HELP! Jewelry box doors twisting

[Lester Sak]

I'm building a jewelry box with storage on either side for hanging necklaces. The doors are frame and panel using sapele and maple with mitered joints, each reinforced with a dowel. I cut the wood to rough size and let it sit in my shop for about a week before milling to final size and assembling the doors on December 18th.

328863

A few days after attaching the doors, I noticed a slight gap at the top of the right door (picture below).
328868

I'm getting back to work in the shop now that the holidays are over ready to work on the drawer fronts when I noticed the gap has gotten larger on the right door, and now there is a 9/32" gap on the left door.

328872

Here's the left door lying "flat" on my tablesaw.

328873

The box has been stored in the house, only going out to the shop when I work on it. I obviously need to build new doors, but I'd like to know how to avoid this happening again. I'm wondering if the rails and stiles need to be thicker or wider (or both) to resist twisting? The panel is floating in grooves that are about 1/16" wider and longer than the panel dimensions. The door is 12"H x 7.5"W x 1/2"d.

Any suggestions would be appreciated!

[Lee Schierer]

The 1/16" gap on the sides seems a bit small. I think what may have happened is the panel expanded and has put pressure on the frame causing it to twist. How long had the wood been n your shop before you started working with it? Where was the wood before you moved it into your shop? Moving it into the house probably created the problem. With the doors closed, the inside can't see the moisture changes the outside is seeing. The humidity level in the house is probably much lower than your shop since it is heating season. This is one of the reasons to have a moisture meter to check moisture levels before starting work.

[Lester Sak]

Thanks for the feedback Lee. The wood has been in my shop since the summer and was in a similar environment prior to that. The point you made about what might have happened with the doors closed makes sense, although the box has been in the house since before Christmas and I just noticed the large gap today. Are you suggesting that once it's in the shop it should stay there until complete? And then once in the house, keep the doors open until it has time to acclimate? How would I have done things differently using a moisture meter?

EDIT: I agree that the humidity in the house is lower than my shop, but how would that cause the panel to expand? Wouldn't it contract in a drier environment?
EDIT 2: Never mind, I understood what you meant as it related to the doors being closed (which would retain the higher humidity from the shop) vs. the lower humidity on the outside.

Thanks for the help.

[Andrew Hughes]

I don't there is any way to prevent wood from Twisting that's going to twist.Since your building new door this is a good way to learn what moved the panels or the miterd frames.I would take them all apart.
Look carefully at the grain of the wood that's the culprit.
I would stay away from maple that's has soo much cathedral grain all flat sawn.And maybe a small one in the middle with quarded or rift on the sides.Still no guarantee.

[Lester Sak]

Thanks Andrew, my plan was to take the doors apart to see what happened. But I'm still trying to wrap my head around all of this.
Are you suggesting that even if the woods moisture content is ideal (I'm not suggesting it was as I don't currently have a moisture meter, this is hypothetical) and the humidity in the room remains stable, the wood can twist at some point in the future due to it's grain pattern?

And because the relative humidity is different in the shop than the house, what's the best way to deal with that? Should the piece I'm working on stay in the shop until it's complete? And then once in the house, keep the doors and drawers open until it has time to acclimate? I would typically bring anything I'm working on back in the house when done for the day with the thought that changes in humidity are potentially greater in the shop than in my house.

[Dan Hahr]

Use a straight edge to see which board warped or twisted. Examine the grain to see if that's why it did. Use the straightest wood you can for the stiles and rails.
Dan

[Lee Schierer]

When making thin panels from thicker material, it is a good idea to remove equal amounts of material from each face to equalize the moisture. If you resaw a piece of thick lumber to get thinner material you need to stack and sticker the sawed pieces for a period of time so they can acclimate. Remember that moisture changes occur slowly in wood. If a board is 1" thick it can take months for an outside moisture change to reach the interior. I would recommend that once your doors are constructed and test mounted that you finish all sides as soon as possible. The finish will mitigate any surface moisture changes. I would also recommend that you not lay the doors on any flat surface and leave them that way for more than a few minutes and definitely not to store them stacked where air can't get on all sides.

[Joe Jensen]

In my experience wood moves both predictably and unpredictably. By predictably I mean wood moves with humidity changes and we know how it will move with humidity changes. If you have a piece with the growth rings perfectly perpendicular to the face of the board it will move very predictably. With quarter sawn you can get good sized boards that have the grain pretty perpendicular to the face. In my experience flat sawn (where the grain runs mainly parallel to the face) is much less predictable. If you had a piece where the grain was perfectly parallel to the face the movement would be pretty predictable too. But only very narrow boards would enable perfectly parallel grain. Usually the gain on flat sawn is very random ranging from quarter sawn to flat sawn and every thing in between. The grain will be in a cupped pattern on the end of the board, and usually that cupped pattern is different on the other end. I suspect we see twisting when the quarter sawn'ness to flat sawn'ness changes along the length of the board.

We also see tension in some boards when they are cut but usually that movement settles pretty quickly.

I always use quarter sawn for parts where any movement can't be contained or movement will be very visible. On a little door like that I'd want quarter sawn for all the door parts. If the grain didn't work in quarter sawn I'd veneer over the quarter sawn with a face veneer.

Once I built a very complex headboard and bed. It had many parts and I was very careful to select boards for their stability. Once built it saw waiting for finish for a few months and when I got back to it one of the smaller boards that was 1.25" wide, 7/8" think, and with width of a king size mattress shrunk in length. yes that's not a typo, it shrunk almost 1/16" in length. I was perplexed and I actually reached out to the author of "Understanding Wood" Bob Hoadley. Bob said that he had seen that before, said it was completely unpredictable and he called it wild wood. This board was a part of a complex molding and none of the other parts shrunk. The molding was glued up before being cut to length so all the parts became one when I glued and then I cut that one to length. I guess you can never be sure.

Understanding Wood is a bible as far as I am concerned. Highly recommended.

[glenn bradley]

First off, nice design. That is going to be a real treat when you're done. I use sapele but, try to stay with the quarter sawn material as I find it can get pretty lively when tensions are release during milling.

Your movement occurred after milling and assembly which makes me target the panel clearance as the likely culprit. There is also the possibility of catching the corners of the panel with squeeze out from the frame assembly thereby gluing it all together instead of ending up with a floating assembly.

I agree with Dan's suggestion to use a straight edge to determine which boards twisted and which way; then mark them prior to disassembly. Once the doors are apart you can recheck the frame pieces to see if they are still tweaked or are less so.

I leave 3/16" clearance all around a floating panel of the size you show and up to 3/8" depending on species and panel sizes. Space balls, panel barrels or even pre-applied blobs of silicone rubber cement (allowed to cure) can serve to keep floating panels centered and act as a dynamic shock-absorber during movement.

Please let us know what you find.

[Bill Orbine]

I notice you mention the doors being 1/2" thick. Did you resaw this sapele from thicker stock when you started making the doors? The same question applies to the maple panels. As if you can get two stiles or rails with from one thicker board? A lot of times, along with moisture balance....... you also need to deal with tension relief. So.....you need to acclimate the wood again for a few days as well let the tension get relieved.

[larry senen]

I notice you mention the doors being 1/2" thick. Did you resaw this sapele from thicker stock when you started making the doors? The same question applies to the maple panels. As if you can get two stiles or rails with from one thicker board? A lot of times, along with moisture balance....... you also need to deal with tension relief. So.....you need to acclimate the wood again for a few days as well let the tension get relieved.


yes,, if the panels are resawn that's where the problem is. a more stable panel is a veneered ply or mdf core . even if the veneer is not highly figured.
blame kiln dried lumber for the problem;you've released all kinds of tension in the wood.

[Lester Sak]

Use a straight edge to see which board warped or twisted. Examine the grain to see if that's why it did. Use the straightest wood you can for the stiles and rails.
Dan

Dan, I plan on taking the doors apart tomorrow, and I'll be checking with a straight edge before and after. Thanks.

[Lester Sak]

When making thin panels from thicker material, it is a good idea to remove equal amounts of material from each face to equalize the moisture. If you resaw a piece of thick lumber to get thinner material you need to stack and sticker the sawed pieces for a period of time so they can acclimate. Remember that moisture changes occur slowly in wood. If a board is 1" thick it can take months for an outside moisture change to reach the interior. I would recommend that once your doors are constructed and test mounted that you finish all sides as soon as possible. The finish will mitigate any surface moisture changes. I would also recommend that you not lay the doors on any flat surface and leave them that way for more than a few minutes and definitely not to store them stacked where air can't get on all sides.

Lee,
After resawing the wood, I did sticker it for over a week. This pic was from Nov. 13th:

328926

When I did final milling, I alternated sides going through my thickness planer to equalize what was removed from each side. Your point about finishing all sides asap once the doors were completed makes alot of sense as well. I'm wondering now if that would have prevented this from happening as it did take several weeks before the twist got this bad. And I learned my lesson in the past about leaving resawn panels from a different project in the past lying on my tablesaw overnight. It wasn't pretty :eek:

[Lester Sak]

In my experience wood moves both predictably and unpredictably. By predictably I mean wood moves with humidity changes and we know how it will move with humidity changes. If you have a piece with the growth rings perfectly perpendicular to the face of the board it will move very predictably. With quarter sawn you can get good sized boards that have the grain pretty perpendicular to the face. In my experience flat sawn (where the grain runs mainly parallel to the face) is much less predictable. If you had a piece where the grain was perfectly parallel to the face the movement would be pretty predictable too. But only very narrow boards would enable perfectly parallel grain. Usually the gain on flat sawn is very random ranging from quarter sawn to flat sawn and every thing in between. The grain will be in a cupped pattern on the end of the board, and usually that cupped pattern is different on the other end. I suspect we see twisting when the quarter sawn'ness to flat sawn'ness changes along the length of the board.

We also see tension in some boards when they are cut but usually that movement settles pretty quickly.

I always use quarter sawn for parts where any movement can't be contained or movement will be very visible. On a little door like that I'd want quarter sawn for all the door parts. If the grain didn't work in quarter sawn I'd veneer over the quarter sawn with a face veneer.

Once I built a very complex headboard and bed. It had many parts and I was very careful to select boards for their stability. Once built it saw waiting for finish for a few months and when I got back to it one of the smaller boards that was 1.25" wide, 7/8" think, and with width of a king size mattress shrunk in length. yes that's not a typo, it shrunk almost 1/16" in length. I was perplexed and I actually reached out to the author of "Understanding Wood" Bob Hoadley. Bob said that he had seen that before, said it was completely unpredictable and he called it wild wood. This board was a part of a complex molding and none of the other parts shrunk. The molding was glued up before being cut to length so all the parts became one when I glued and then I cut that one to length. I guess you can never be sure.

Understanding Wood is a bible as far as I am concerned. Highly recommended.

Joe, Interesting point about the effect of grain pattern changing from one end of a board to the other. And I agree the best chance of success would be to use quarter sawn wood. Thanks for the recommendation on the book, I'll check it out!

[Lester Sak]

First off, nice design. That is going to be a real treat when you're done. I use sapele but, try to stay with the quarter sawn material as I find it can get pretty lively when tensions are release during milling.

Your movement occurred after milling and assembly which makes me target the panel clearance as the likely culprit. There is also the possibility of catching the corners of the panel with squeeze out from the frame assembly thereby gluing it all together instead of ending up with a floating assembly.

I agree with Dan's suggestion to use a straight edge to determine which boards twisted and which way; then mark them prior to disassembly. Once the doors are apart you can recheck the frame pieces to see if they are still tweaked or are less so.

I leave 3/16" clearance all around a floating panel of the size you show and up to 3/8" depending on species and panel sizes. Space balls, panel barrels or even pre-applied blobs of silicone rubber cement (allowed to cure) can serve to keep floating panels centered and act as a dynamic shock-absorber during movement.

Please let us know what you find.

Hi Glenn,

The more I think about it, I'm thinking that my clearance was greater than 1/16", but I'll know for sure once I've got the doors apart. Good idea about marking them before, thanks. And I'll definitely report back. Hopefully, someone else can learn from this :)

[Lester Sak]

I notice you mention the doors being 1/2" thick. Did you resaw this sapele from thicker stock when you started making the doors? The same question applies to the maple panels. As if you can get two stiles or rails with from one thicker board? A lot of times, along with moisture balance....... you also need to deal with tension relief. So.....you need to acclimate the wood again for a few days as well let the tension get relieved.

Yes Bill, all door parts are from resawn stock which I did sticker and allowed to acclimate for over a week before final milling.

[Mel Fulks]

I just read whole thread but don't see how thick material was to start with. If you were using 8/4 to net two 4/4 pieces I would say most of us try to avoid that. But when you can't avoid it ,remove most of the wood from convex side . Better chance of success since the immediate movement is more likely to help you. Equal wood removal on both sides is a rule more suited for stock that is pretty straight in the rough.

[Lester Sak]

- Post Mortem -


I checked all frame members and the panel with a straight edge, and all were flat once the door was taken apart, so it appears that the twist was caused by the panel expanding, but being restricted within the frame as Lee suggested. The miter joints were pretty strong due to using dowels to reinforce them. If they were weaker, I think one of the joints would have opened up instead.


The problem was caused by (a) not leaving a large enough gap for expansion in the frame and (b) tack gluing the center of each side of the panel into the frame. The panel had a pretty wild grain pattern so predicting how it might have expanded/moved would have been difficult. I don't think I would have had the problem if there was enough room for the panel to expand by providing a larger gap in the frame, and using space balls as suggested to control movement of the panel.


I also think that this might have been avoided if I had put a finish on the doors once they were completed as also suggested. This twist didn't happen until several weeks after the doors were assembled, so it probably was caused by humidity change which was enough to cause a thin panel to expand. Putting on the finish might have prevented the humidity from affecting it, or at least minimized the impact.


I spent quite a bit of time reading about moisture meters and decided it's time to invest in one. I believe you get what you pay for, so I ordered a Lignomat Scanner SD Moisture Meter (http://www.amazon.com/Lignomat-Scanner-SD-Moisture-Meter/dp/B003CLQIGY/ref=sr_1_33?s=hi&ie=UTF8&qid=1452442863&sr=1-33). I also plan on picking up Understanding Wood: A Craftsman's Guide to Wood Technology (http://www.barnesandnoble.com/reviews/understanding-wood-r-bruce-hoadley/1030868053), which Joe suggested. I was up pretty late last night reading chapter 13 ("Drying and Control of Moisture Content and Dimensional Changes") of the Wood Handbook - Wood as an Engineering Material (http://www.fpl.fs.fed.us/documnts/fplgtr/fpl_gtr190.pdf) from the Department of Agriculture. Alot of good information there too.


Hopefully my next set of doors will be stable thanks to everything I've learned here, and maybe someone else can avoid a similar problem by reading this thread.


Thanks again for all of the help and suggestions!

[Dave Lehnert]

Only time I ever had a problem like you is in the winter time working in the shop and bringing the project into the house to show others the progress I was making, Learned very quick to keep project in the shop start to finish.

[Dan Hahr]

I would not make rails and stiles from resawn stock unless they were allowed to dry for several months. If there was any stress inside from drying, it will be released slowly once the inside is exposed. The inside and outside are usually not equally dry, and it will warp slightly as it dries. If the panel swelled up it would separate the joint between the rail and stile.

Dan

[Mel Fulks]

I notice Dan's reply is close to mine, so I want to build onto it. Most resawing I have done and seen done over the years in commercial shops has been to make veneer or pieces thin enough to bend for laminated curves. Using it the way OP did is avoided. Sure some here have made large mouldings from thick stock and noticed how how convex the bottom of moulding gets and requiring a relief cut on back almost edge to edge to be usable. Some wood suppliers print statements to asure customers they take great pains to prevent "case hardening". I translate those claims as "our product is as good as is available".

[Lester Sak]

Dan/Mel,

Thanks for your thoughts. It doesn't appear that tension released from resawn rails & stiles caused this, as all pieces remained straight and flat after the door was taken apart. But please let me know if there's something else I should be considering.

[Andrew Hughes]

Hi Lester,There is on more thing I thought you can double check,Is the miter cut square to the face.Hopefully you have a good little square.To check. Miters are tricky my saw only gets me close so I handplane for the final fit.I'm pretty fussy about it so maybe I'm going overboard.

[Lester Sak]

Hi Andrew, I'm the same way re: the miters. I have an Incra miter gauge and finish off with a shooting board if needed. Thanks for mentioning it though.