisingly quickly. I typically flatten in the CNC at 350 IPM, each side usually takes under a minute. I built a pair of vises for the CNC that make holding the pieces extremely easy and fast.

Working with a cnc on a regular basis requires figuring out what is going to work best for what you want to do. When I first started I did like most folks, put a piece of material on the machine, do the interior details then cut the piece out. That meant I always had to use material that was larger than the project. Having done a lot of woodworking I knew that cutting pieces to size on a table saw is much MUCH faster. Cutting them to length is much quicker on a radial arm saw. The problem was accuracy. Now my table saw has a digital readout and so does the RAS.

The vises are extremely useful, especially for smaller pieces, and really useful if the parts I am making require work on more than one side.

So, I guess the idea is that it pays to do some tooling up if you are going to do a lot of cnc work.

BTW, cutting stuff into strips and gluing them up as you wrote about solves a lot of issues, but it does not always result in the best looking project. Sometimes using a gel stain helps with that.

The next time I do a thick project I will document the process that I use and post it up here. There are bits and pieces of how I do things scattered throughout my posts but it might be nice to have that information all in one place.

[litzluth]

Short answer: You can typically assume growth rate is correlated with the width of the grain lines. Narrower grain lines mean slower growth.

Much longer answer and a little explanation about why wood warps, and what one can do about it:

Wood is mainly composed of lignin and polysaccharides--cellulose and hemicelluloses--that form concentric cell wall structures along the longitudinal axis of the tree. Each narrow dark grain line represents fall growth with a much greater lignin concentration (the tree's response to the coming winter). The areas in between the dark lignin rings are the faster growing spring and summer growth, and are predominately cellulose. The lignin ring/area is harder, less abundant, and less likely to wick moisture. The cellulose area is softer, more abundant, and more likely to wick moisture.

Wood shrinks when it dries and expands when it becomes more moist. Wood warps when the moisture content in different areas changes unevenly. A relatively drier section will shrink more than a section with higher moisture content. That creates stress (tension) that will deform the wood. The converse is true. A relatively moister section will expand more than a drier section and create stress (compression) that will deform the wood. Exposed endgrain exchanges moisture much more readily (up to 10X). Wood can warp in any direction, but the most common problems are bowing, cupping, and twisting.

Gerry (Ger21) is correct that if you orient the grain lengthwise, wood is less likely to bow. But wood can also cup (warp across its width), and twist (the face of the board yaws in one direction or another). If I am interpreting the pictures and what you said correctly, the wood in your project, particularly the middle piece, cupped. Using quartersawn or riftsawn wood helps protect against cupping. I don't really know the cause/best remedy for twist, other than to ensure the wood does not encounter changes in moisture content.

Before you milled your piece, the wood had an equilibrium moisture content (EMC). As you milled it, sections that had not been exposed to air were suddenly exposed. The EMC of the newly exposed sections were impacted by the relative humidity of the surrounding air. I try to keep my shop right at 45% relative humidity (not all that easy in Pensacola FL, but 3.5" of closed cell foam on the walls and ceiling along with a dehumidifier makes that doable) and work with wood that has had at least 6 months to equilibriate. That 45% RH translates into about an 8% EMC in the wood (assuming the wood has been sitting in that 45% environment long enough to stabilize). Assume, however, the RH in your shop on the day you milled the piece was 75%. That translates to a 14% EMC, which means the sections exposed to that higher moisture content will want to expand as they absorb water from the surrounding air.

The best solution is humidity control and working with pieces that have equilibriated to the surroundings. If you can't achieve that, as Ger21 points out, you can finish and clamp quickly to try to keep the EMC of the wood from changing.

Hope this helps.
Best,
Mike

[CarveOne]

One other method is to glue two boards together back to back, using slow cure epoxy such that the annular rings are opposing each other. Don't use a water based wood glue. (Which adds moisture.) This has worked well for me on a few projects that justified doing it this way.

Also, I recommend having both a surface planer and a drum sander. Surface plane to make boards thinner, then drum sand to final dimension. This is what they are designed to do. And, you can't have too many tools in the shop ...

CarveOne

[Rcnewcomb]

A few related articles:
Avoiding Cupped Panels
Glue Ups & Grain Direction
Gluing Up Panels

[chemstock]

from the image, it looks like the warp was concave.
likely the internal stress of the piece was relieved or the interior of the piece was not as dry as the outside. The warp suggests the latter. you could try fixing the cross to a table or slab to pull it square and laying a lightly damp towel on the top of the cross to have the cell structure reabsorb moisture. if that works, seal the wood both sides with sanding sealer before it has a chance to redry out
When I use water based dyes and stains, I get warping and solve that issue by using the same product on the flip side; so maybe this process would be the reverse.

going forward, try a two step rough out with a large machining allowance on the first then give time between for the wood to relax.

[gwilki]

I realize that I am late to the party, but in my climate, using cherry that is 8" wide will always result in warping and cupping. When we are gluing up panels, we never go with boards wider than 5" and we try to stay narrower if we can. This does make grain matching more difficult, but more importantly, it results in more stable panels.