Official Luthiers Forum!

I did some testing:
Woods used: italian walnut and pearwood, the first is open pored, the second close pored. Pearwood had some fiddleback figure in it. 4 blocks each. None of these woods is oily.
Surface preparation: both planed with a fine tuned veritas smooter plane. Two pieces of both pearwood and walnut have been sanded together (to ensure flatness and homogeneous results - they were paned together too) with a flat block with 120 grit sandpaper.
Glued with Titebond Original and kept clamped 24h. Blocks were 200x30x10 mm, grain running lenghtwise on with the 200mm side. They have been glued with 70mm overlap lenghtwise, glued on the 10mm face to ensure I could provoke a joint failure pushing the extremities with clamps.
After 24h from gluing, I clamped one side to the bench and pressed the other side with clamps until joint failure.
Results:
walnut: 120 grit: substrate failure (wood snapped along fibers); planed surface: substrate failure (wood snapped along fibers);
pearwood: 120 grit: substrate failure (wood snapped along fibers); planed surface: glue joint failure;

My PC is refusing to recognize my camera memory card, as soon as I can I'll post some pictures I've taken of the procedure.
I'm not writing this to prove anything, it's not meant to be a scientific test as I used only 2 wood species and didn't make many trials, I just wanted to share a simple try I did. Anyone would better try themselves with the methods they prefer as results may vary using different procedures and species.

My conclusion is that some degree of glue absorption is needed for the glue to achieve the best possible bond, or at least something for the glue to "grab" on, even if it's at a microscopic level. A planed surface (or scraped, for the matter) is probably less prone to absorb glue as surface will be smoother and fibers will be pressed down somehow, even with the sharpest blade (it's just a mechanical result of this kind of cut). Surely it's smoother though, at least at a molecular level.
Walnut is porous and it may be enough to ensure a sufficient glue, but it's also weaker than pearwood to stress along fibers so the failure at wood fibers isn't necessarily due to the joint being stronger.
Pearwood is a close-grained wood. When being planed it gets very smooth, a nearly flawless surface. It should be the best possible if titebond just required chemical attraction at molecular level, but in this case (and this alone, as I said it's not a scientific test) it didn't prove right: the planed surface on pearwood have been the only one to have a glue joint failure.

Moreover, if I think of obtaining the perfect match (for gluing) on a curved surface, I find that it's easier to achieve it with a radius block with sandpaper on it than scraping it by hand. With a careful work the curve can be close to perfection when done with a radius block. Considering that a scraper works with a curved edge (curving it when you push it to cut), it may be harder to get a perfectly level surface on a curve made with a scraper, even more considering the smoothness of the surface obtained (that enhances any imperfection in relation to perfect surface contact, possibly leaving gaps where a curve from the scraper cut pattern is left). I'm not saying that it's not possible to achieve a perfect surface (the best surface flatness is achieved with hand scraping on reference surfaces, indeed), I'm only thinking about which one is faster given what specs of surface flatness have to be achieved.

The conclusion is that, considering how I work and the woods I use, I find safer to sand the surfaces prior to gluing.
You may want to test your method too (and I hope you will and share your results! ), I think that knowledge is often better when achieved with experience... (not always anyway!)

Best regards!
Marco

Marco, thanks for the effort. But would you try the pearwood test again, at least 2 more times? You could also try more relevant wood combination such as spruce on spruce and spruce on rosewood.

_________________
Build log

It would be interesting to know the amount of force in marco's experiment required to make the break. If for example the planed joint which broke on the glue line took 40 pounds more force before the break then it's up in the air once again as to what technique is better.

I just glued 2 more couples of planed pearwood to see what happens... results tomorrow!
In the meanwhile my PC finally decided to acknowledge the memory card, so you can see the pictures below.

Given the shape of the test piece, the pressure applied was the force needed to make these pieces fail along fibers. In this case I clamped one side to the bench with one of those clamps shown in the picture below, then I applied a clamp on the side that was hanging higher on the bench and applied pressure until the pieces snapped. To be more clear, the pieces was clamped so the arrows on the sides with writings on them where pointing at the floor (for one of the pieces of the couple glued together) and to the ceiling. Dunno if it's clear, I'll eventually take a picture of the pieces getting snapped tomorrow.

I don't have scrap pieces of hard oily woods right now (not thick enough at least). When I'll have to make a strong joint on some particular woods combination, I'll surely test again with the woods to be glued. I had some olivewood (hard as pearwood and close-grained too, but much more oily) that could have been interesting to test, maybe if I can find some scrap pieces around the lab I may test it soon...

I tested the remaining planed pearwood blocks: one joint failed along wood grain, much alike the 120 grit sanded from before, another was in between those with a bit of glue failure and more wood grain failure than the planed one on the first test.
It seems it also depends on grain orientation, as the glue joint tends to fail where the grain is on average parallel to the border (average because being figured the grain waves noticeably, as you can see from the pictures), while with average grain runout both times it snapped along grain in these cases. Probably the waving patterns getting in and out the joint surface don't allow a full length wood grain fail, even if in some points (especially on the first pearwood test) it just seems basically a glue joint fail.
My personal conclusion is that I'd better do some other test with the specific woods I'm gonna use when I'll have to glue up some oily hardwood, possibly reproducing the exact conditions of the final joint. Overall, I doubt that fresh sanded surfaces that perfectly match each other will ever give problems with joint strength, but I'd rather be on the safe side.

Maybe it's just me but I find a repair business' discussion of a fellow builder's guitar on an open forum to be in bad taste. I would hope that there be at the very least an implied professional consideration of confidence to be honored. The issue could have been reviewed and the point made without naming the builder. And in light of Kevin's recent death makes this post even more disrespectful toward a man who gave so much and was a friend to many on this forum. I hope it's not just me.

_________________
JJ
Napa, CA
http://www.DonohueGuitars.com

Mr. Donohue your criticism is both ironic and misplaced...

Regarding professional curtesy and respect for fellow builders and more specifically the recently deceased Kevin Gallagher please be advised that "Gallagher Guitars" is not Kevin's brand... he was Omega guitars. I speak of a Gallagher guitar in my comments not a guitar made by my friend Kevin. Anyone in this business should know that Kevin's brand is Omega.

Further you speak of "implied professional consideration of confidence to be honored" which is also irronic if not phony as hell coming from you. I seem to recall posts here on the OLF and other forums in which you self appointed yourself the quality police sitting in judgement of the work of others when, well... sitting in judgement of the work of others...

Here is an example of your very own words:
http://www.luthiersforum.com/forum/viewtopic.php?f=10101&t=26187&start=0&st=0&sk=t&sd=a&hilit=shunned&p=352184&view=show#p352184

Where is the "implied professional consideration of confidence to be honored" in self appointing yourself the one who will "expose and shun" builders as you see fit.

Truth be told I stopped being your friend nearly 1.5 years ago largely because my experience with you was that you keep looking for bad guys, conspiracies, and ill intent in others to the point where I just could not deal with it... Now you wish to shun me for reasons that have nothing to do with my words here in this thread as you have quoted and responded. You do know what your motives are and so too do I am others - please stop with the games....

Please don't attempt to do with me as you have with other builders when they avoided you like the cable guy as I now do too.

Many thanks, bro....

I too read Hesh's comment about Gallagher; I didn't believe he was talking about Kevin Gallagher, but rather I believed the reference to be Gallagher Guitars of Wartrace Tennessee.

I don't believe I would have mentioned Gallagher's name, but it's already done, and giving Hesh the benefit of doubt, I don't believe he intended any ill will.

For what it's worth, I have worked on a couple of Gallaghers ( the Tennessee Gallagher) and both were, in my opinion, fine guitars.
regards,
jack

-

Gallagher or Gallagher, Psycho-drama and personal attack aside...the point is:

Is it good practice or ethical for a repair shop to reveal a builders name in an open forum discussion about a serious problem that's in for repair? The repair issue could have been reasonably discussed without identifying the builder. Furthermore, would it also be proper to call the builder and describe the issue he's seeing so he has an opportunity to both be involved in the repair as well as make corrections to his building process if needed? I personally would appreciate a call and heads up were it one of mine to find its way to a repair person. It just seems like the decent thing to do.

_________________
JJ
Napa, CA
http://www.DonohueGuitars.com