Not sure the best way to describe this but here goes.
The joint line on some stuff that I have made is now not smooth. After glue up , sanding and smoothing they where perfect the two pieces of wood used where indistinguishable to the touch. I finished them and now about 4 months old I can feel a tinny ridge of glue that seems to be squeezed from the joint. What is up with this and how can I keep it from happening in the future?
Discussion Forum
Get It All!
UNLIMITED Membership is like taking a master class in woodworking for less than $10 a month.
Start Your Free TrialCategories
Discussion Forum
Digital Plans Library
Member exclusive! – Plans for everyone – from beginners to experts – right at your fingertips.
Highlights
-
Shape Your Skills
when you sign up for our emails
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply. -
Shop Talk Live Podcast -
Our favorite articles and videos -
E-Learning Courses from Fine Woodworking
-
-
Replies
Is it a ridge of glue, or could it be that one side of the joint has moved with respect to the other, because of uneven shrinkage or the wood? (See attached sketches.)
I think the latter is more likely, and is usually caused by having different grain orientations on either side of the joint.
-Steve
Saschafer
Its glue for sure. You can feel it when rubbing in both directions, and you can also smooth it out by dragging you finger nail over it and taking off the edge.
"...you can also smooth it out by dragging you finger nail over it and taking off the edge."
Aha. What kind of glue, and what kind of finish? Is it possible that you had a relatively thick glue line, and applied the finish before all of the moisture had escaped from the glue? If that were the case, the glue could have continued to shrink, eventually pushing the finish up into a ridge.
-Steve
Saschafer
Just regular old titebond 3 wood glue. As far as the glue not drying before it was cured. The first one sat probably 2 months after glue up before it was finished. The finish is just a danish oil with a wax coating.
I does seem like the wood has maybe shrunk a bit and the glue has squeezed out the top. but how to i keep this from happening?
"I does seem like the wood has maybe shrunk a bit and the glue has squeezed out the top."
I can't envision that actually happening. I can imagine the glue absorbing moisture from the air and swelling, or possibly even absorbing some of the finish and swelling, but it's not something that I've ever experienced. I'm afraid I don't have any brilliant (or even stupid) ideas at this point.
-Steve
saschafer
Ha thanks for trying. I guess we (mostly I but maybe you) should be working instead of worrying about wood anyway.
WT
When you get to the Titebond III level(waterproof, waterresistant.....) its new ground for some on this thread. I use Titebond original after having some troubles with Titebond II. If I am not using outside or in a place with some really unusual atmospheric conditions, I don't understand using T-III.
Could you say a little more about the joint: what are the woods used in the joint? How thick is the board? What did you finish the surfaces with before applying the oil?
Dan
The joint is for the top of a dresser. Its just under 2 inches think. And i din't treat the wood with anything before i fininshed it.
what the ,
Perhaps after several months of sitting around in an unfinished state the wood may have changed RH factors several times , warm , moist , cold .
Maybe the glue did not move just the wood changed .
Did you sand the top again before finally finishing ???
I have noticed some woods do require re sanding after much time has passed the grain gets a raised feeling .
dusty
dusty
Yeah it did sit around but i did sand it prior to finishing and it was very smooth. It changed once it was finished. The only think i can think is that the wood has shrunk slightly but the glue has not.
makes sense , maybe .
I think i am gong to try and light sand and then re finish it. Its a oil wax finnish so it should be pretty easy
At this point if it were me , I would scrape or heavens forbid knowing me , belt sand and re orbital or finish sand the piece to give the entire surface a new leveling so to speak .
Was this extra soft wood by any chance , sorry if you already said what wood it was.
dusty
I've always heard this problem described as glue line creep. I belive the glue nearest the surface actually absorbs some moisture and swells proud of the flush face. I had a bed with 3" bedposts that exhibited this along the glue lines. I think I used TB II or I. I have noticed that I have only noticed this on the bed and it is the only piece of furniture that I used a "Danish oil" finish on Maloofs poly/oil blend). Pieces with varnish do not seem to exhibit this.
Now that I think of it, I remember a rocking chair with cold laminated rockers that exhibited the same problem. I haven't had a chance to worry about it, but my next laminations with be with plastic resin glue.
Once I scraped the glue lines with my fingernail, they never returned.....I would not go so far as to sand, but a light scraping with a scraper or razor blade and another coat of finish and I'd call it.
Dan
wiskytango,
I've had a similar problem...the only other guess that has not been mentioned here is grain pattern along the joint. The theory in my case was differences on expansion/shrinkage along the joint and next time try and avoid gluing up flat sawn to rift sawn pieces....maybe
I've heard various theories for this:
-- The wood adjacent to the joint absorbs water from the glue and swells. You sand it smooth, and when the wood dries out again, the wood shrinks and the glue line is left to stand proud.
-- PVA "glue creep" causes it to keep on oozing. I think this is likely to happen when one follows the oft-heard but erroneous advice "only apply enough clamping pressure is get a little squeeze-out", rather than following the glue manufacturers' recommendations of ~150-250 psi clamping pressure (see this month's issue). Some folks also use only urea formaldehyde glue on panel glue-ups to eliminate this "creep" problem.
Some glues remain slightly elastic even after their full cure. Polyvinyl acetate (PVA) glue is one of the classic ‘creepers’. Under stress the glue line gives and parts move out of alignment. For example, bent wood laminations glued with PVA tend to straighten out long after their removal from the bending form.
Where the bend is slight and the stress on the glue line also similarly small the springback and long term creeping is often insignificant. Severe bends where stress occurs in both the glue line and laminations themselves show greater tendency to ‘springback’ or straighten.
Another form of creep involves moisture cycling, again with PVA glue. In this second version both the wood and the glue in a joint swell as they take on moisture. The wood later shrinks as it dries but the glue doesn’t and sits above the joint line. Where this happens to a solid wood glued up table top the result is a line of small pimples along the glue line that is just perceptible to the finger tips although, quite often, the pimples are also visible.
I’ve occasionally seen the ‘row of pimples’ fault where rather wet planks, 13% MC or so, were edge joined to make a panel which later shrank as the wood adapted to internal house conditions by drying to roughly 8% MC. Glues that never really cure hard and always remain a little elastic allow a little movement in the joint. PVA glues, being water based and affected by humidity sometimes allow adjacent planks to shift a little as the seasons change.
Also consider that mis-matches in grain orientation as seen from the end can cause steps at the join line. The sketches below give the general idea, even though I've rather exaggerated the movement and end result. The arrows indicate the primary direction of shrinkage or expansion, ie, it moves the most following the line of the growth rings although there is also, on average, about half as much expansion and contraction that follows the medullary rays (radially.)
View Image View Image
Adjoining boards in a table top can, and do, drift out of line in the thickness and form a step. Even the weight of a TV or computer monitor will do it with a foot placed next to a glue line and left in place for long enough. This is the case with PVA glue that creeps, see footnote on page ?.
For the reason just described, and in many other situations, many woodworkers like to include an alignment aid or reinforcement of some kind into the basic sprung joint, eg, biscuits, dowels, loose tongue, finger joint, etc. They all reduce the strength of the plain edge joint, but compensate for this loss through increased glue line length and mechanical locking. The alignment aids and machined edge joints work best when the joined edges are as carefully and meticulously prepared as if doing a basic sprung edge joint.
There is a further wood movement factor to consider in edge joinery. Water based glues such as PVA, hide glue, urea formaldehyde, etc, all introduce water to the glue line. This water is very localised, but it does cause wood to swell creating a slight ridge at the line itself, perhaps no more than 6- 10 mm (1/4”- 3/8”) wide. If you glue up a panel with PVA glue at 9 am, take it out of the clamps at, say, 12 pm and immediately get to work preparing the panel for polishing, and actually follow up with polish you’re likely to encounter a problem. A week or two after polishing you’ll notice a slight furrow tracking every glue line.
If you’ve used a series of glued in biscuits to reinforce the joints you’ll see a slight furrow, and evenly spaced roughly elliptical shaped depressions that match the position of each biscuit. The chances are that you’ll get away with this error on vertical surfaces because there is often little or no light glancing off the surface. But you won’t with table tops, and the glossier the polish the more the furrows and depressions stand out. Because of the way they’re manufactured of compressed beech wood, biscuits work in part by swelling with the water introduced during the gluing up procedure. This makes them fit tightly in the slot and can exacerbate any swelling caused by water alone.
Sufficient time is needed between introducing moisture to the wood causing it to swell, and preparation of the surface for polishing. The wood around the joint line needs two or three days to lose the added moisture and shrink down to match the level of the rest of the surface. I like to leave edge joined panels at least three days to settle, and a week is better still.
I have only discussed biscuits, but similar considerations are required when using other edge joint reinforcement or alignment aids, eg, dowels or loose tongues. Non water based glues such as epoxy resin or polyurethane types don’t cause the same problems, but these types aren’t always appropriate for the job.
That'll do for now. Slainte.
Richard Jones Furniture
Edited 10/4/2007 2:59 pm by SgianDubh
This forum post is now archived. Commenting has been disabled