Crown Moulding Expansion/Contraction Problem

Who or what is responsible for gaps appearing at joints after installation? June 4, 2004

I was called to a jobsite today to look at some crown moulding that we provided (did not install) for a tenant finish project. We provided other millwork and casework on the job as well.

There is a problem with the end to end crown joints pulling apart (see photo). This is occurring all over the building, but does not occur on any base, casing, etc. The base and crown were run from the same lumber.

The GC said it couldn't be anything other than the wood shrinking. I advised him that wood expands/contracts very little along its length, and if there was enough moisture in the wood for the crown to shrink that much, we couldn't have run it through our moulder (it would have had to be green). And if it had shrunk that much along its length, the width would have been much greater (but the width was the same). The species is African mahogany.

We didn't resolve anything, but I could tell the GC (PM and Super) doubted what I was saying. We brought along two references indicating the fact we were stating, but it didn't seem to help.

1. How would you handle the situation?

2. Any ideas on what is causing this joint to open?

The cause is not poor installation. Something has moved, and I'm guessing it is the building, sheetrock, or something else. I am not an expert in those materials, so any help would be appreciated.

To give you an idea on what I'm dealing with, after discussing the crown problem, they had me look at some trim where the "finish was coming off." They scraped the face with their fingernails and something flaked off. At first I was worried and suggested they might have used too much water while wiping the wallpaper paste off our millwork. But, after scraping a little bit more, we saw that the flaking *was* wallpaper paste. When we scraped enough, we got down to the nice shiny finish. They both accepted this problem as theirs!

I'm confident the joint problem is not ours, but proving it to them is another story.

Forum Responses
(Architectural Woodworking Forum)

From contributor M:
From the looks of the picture, this crown was just but jointed together. Was it glued? It sounds as though your millwork was pre-finished. I have seen moulding shrink enough to get gaps like that. I might be wrong, but I'd say that that joint needed to be mitred together and glued. That is how I would have had it installed. End grain to end grain, even glued, will not hold very well. The mitre will allow a little cross grain gluing, thus creating a stronger joint. Does that sound right? I'm not 100% sure.

From contributor C:
While I don't pretend to have answers to your questions, I'll offer some comments...

> I advised him that wood expands/contracts very little along its length...

This is true, and we have a Knowledge Base article at WOODWEB that discusses this issue. It can be found at:

I've copied the content at the end of my post... the second sentence of Gene Wengert's answer seems to sum up the relative movement when wood shrinks lengthwise.

My feeling is that it would be unlikely that longitudinal shrinkage was the culprit, particularly since you mentioned the crown was run from the same batch of cherry that the base was run from. If longitudinal shrinkage was the culprit, it would likely also have caused problems with the base. I'm also interested in whether the other joints in the crown opened up. I'd be more inclined to point to building movement as the culprit, but if that were the case, I'd also expect there to be evidence of building movement (i.e. drywall cracks).

I guess the real issue is "How would you handle the situation?" Being right doesn't always translate into being off the hook. I've been in situations where I felt very strongly that the problem was not my fault, but also understood that drawing a line in the sand was not the best way to handle the problem. I think your actions at this point might be influenced by your past/ongoing relationship with the contractor. Do they provide a substantial portion of your annual work? Are they good to work with? If so, working with them to fix the problem might hurt initially, but pay off in the long run. Do they provide little work for you, and are they problematic to work with? This might convince you to take a tougher stand.

Another shrinkage related article at our Knowledge Base that you may want to review:

From "Calculating Longitudinal Shrinkage":
What factor should we use for calculating longitudinal shrinkage in hard maple strip flooring?

I do not have any good ideas other than to use the traditional number of 0.2% (green to dry) which amounts to about 2-1/2 inches per 100 feet. However, since the shrinkage change occurs over a 30% MC change, then we could also say that the wood will shrink longitudinally about 0.1 inches per 1% MC change per 100 feet.

The problem comes when we consider the cross grain in the wood, plus any juvenile wood, all which can shrink up to 30 times more lengthwise. I haven't seen much discussion of this lengthwise shrinkage.

Professor Gene Wengert is Extension Specialist in Wood Processing at the Department of Forestry, University of Wisconsin-Madison.

From contributor T:
Was that wood acclimated to the job site before it was installed? Was the A/C on? Does the crown have a sealer coat on the back? It really looks like shrinkage of the wood. Too bad it's a nice piece of mahogany. I'm not too crazy about the butt joint, but either way it would have come apart, probably from a moisture change at the job site. You have a good picture of the grain pattern. Replace it and move on. No sense in trying to pass the buck. Wood is a living thing and it does things like this. I hate to be the one to suggest this, but because you can see the continuous grain pattern in the wood, I would replace both pieces and move on. You will come out of it with your rep intact and probably get more work because you did the right thing.

From the original questioner:
We did not install the job, and have no way of knowing if they let the wood acclimate (I doubt they did). The A/C was on, and we always seal the back of trim.

The wood would not shrink that much (this gap is 1/8") for length. And if it was a moisture issue, the width would have drastically changed, and it is exactly the same as milled (6").

I'm not trying to "pass the buck" in any way. If I was at fault, we would replace it in a heartbeat. There is about 3,000 lineal feet of crown moulding on this job, and I'm not about to replace it when I'm sure we are not at fault. Wood does not move that much in the direction these gaps are appearing. If it had, the material would have been green, and we wouldn't have been able to machine and finish it. The gaps are not appearing throughout the space, either - just in about 1/4 of the joints.

From contributor C:
I just realized that you mentioned you didn't install it. Don't know why I missed that on the first post.

At any rate, were you able to observe the job before the joint opened up? Or can a neutral third party confirm that the joints were tight to begin with?

Also - a thought... by any chance is the ceiling that the crown is installed on fastened to roof trusses? Expansion and contraction of the bottom truss chord is a well-documented phenomenon, and could easily create enough movement to open up a joint.

From contributor W:
I hate to tell you this, but it looks like shrinkage to me. I do think it will shrink that much. The crown will shrink more than the base. Heat at the ceiling, moisture at the floor. Also, I am not happy with the butt joint, though I do not expect glue would have helped much. As sometimes happens, your building must have still been wet when you delivered the material. No easy way to fix this.

From the original questioner:
I'm not sure I can verify that the joints were tight when installed, but they would had to purposely spread them apart that distance, which makes no sense to me. Something is moving. I'm sure it is not the wood, though.

The walls are metal studs in a commercial building. Concrete floor, and sheetrock both faces. The ceiling is drop-in grid. In my opinion, it is the ceiling that is moving. But I don't know enough about it to know for sure. What I do know is that it is not the wood shrinking for length. The width would have drastically changed, too.

From contributor C:
"The width would have drastically changed too."

I crunched the numbers. According to the Knowledge Base article above that contains Professor Wengert's comment on longitudinal shrinkage: "then we could also say that the wood will shrink longitudinally about 0.1 inches per 1% MC change per 100 feet."

Let's apply this to your situation. Assuming the length of two pieces of crown that meet is 10 feet (each), his comment above would translate into the crown shrinking 1/100th of an inch (.01) for each percentage change in moisture content. In order for each piece to shrink 1/16th of an inch (one half of the gap), there would have to be a moisture content swing of over 6%.

I used WOODWEB's wood shrinkage calculator
which is for *width* shrinkage, and ran the scenario through the calculator, using 6 inches for the width, mahogany for the species, and a moisture change of 6%.

The result was a predicted shrinkage of 1/16th of an inch in width, a shrinkage amount that would be tough to confirm in the field.

But this triggered what I think is a more important question: are there any examples of two *short* pieces pulling apart, or are the gaps only limited to long runs? If the gaps are limited to long runs only, it seems this weighs in favor of crown shrinkage. If the gaps also occur in short runs, this would seem to me to favor building movement.

From the original questioner:
If, in fact, the crown is shrinking that much in length, why not the base that came from the same material? And why is it not consistent for all the crown?

Another point I forgot to make is that this mahogany was finished in our shop in the Midwest during the winter (with our shop heaters drying the air out). The material, if anything, would have been much dryer in our shop than in the field (where they have humidity control). If anything, I would expect the material to expand, not contract.

To answer your question concerning material lengths, I didn't catch if this was only occurring on long pieces. What baffles me is that it is happening on a few joints, and not all of them. On a run of crown, two joints might be fine, and the third is spread apart.

From contributor C:
Reviewing all that's been posted, I see things narrowed to these issues:
> I'm not sure I can verify that the joints were tight when installed.

> If, in fact, the crown is shrinking that much in length, why not the base that came from the same material?

> And why is it not consistent for all the crown?

> What baffles me is that it is happening on a few joints, and not all of them. On a run of crown, two joints might be fine, and the third is spread apart.

Assuming that the material came from the same run of lumber, and the moisture content was acceptable at the time of delivery, it seems that there are numerous supporting arguments that shrinkage is not the cause.

This seems like a situation where who's right and who's wrong will never be agreed on. I don't envy your position: figuring out a way to get closure on an unresolved problem.

From contributor T:
I am not critiquing your ability. This happened elsewhere on the job? Looking at the picture - the gap is so perfect that I wonder if someone is pulling your chain? Being that you didn't install it, I would get on someone about that butt joint. That is not standard, and no glue to boot? I think you have a good case with the installer. I worked for Harry Rich in the ceiling division on drop-in ceilings around the world and I'll tell you that if that ceiling moved, that tile would have fallen out. I say the installer has got a problem. That is a perfect piece of wood - you can see it in the picture, no warping at all and there is no way it could shrink that much in 100 years. I don't think there is any way the GC can prove that your product is defective. Not only that, but it sure looks to me that the cut was made in the same board. How's that?

From contributor S:
First, the simple fact that we are looking at a butt joint in that location should indicate to everyone that the installation might be below acceptable quality standards. The crown should have a glued compound angle cut in it at the joint, not only for additional strength, but to also transition the grain pattern from one stick to the next. Are the joints still tight at any inside or outside corners in the rooms? Are those inside corner joints coped, or simply mitred? I would bet they are mitred, another indication of amateur installation.

The wallpaper paste story is another "tell" on the folks you are dealing with. Myself, I'd try to get the money I'm owed from them, if any, and tell them to bite me.

From contributor Y:
Many good theories, most all seem relevant. I have been on many trim installations over the years. Can wood shrink this much? Yes. I would agree that the work might not be up to snuff. However, if the wood was dry to specs when it was delivered, it might have been stored in a cold climate and not brought up to temp for a few days before install. Here in Vermont, we tell the contractor to let wood acclimate to building temp for one week.

I have seen the same thing happen with joints when these rules are not followed. Print this thread out and give it to the owner and GC and see what their response is. Having only supplied this trim in good dry condition, I think the issue would be with the installer and GC. Did they follow the trim specs in the job spec book?

From contributor J:
I see it this way - bad install. If I was installing this, I would have mitered the joint, glued it and glued a backer strip to the back of the crown, then made sure I was a little long and sprang the cope to the next piece and the corner of the wall. This way, if it had any shrinkage, I'd have been able to tell.

But the big problem is where the finger stops being pointed. I would tell the installer I'll get another length of crown for him and he can install it without each of us charging the builder. This way everyone is happy and it could end up being a good relationship in the long run. Just make sure that they install it the right way the next time.

From contributor A:
I was a GC for years before I started to specialize in woodworking. If the ceiling has expanded due to movement of the bottom chord of the trusses or ceiling joists, the gaps would occur on the walls parallel to these members only. As to the ceiling tiles falling out, that just wouldn't happen. The tiles sit on the grid approximately 3/8 of an inch. And if this is the problem, it is likely that seasonal movement will bring it back together again. I am addressing this problem on a job right now. A 30' gable truss is moving about 1/8 of an inch seasonally, making an unsightly crack at the ceiling. You will only know if this is the problem by waiting to see what will happen.

Why not offer to fill the gaps as good as possible in the meantime, using a soft, pre-tinted, oil base filler?

From contributor L:
They are looking for a fall guy, and they are not very good at it, re: the wallpaper paste. You did not do the install, so how do they expect you to redo it? I buy trim from HD or other, have it installed by trained monkeys, then go after HD because of a very poor install? To resolve? Maybe smack the GC with the great common sense stick.

From contributor R:
Definitely poor install with the possibility of delayed shrinkage in the building itself. I've seen this happen when crown is finished and/or stored in an unheated garage, etc, then installed in a building in which the heat is cranked on. Was the finish you used waterbase? If you're using waterbase material, you'd expect shrinkage if the material was not allowed to dry out long enough. As for why the baseboard did not shrink, it is probably warmer and dryer on the ceiling and the baseboard was probably popped in under pressure, cut 1/8 long, while the crown, not being coped, probably wasn't under pressure.

From contributor P:
You may want to check the MC of the wood and then the EMC of the house that it was installed in. We installed a built-in unit in a house that has the same problem with their trim and our door panels shrunk over the norm. The MC of our wood is at 7% and our shop has an EMC of 6.9%. So we measured the house and it has an EMC of 3.2%. Anything more than a 2% change within the MC of the EMC and you will get these problems. I went down in their basement to see if the humidifier was running. Guess what they don't have? The guy has a company in China that makes a misting unit that attaches to your shower head and he thinks it gives enough humidity for the house. A 72 degree house should have a humidity level of 40% in a perfect world. 30-50 is an average.

From contributor T:
Contributor J's suggestion is an amicable solution to the problem. Get the GC a couple of new boards and he'll install it correctly and everyone charges each other back. Just sounds like good common business sense.

From the original questioner:
If it were just a couple boards, they would have already been re-made. This is probably 20 joints or more, so it's not a small deal.

From contributor R:
Let's get specific - what kind of framing? Steel stud? Wood? Climate control? What type of heat was used during construction? Open flame such as propane? What temp was the building kept at until the molding was installed? What temp after? How long after painting was the molding installed? What temp was the molding stored, and for how long? My guess is that the building was slammed up quick, then drywalled and painted with almost no time to properly dry out. The drywall was still slightly damp, the paint didn't help, the molding was thrown up - no scarf joints, no glue, mitered corners instead of copes put in under tension. This was probably followed by the full heating system being turned on for the first time - maybe there was a dry, cold snap at the same time. Without a properly working humidistat, the humidity can drop to 10% easily! Anything will shrink! Bottom line is, how can a company that's just providing moldings possibly be responsible?

From contributor H:
Well, if you decide to do it, how much is it going to cost? My guess is $800. That's about $0.42/hr over the next 12 months. This is an overhead expense and should be built into your cost of doing business. All future jobs have to pay a little more now. I don't think it should be your problem, and make sure you eventually get paid for it. If you do it, of course.

From contributor U:
Ran into this some years ago - temperature difference between ceiling and floor was 7 degrees, and the RH was 10% different!

The comments below were added after this Forum discussion was archived as a Knowledge Base article (add your comment).

Comment from contributor X:
Your shrinkage problem might be the building itself and not your molding. In conventional wood framed buildings, the framing is not kiln dried and many times I have seen the walls shrink in thickness and look as though the crown has shrunk. In your case, the length might be affected if the building had long walls that are steel framed, which will move in length.

Comment from contributor Z:
I would have to agree with contributer A. The first place I would look is the tendency of trusses to camber up in the center. If the problem isn't as bad on the perimeter walls, and worse on inner walls, that would most likely be the culprit.

Comment from contributor V:
If the framing is metal studs, and the trim was installed before the HVAC was fully operational, and the install was in the winter, it is certainly possible that as the building temperature was brought up, the metal framing expanded, thereby causing the longitudinal stretch.

A lot of ifs, but the only reasonable culprit is the metal expanding wall dimensions. There is no way that the moulding is shrinking along the grain.

Comment from contributor Q:
It seems to me that if the molding did shrink after installation, there would have been a much larger movement across the grain at the compound miter joints. This would tend to open the joint due to an actual change of the cut angles. If there was no complaint about the miter joints in the corners, I assume there was no change in the length of the molding.

Comment from contributor B:
I'm not a trim carpenter, but an architect. Here are my observations. If the open joints are only found in midspan, then the corners must be very soundly anchored and have overpowered the midspan anchorage, which I doubt is the case. So, for uniform shrinkage throughout the member (temperature or moisture), there should be gaps at either corner (molding end) that are each 1/2 the size of the midspan gap. It isn't going to "shrink" in the middle only! If these gaps exist, it implies (but not absolutely) movement possibly by water loss, or more likely, by a large change in the environment (temp, moisture) between storage and installation.

Also, I would investigate the anchorage of the molding. Is it anchored to the wall, to the ceiling grid or both? Decide which surface or system is the predominate support and analyze the condition from that aspect. Don't forget the ceiling grid is metal, which is very expansive and so are the metal studs. Depending upon the environment, these systems will move around and take everything attached with it. That is why specifications usually require that the building HVAC system bring the building temperture and humidity to a normal operating level at least a week in advance of hanging any gyp board and other materials. So check the specs and ask to see the daily log (if available, or check with the local weather bureau for a history) to get an idea about the conditions in which it might have been installed. Finally, did the contractor run auxiliary heating units to dry out the building quicker during the drywall installation and finishing process? This would throw an unusual amount of heat at the ceiling, making things move around, if nothing else.

Comment from contributor D:
Have I seen this type of shrikage before? Yes, often. What season was the shrikage noticed? My bet would be the begining of winter, with the heater turned on and the building truly dried out for the first time. Let it stand till spring and the gaps will close up somewhat.

I have run into this many times. One bank called up about the cracks in the crown molding (miter, glue and pinned) and they still separated. We puttied the cracks, applied another coat of finish and then got called back in the spring when putty was pushed out of the cracks.

The building and the trim must be climatized before any installation can begin, otherwise no warranty can be given. How much water is in a 5 gallon bucket of sheetrock mud? How many fives applied in the building? All that moisture does not go strictly to the surface. Was the rock delivered in the rain? Was it stacked on the second and third floor before the roof was completed and windows in, and then hung later? Many commercial buildings will not turn on the heat and/or air until the rock is sanded in order to save the units from the dust.

And you are always asked to install just as the drywall sanders are through. We are always installing when they are still hanging on the next floor. "We have to get the job finished. If you don't do it we will get someone else." I simply send a certified letter to the general contractor stating that all warrranties are void if we are told to proceed in an unclimatized building (as stated in the specifications and contract documents). No one worries about the letter. They tell you to go ahead and put it up, but when warranty issues come up, that $3.25 and five minutes of your time answers all the questions! And nine times out of ten, they will pay you to repair the problem.

Comment from contributor E:
If the wood was 7% MC, installed in a cold house built with metal studs, had moisture (from taped joints, tile jobs, paint, rain, fog, beach, plaster, or beer parties) and the house was subsquently heated, the metal framing would expand some and the moulding, after absorbing moisture, could have shrunk slightly. The butt joint is a lousy joint. A scarf joint (45 degrees or more) is better. However, when gluing the joint, use glues that withstand 160 degrees temp without cold flow. Regular polyvinyl glues stretch under tension (cold flow). Use an aliphatic glue (only some of the yellow glues are). They are sandable, also.

Comment from contributor F:
Not sure, but by the picture the moulding looks to be about 8" in depth? If so, that's a considerable amount of wood and shrinkage that is possible. More than likely it's not attached to the ceiling tile and only the wall. I would agree with the architect in that approach that the corners should also show gaps, butt joints or not. Not all pieces would shrink. Density would be close in the same species of wood, but most would know, not all are. Sometimes you have to accept the fact that wood does this now and then regardless of how much climatizing is given to the room/area.