Problem gluing EI Rosewood

[Woodsy23]

I am building a Trevor Gore falcate steel string as my second acoustic build and encountered a problem with the glue joint in the East Indian Rosewood back plate. I joined the two halves using Titebond original and completed the final thicknessing to 2.2mm then cut out the shape. While I was sanding the rough edges of the back, the two halves virtually fell apart.

I had heard about the oiliness of EIR and potential problems gluing. Before gluing, I wiped the edges of the plates with acetone. There was enough glue as there was some squeeze out all the way along the joint.

I would appreciate any advice about what has gone wrong and how to achieve a good glue joint in EIR. Should I use epoxy? Will any special gluing technique be required for the back braces?

Also, I am concerned about the glue joints for the top and back linings which are already complete. They seem fine.

[peter.coombe]

Never had any problems gluing EIR, but also never wiped it with acetone so maybe that is the problem. How old is your Titebond?

Don't use epoxy!

Peter

[simonm]

I haven't done a lot of EIR - maybe 4 - but I haven't had a problem either. I always shoot the edges again with a plane shortly before I glue up. While I don't open a fresh bottle of titebond for each build, I won't use it for critical joints if it has been open for a long time. I also keep opened bottles upside down. Some time back I decided I would only buy smaller bottles (8oz).
I always rub the centre joints before I clamp to make sure everything has some glue on it.

[Trevor Gore]

Wiping with acetone can and does work provided that all the residual oil and colour is fully removed from the joint area. However, if not done thoroughly, it is easy to just bring more oiliness to the surface, which can make matters worse. An alternative is just to glue the very fresh surfaces immediately off the plane (within 5 minutes, preferably) and that seems to produce a good, strong a joint. I've used both methods and have never had a failure joining EIR with Titebond (and I've done plenty!), but I find the second method more convenient.

Likewise, I've glued lots of braces and linings to EIR using Titebond and never had a failure. Head and tail blocks I've glued in using Techniglue (structural epoxy) and not had any problems there either, using a variety of "cleaning" techniques. These joints have much more surface area, so arguably are not as critical. If your EIR has been left a while prior to gluing, the surfaces should always be freshened up with a plane, scraper or abrasive paper where you have decent access to do that. I clean off the dust with a microfibre floor wipe. For less accessible areas, I swab down with acetone or metho.

The joint that gets routinely destructively tested is the back panel marriage strip, which gets cut away where the braces cross and at the ends where it butts against the end blocks. It always amazes me how hard the glue is and how difficult it is to get the marriage strip and glue off the EIR. Anyone who thinks Titebond doesn't set hard enough should try that test! The glueline is always a lot harder than the EIR, so I don't think anything harder is required.

[Bruce McC]

When you joined the plates is it possible that you applied too much pressure and starved the joint of glue?

[kiwigeo]

I've never had issues with Titebond on EIR. Id suspect starved joint as suggested above.

[Woodsy23]

Starving the joint of glue may indeed be the problem. I used back to back wedges to apply pressure to the joint and squeezed it hard to get a close joint. I can do some trial joins with the off cuts, applying less pressure. Also, I think that while cleaning with acetone, I stopped while there was still colour coming off. If I can get a good photo of the failed joint, I will post it.

I'll also take up the suggestions about preparation of the surface to be glued (planing, cleaning etc) when I re-glue the two halves.

Thanks everyone.

[kiwigeo]

I haven't done a lot of EIR - maybe 4 - but I haven't had a problem either. I always shoot the edges again with a plane shortly before I glue up. While I don't open a fresh bottle of titebond for each build, I won't use it for critical joints if it has been open for a long time. I also keep opened bottles upside down. Some time back I decided I would only buy smaller bottles (8oz).
I always rub the centre joints before I clamp to make sure everything has some glue on it.

I take the same approach.....these days I generally buy my Titebond in smaller bottles and after 6 months the stuff gets relegated to jig glue up duty or if its more than 12 months old it gets turfed.

[Bruce McC]

Woodsy

Check out this post re Titebond glue and shelf life.

viewtopic.php?f=1&t=6405#p72311

[Trevor Gore]

I seem to recall a guy on the OLF who used to work for Franklin (Titebond) in their tech dept saying it was virtually impossible to starve a Titebond joint. I think the argument ran along the lines of having properly wet out the wood of the joint (difficult with oily contaminants) the water moved quickly into the wood, raising the local viscosity of the layer of glue immediately adjacent to the wood, to the extent that it wouldn't readily flow, so there was always enough left even when squeezed hard. The guy iirc had been retrenched by Franklin, so had no specific requirement to say that. So long as you don't thin the stuff, I'd be inclined to agree with him. I've starved epoxy joints in my time, but so far never a Titebond one.

Assuming a good fit and all of that, I would be suggesting contamination as being the most likely cause of the problem.

One other thought...if you close a joint, find it out of adjustment, open it again, adjust and close again you may well end up with a compromised joint.

[charangohabsburg]

Similarly like Trevor I also think that it would be rather difficult or even impossible to starve a Titebond joint using any plate joining jig.

Here, Franklin recommends for their wood glues a clamping pressure of 200 - 300 psi (ca. 14 - 20 bar or kg/cm²). Starving a Titebond joint would require more than that, which would mean to apply more than 400 kg on a 4 mm wide and 500 mm long back panel joint. That sounds like a lot of pressure. When stepping on a 10 X 30 mm piece of balsa wood I crush it completely (I weigh 60 kg => pressure on balsa piece is 20 kg/cm²). An equally dimensioned piece of balsa wood only gets slightly dented when put between the wedge and a 4 mm back plate in my joining jig.

[Woodsy23]

Woodsy

Check out this post re Titebond glue and shelf life.

viewtopic.php?f=1&t=6405#p72311

From the information in that post, the Titebond I used was made in October 2012 - just over 2 years old. It may be getting a little old but I don't think that is the problem. Here are a couple of photos of the failed joint.

Clearly, there is not enough glue left in the joint.

I seem to recall a guy on the OLF who used to work for Franklin (Titebond) in their tech dept saying it was virtually impossible to starve a Titebond joint.

Perhaps in this case, both reason offered for the failure are correct. Because the gluing surface was not prepared properly (by planing or cleaning thoroughly with acetone) and therefore remained oily, the wood was not "wet" by the glue. Therefore, when pressure was applied (possibly too much) most of the glue squeezed out.

I have just tried two test pieces. Both were joined with much less pressure. With the first, I kept the original surface but wiped it thoroughly with acetone. With the second, I planed fresh surfaces but did not clean them with acetone. This is effectively trying the two methods that Trevor suggested but I also applied less pressure so it will not prove either cause conclusively. I will carry out an un-scientific "break test" tomorrow and post the results.

[Hard Road]

Does anyone moisten the edges to be glued with a damp cloth just before applying the glue? Back in my furniture making days we were told to do this with white glues as the damp surface would allow the glue to penetrate deeper.

[Trevor Gore]

Whoa!

Looking at that glue join, to me it seems that the joint was not closed. i.e. did not fit properly. It looks like there are gaps in the glue rather than any wood pulling out. A good Titebond joint should have such a minusculely thick glue line that you can barely see it. There should never be any gaps of that depth in the dried glue film. So it's not a case of not enough glue, rather not enough fit. Put another way, if you plane parallel to the joint with a very fine cut, you should see wood shavings, then wood soaked in Titebond, then wood soaked in Titebond on the other side of the joint then wood again. You should never see a "shaving" of just glue. From the way the photos look, I would be getting a few pure glue shavings from that joint.

The bubbles that can be seen form when the water phase of the glue evaporates away. The glue shrinks and leaves voids like you see if there is a space for voids to form. But there should be no space for voids to form on the macro level. If they do form, suddenly half your gluing area has disappeared and you're looking at the cohesive strength of the glue rather than it's adhesive strength. And what happened happens.

Here, Franklin recommends for their wood glues a clamping pressure of 200 - 300 psi (ca. 14 - 20 bar or kg/cm²).

The same guy I mentioned before was asked why Franklin recommend such high pressures. His answer was along the lines that Franklin assumed a certain surface roughness (different from fit) then figured out how much pressure was required to flatten all the "lumps" in the surface so that there was surface-to-surface contact everywhere. The 200-300 psi was the answer. There is no way that most guitar builders use anything like that pressure, yet most people get joints at least as strong as the wood. Which I guess means they are able to produce very conforming surfaces that don't rely on high pressure to bring them together.

[peter.coombe]

Clearly, there is not enough glue left in the joint.

I disagree. There is plenty of glue there, but the surfaces are not meeting properly. I agree with Trevor. The reasons they are not meeting properly is likely to be due to poor fit, contamination, or the joint has been joined, taken apart and then re-joined. To me it looks like a typical re-join, but then again it might not be. Also, chuck your Titebond in the bin and get a fresh batch. It might be ok, but it is not worth taking the risk.

I always glue immediately after planing the surfaces, so that may be why I never had any problems.

Peter

[Woodsy23]

Thanks Peter and Trevor for the additional advice. I will re-prepare the surfaces and ensure that I get a good fit before regluing with new glue.

I carried out my non-scientific break test on the two samples I mentioned previously. Both broke along a grain line, not through the glue joint. The fleshly planed sample broke some distance from the join (through a weaker part of the timber) so I was able to break it again. This time it broke along the glue line for half the join length and throught the wood for the remainder. The glued surface that was exposed by the second break looked very uniform, nothing like the pictures I posted above. The pieces tested were very short (60 to 70mm) so wouldn't reflect any fit up issued that were possibly the problem with the 500mm long back pieces.

Not much can be concluded from these tests other than the fact that satisfactory joints can be achieved without excessive pressure applied. Good fit up, proper joint preparation and fresh glue are all important so I will address all these matters for future joins.

[Hank]

Hi Woodsy

Check your shooting plane, I've just starting my 4th acoustic build and while jointing the top with my trusty 39 year old No 5 plane could not get a tight fit, you know, perfectly straight not hollow or bowed, anyway I asked my business partner Peter who is a German master cabinet maker and knows everything
So it turns out that after 39 years my No 5 wasn't flat anymore
After about 6 hours of sweat,different cutting compounds and some voodoo she's amazing!
This luthier gig really relies on well tuned tools

Cheers
Hank

[Trevor Gore]

The guy I mentioned before, whose name I couldn't remember, is Hugh Evans. You might want to do a search around his name here to get some insight into how various glues perform, particularly the Titebonds.

[Woodsy23]

I have new re-joined the back pieces. During the preparation of the edges to make them straight I definitely had to remove more material from the ends of the joint than in the middle so Trevor's and Peter's assessment seems spot on. I used new Titebond too.

When initially preparing the edges, I recall that there was a small amount of light showing showing through in the middle, indicating a small gap, but I thought that would be OK, noting the advice in Section 6.3.1 of Trevor's white book about the leaving the edges slightly concave to allow for wood swelling. However, I guess the gap was too big.

Trevor,
Is leaving the edges slightly concave more important for the spruce top, which seems to me to react more to the moisture in the glue than, say, rosewood? Have you ever measured the gap created by the concavity using a feeler gauge? Maybe beginners should just aim for perfectly straight edges.