Raised Glue Lines after Sanding and Finish

Moisture in glue can cause swelling of wood near the joint, which may result in noticeable lines at the joint (depending on the timing of sanding and finishing). October 19, 2014

My last two projects have yielded glue lines that appear cracked through the finish. I use yellow glue and Enduro Pre Cat waterborne finish. I fill with Crystalac. This process has produced excellent results for the last several years. Now these cracks appear on some of the glue lines. Some are as they have always been. It is as if the finish is repelled from some of the glue lines. Is anyone willing to share how they solved this problem?

Forum Responses
(Furniture Making Forum)
From Contributor J:
Is the glue line protruding? Could it be spots that just weren't sanded fully? I also wonder if you've got some contamination going on? Did you use a tack cloth or something that might have contaminated the surface at the glue line?

From the original questioner:
No tack cloth used. I tried putting the corner of a razor blade into the glue line and it went in a very small bit as if the glue line was open. A 10X loupe shows the finish piled up along some of the glue line. I wonder if I have some starved joints. I would have thought that the filler would have taken care of that.

From contributor K:
My take on this is that the water in the glue expands the wood along the glue line, which would cause it to be a little wider or thicker than the rest of the plank. You probably sanded it while those cells were expanded by the water. As that moisture absorbed into more of the boards and then made it back to EMC they would end up lower than the dry wood back from the joint. Then those cells which were permeated by the glue were locked or fixed at the expanded size, which leaves the glue line raised when those cells just back from it contracted. I'm not sure how to avoid this without knowing more. How long between gluing up, before sanding, etc.?

From the original questioner:
For table tops I let the yellow glue set for 12 hours.

From Contributor P:
I had similar problems in the past when using Titebond (regular) yellow glue. Once I started using Titebond II, the problem never re-appeared. I don't have a scientific explanation for this, just my experience.

From Gene Wengert, forum technical advisor:
It is possible that the sanding operation is hot enough to melt the adhesive and then this more liquid adhesive runs into the wood a bit, creating a slight void. TB II is much more temperature stable, so it would not do this. I have also seen this occur when the pressure that is initially applied to a joint is reduced a little bit (probably by mistake or malfunction), so that the initial pressure squeezes out the excess adhesive and then the slight reduction in pressure (for whatever reason) creates a slight void and there is no adhesive to fill the gap.

From Contributor C:
Yellow glue tacks enough to remove most clamps in an hour or two but it doesn't reach full strength for a day or more. The moisture introduced by yellow glue doesn't equalize for multiple days. I have had joints that sunk when I have tried to do final sanding less than a couple days after a major glue-up. Do I cheat and rush sometimes? Don't ask. But I pay for it with less than perfect surface and joints.

From Gene Wengert, forum technical advisor:
Certainly some formulations of PVA (white or yellow colored, as the color has nothing to do with the properties of the adhesive) that cures enough to hold the joint together after two hours. However, there are some formulations that will not do this. Also, should the individual pieces not be perfectly straight, but the clamps bring them together, the added stresses on the joint from this straightening can be more than the two-hours-old joint can withstand and so some movement and opening can occur.

From the original questioner:
I got the errant glue lines to fill in by running a razor blade along the glue line through a wet coat right after spraying.