Official Luthiers Forum!

I've been building electrics for a while and Im currently working on two dulcimers as kind of an transition to acoustic guitars. I used titebond 2 for the bracing. After doing some research into hide glues, I read about how titebond 2 creeps and that its not great for instrument building. Im now nervous that the creep might allow the top to give a little under the tension of the strings. Should I re-do the bracing with titebond 1, or should I be ok?

Redo.
Titebond 2 has very few uses in instrument building and securing braces is certainly not one of them.

Steve

I think Hugh will be along soon (username hugh.evans). He used to work at Franklin Adhesives.

In the other thread about the various Titebonds, he says that 2 will creep but doesn't think it's a concern in the stresses related to guitars.

Kevin Looker

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

A Dulcimer top is not under a great amount of tension as the fingerboard carries the total load from the strings. I doubt you or anyone else will be able to tell the difference between TB1 and TB2 on it. It will be plenty strong as long as it is made like traditional versions. I shouldn't go here but the top does not function like a guitar top and is more like the bottom. The fingerboard is much more involved in the sound quality.

My ears were burning... You guys already got it right, though. There's no reason for concern in this case. Yes, Titebond Original is less susceptible to creep, but more importantly it is easier to steam apart for repairs.

Now if I can just get Elmer's or Ashland to offer me a job I can stop being the ex-Franklin guy.

What I don't understand is why is there ANY concern about creep? I have guitars I've made ten years ago with titebond 2, no creep.

Can you de-mystify the whole "creep" thing?

What are the facts......?

_________________
Old growth, shmold growth!

I would not be concerned. I used Titebond II in the first few mandolins I made when I did not know what I was doing, and as far as I know they are still going fine nearly 20 years later. I think I also used it on a few dulcimers and they are also still doing fine.

Glues like Titebond and and Elmers wood glues don't dry and harden crystalline like hot hide glue does. They retain an almost fluid like elastic state. As they age the joints can and will slip some. Maybe not a lot but it can be enough to move a bridge a few hairs to alter the intonation of an instrument. It can cause braces to move enough to effect the stability of the top. Ten years may seem like a long time but it's really not as it may take longer. I believe different conditions like hot/cold weather conditions and such can have more of an effect on it. Now this is just my two cents. Hugh can explain it better and probably will....Mike

_________________
Guitars, guitars and more guitars.

That's what I'm hoping, I'd like his take on it, since wood is itself an elastic material that is in and of itself prone to creep over time.

_________________
Old growth, shmold growth!

I would be happy to. Let's start off with the technical definition and go from there. Creep is the permanent deformation of a solid material resulting from a sustained load. All solids are susceptible to creep, and creep always increases with temperature. For the purposes of this discussion I will focus on the creep performance of thermoplastic polymers, since they describe the majority of adhesives using in luthiery. Thermoplastics are simply plastics that can be melted. When such a material is exposed to a load (stress) it will begin to stretch/deform (strain.) As long as the applied load does not exceed the proportional limit of the material, it will return to its original state when the stress is removed. The onset of creep occurs when the proportional limit is exceeded. Even if the stress is removed, the material in question will never return exactly to its original state. If the stress is maintained the material will continue to deform and will eventually fail, although depending on the material and variables such as the amount of stress and temperature fluctuations, it could take decades or a few minutes to actually occur.

A good real world example most people have experienced is a gallon of milk being carried in a plastic grocery bag. You have most likely seen the handles of such a bag either begin to stretch out or actually fail. The relationship between elevated temperature and creep can be easily understood by considering that as temperature increases a material approaches its melting point, even if it is still hundreds of degrees away. Plastics such as PVA generally melt around 350°F (although additives such as plasticizers can make this much lower) so a hot car in the middle of summer, which can easily exceed 150°F, combined with string tension has the potential to fail. I have personally tested PVA based white glues, that could be easily stressed to failure with hand pressure at 150°, while others routinely retain sufficient strength to cause extensive wood failure under stresses exceeded 2000 PSI.

One of the great challenges facing designs when using thermoplastic adhesives is the absence of useful strength data. If you browse the websites of a few adhesive companies you will find several references to ASTM D905, hard maple, and 100% wood failure at anywhere from 2500 to over 4000 PSI. The are several challenges in the process of testing the strength of wood glues, and the most obvious is that fact that wood is being tested... and its strength inherently varies by huge amounts. When testing an adhesive on blocks glued up from the same piece of wood, good data often varies by ±20%. There's a good reason I didn't quote numbers from the published data sheets while I worked at Franklin: on average, hard maple fails under shear stress somewhere between 3200 and 3400 PSI with 100% wood failure. Titebond III has a listed strength of 4000 PSI, while Titebond Original is only 3600. Both products have held up to over 4000 PSI, so has Elmer's Max, Gorilla Wood Glue, and Titebond II. Part of ASTM D905 is shear strength at 150°F, and this is a good indicator of creep performance. Titebond III is the only product in the wood glue line that does not have a published value at this temperature. I can tell you that when it is: the average is under 1000 PSI. Titebond Original and II retain about 50% of their room temperature strength, while Titebond Extend usually keeps about 80%. In all cases, when they cool back down to room temperature the adhesives return to their original strength.

What does this tell us about the true bond strength of PVA wood glues? Very little. The good ones are stronger than wood, and during a test that lasts less than 10 seconds they can withstand stresses of greater than 4000 PSI. ASTM has several methods used to evaluate structural adhesives. One of the most common uses a fixture that looks like a shock absorber into which a complex assembly of wood blocks is inserted, the assembly is loaded to a predetermined stress and a nut is tightened down to maintain it. Unfortunately, since it is specified as a pass/fail test, no one ever takes measurements while it is running to determine the rate of deformation... and failure usually takes several weeks to occur, so useful data could be obtained from it. By testing different loads, it would be possible to determine the true proportional limit of every adhesive and they could be treated as actual engineering materials. Guitar stresses could be used to establish minimum bond areas for given design safety factors. In collaboration with a few coworkers I developed a load and hold hysteresis method for pieces at elevated temperatures. Running the test over a range of temperatures would allow the proportional limit to be determined via regression. Unfortunately, for the time being, no one knows the true strength of any PVA wood glues. Beware anyone who claims to know the actual strength of their adhesives... they are either ignorant, liars, or both.

What we do know is that sufficient historical data exists to support the claim that high quality PVA wood glues will not fail due to creep when used to assemble musical instruments such as guitars and pianos. High quality PVAs should all contain some type of resin-based tackifier, and be minimally plasticized. Fillers such as wood flour (essentially super-fine sawdust) further enhance performance. Since PVAs derive their strength from huge numbers of weak electrostatic interactions at the bonding surface interface, tiny particles of wood effectively decrease the gap between bond surfaces. This increases bond strength and improves performance at higher temperatures because the adhesive is allowed to work as a binding agent in a system dominated by wood fiber, which is what it does best. Manufacturers usually won't tell you which of their products contain it: but a common sign is that bottles left sitting for several weeks will separate into two layers of slightly different colors but roughly the same thickness. Good mechanical design, which transfers loads directly between pieces of wood with minimal need for adhesive or does so via fasteners will remove adhesive creep from the equation. Good design, good adhesives, and good storage conditions will minimize any risk of creep. Unless you use the wrong glue (white "school glue" or TB melamine glue are good examples) your guitars will not experience creep induced failures.

As a side note regarding an extremely common misconception: glue creep is not a ridge of adhesive that protrudes from the edge of a bond line. This is instead the result of an excessively thick glueline that is able to expand and contract independently from the surrounding wood. The only remedy for this is to not have thick gluelines. In the case of PVAs, gaps should never exceed 10 mils (0.010") and thinner is always better. PVA joints, unlike hide glue, cannot be starved by clamp pressure. Therefore, clamp pressure should be applied as quickly as is feasible during assembly and with as much force as possible without damaging the wood.

I hope this explanation has been helpful. Let the questions begin.

I thought a couple of posts might slip in while I was writing my little tome, and I wasn't disappointed. Only adhesive performance was addressed in my explanation, while wood is the main component of a guitar's structure. Wood is, from a materials standpoint, the original high performance composite. Cellulose in the cell walls makes up most of the strength while lignin can be thought of as a thermoset structural resin (seriously, it even crosslinks in kilns which is a big part of the reason wood is more stable and rougher on tools after being dried this way.) If a top is made too thin, braces too light, or much heavier than intended strings are used, wood can effectively creep over time. With respect to bridges, especially when tops allow them to rotate towards the nut, the bond is gradually shifting from shear to tensile stress. If you've ever wondered why no one ever lists the tensile strength of wood glues it's because wood is extremely weak in the tensile mode. In a creep failure adhesive should be found in a uniform layer on both side of the failed bonding surface. There's a very easy way to see if this has occurred: pick up a bottle of first-aid iodine and apply a little bit to the bonding surface: PVA will react with it very quickly and turn a dark purplish brown (it's a permanent change so don't do this on any areas that will be visible after the repair.) Everywhere else it will have nothing more than a light yellow tinge, and it's okay that the bridge will likely be too dark to see anything because you only need to examine one side of the bond. You might find a couple of small dark spots, but I've always seen wood failure when a bridge pops off. Only if you observe a dark reaction over the entire bond area is it an adhesive failure resulting from creep.

Hide glue is completely different from PVA although they work on similar principles. In both cases bond strength comes from weak electrostatic interactions with the interface. PVA is a synthetic long-chain polymer that remains flexible when dry due to plasticizers, and emulsifiers added to improve handling properties during use. Without such additives it would not mix with water and would literally be nothing more than tiny balls of hard, tightly coiled balls of plastic. Hide glue is a mixture of long polypeptides (aka proteins) consisting primarily of collagen and elastin, which both serve connective functions inside of animals. Unlike PVA these peptides don't just like water: they are hygroscopic and will pull moisture from the atmosphere if given the opportunity. Water used in the preparation of hot hide glue works as a carrier, heat denatures the peptides... which is a fancy way of saying it makes them unravel, and therefore able to move around one another more readily. Additionally, as they cool the peptides begin to fold again spontaneously which is why hide glue seems to pull joints together on its own. Although hide glue becomes very rigid when it is fully set, it is no more crystalline than PVA. This is because crystals are extremely well-organized and regular structures. Big complex molecules such as PVA and polypeptides require extremely high purity and special conditions, that include a lot of time, for crystals to form. It's so difficult to make some polypeptides crystallize that in some cases they require the microgravity of space to reach such a state, and indeed such biochemical experiments are sent to the international space station for this purpose (in case you're wondering: I double majored in Mechanical Engineering and Biochemistry.) The structure is rigid yet amorphous. Generally speaking, protein based adhesives tend to have structural (creep resistant) properties, casein is still considered a benchmark structural adhesive.

Thanx, I copied and pasted this into a word document for reference when I discuss adhesives with my custys.

So I gather it's safe to say that with Titebond 1 Extend especially, unless the guitar is stored improperly, effectively any creep will be due to underbuilt wood components?

Do you have any data on temperature and it's effects?

Is there a particular "critical temperature" that a guitar should not rise above, IE, as long as it stays below 115 degrees then there should be no glue creep at all, or anything like that?

Also, I'm having trouble finding the Titebond Extend version. Who sells it?

_________________
Old growth, shmold growth!

That would be an extremely safe assumption. Especially when one considers that guitars use a variety of woods, many of which are softer hardwoods. I once ran a set of D905 blocks using alder as part of a project requested by a guitar company using Titebond Original and Titebond 50 (a product from the industrial line that is similar to Titebond Extend) and achieved 100% wood failure with both products at room temperature and 150°F. Give me a chance to look through some of my old files and notes. I think I might have some strength vs temperature data.

Thanks for the excellent answers everybody, especially Hugh. Quite impressive!

Hugh thank you so very much for your detailed reply abt titebond. Am gradually weaning my reliance on it in favour of other glues.

perhaps Hugh would be kind enough to critique/verify the following logic?

TII and TIIImakes an excellent joint as far as strength goes...can't in any way argue with that...I will go so far as to say that because of their slightly lower viscosity than TI it makes a better unclamped joint...

but, after picking up on the concept of cold creep and considering many things I've seen over the years, it makes an inferior joint in certain aspects...

1) it is not as hard as TI and therefore will not transfer vibrations (energy) as well
2) because of its comparative elasticity it will allow for more movement of joints as caused by normal wood expansion/contraction...I've seen this repeatedly over the years and will confidently say that TI shows less of this

let's consider the already covered concept that a good wood glue is stronger than the wood...taking this as fact then it follows that for a luthier the hardest dried glue is the best as this leads to better transference of energy

using TIII for gluing purflings to binding before bending is one exception to the above that I see, and I'm sure there are others

Prior to my removal from Franklin I was actively designing a series of experiments, with the intention of determining the effect of adhesives on acoustical vibration. No empirical studies have yet been conducted. I was the only guitar expert in the technical group, and since it was one of my undocumented pet projects I am quite certain no one has even considered the topic. What I can offer is anecdotal evidence and expectations of performance based on the established science of vibration analysis.

Ideal physical characteristics dictate glue joint with no mass and infinite rigidity. Since this is, regrettably, impossible we must instead seek the closest approximation. All of the PVA wood glues thrive on close proximity between bonding surfaces, and professionals can consistently achieve gluelines that are less than 0.005" in thickness. Even over a large area, the mass of the adhesive film is effectively negligible. Therefore, assuming the execution is perfect the only real variable left in this question is rigidity. The hardness of PVA wood glues is not typically measured, because in order to obtain the measurement requires casting a film that is thousands of times thicker than any glueline in the real world. It is also impractical because they typically have between 40% and 50% solids content, with the remainder consisting of water... So not only would it take forever to dry, it will also shrink and further complicate useful measurements. Fortunately, it's relatively easy to cast films of adhesive that are of appropriate thickness to feel the dried films. My favorite way to do this is quite simple, and requires the following materials:

*Samples of the adhesives to be tested
*Masking tape
*Release paper (wax paper, or silicone backed paper if you have it, both will do the same thing)
*Tongue depressors, popsicle sticks, or old plastic cards to smooth the west films
*A sharp edge such as an x-acto blade, chisel, knife, or shiv
*Measuring implement with a resolution of ±0.001" (optional but handy)

Cut off a sheet of release paper and start laying out a grid with the masking tape such that you form a rectangle for each adhesive. I personally like making them 2-1/2" to 3" wide and 6" to 12" in length, the key is to make them big enough to easily handle when dry. Masking tape varies quite a bit in thickness. As you build up layers to create wells for the adhesives use plenty of pressure so that variation in thickness is kept in check. Build up the same thickness for each well, 6 to 9 wet mils is considered the optimal average spread rate... But these will be less than half that thickness when dry and therefore can be very difficult to handle. I usually build up 8 to 10 layers, which often works out to the better part of 20 mils, that's on the thick side of acceptable when dry at somewhere around 8 or 9 mils (0.008" to 0.009"). Pour in the adhesive, spread it to reach all of the corners and edges, and then scrape over the top using the tape as a guide to create uniform depth. Do these one at a time, and when they've all been cast let them dry for 24 hours. They should all be fully dry after 24 hours and ready to be gently cut along the edges, which should allow you to remove them in-tact.

Now for the fun subjective learning experience. Feel how the films differ from one another, gently bend pieces while gripping the same length of material until them snap. Warm a small area between your thumb and a finger before trying again (this will cause slight softening in some PVAs.) Also, try setting pieces onto a hard surface such as a countertop and tapping them with a fingernail, even try digging a nail into it to see if you can leave a dent. All the while, keep in mind that the film is very close to what is actually inside of joints assembled using it. The only difference is that the bond it forms with each surface is extremely tenacious. Most people have never tried this experiment, and will likely experience more than a few surprises.

Given all of the assumptions I outlined before, rigidity is the one variable that can safely be assumed to facilitate acoustical energy transfer between interfaces. You will find that Titebond Original is indeed more rigid than Titebond III (especially when warmed between your fingers.) Titebond Extend is even stiffer. With respect to creep, and therefore the thermal plasticity observed at 150°F, results from ASTM D905 can further serve as reasonable guidelines. Superior retention of strength at elevated temperatures is a valid indicator of a more rigid bond line.

If you want to have even more fun, try laminating some decent sized but uniform scraps of wood. Make sure they are large enough to easily hear them ring when tapped. Most trained ears can not only hear subtle variations in resonance, under the right conditions you can hear the adhesive drying. This might sound crazy, but there's good science behind it: PVA wood glues take time to fully set as they release moisture into the surrounding wood and it gradually makes its way out to the air. Well over 90% of ultimate bond strength is achieved within 24 hours. The actual development of strength can be observed continuing to increase over the course of approximately two weeks. Localized increases in wood moisture content at the interface take up to six weeks to return to equilibrium with the surrounding wood!

I tried checking tap tones over time after being informed of such a result from a luthier over the phone one day. The correlation between what he heard and what is known about the curing process was too uncanny for me to not investigate. There was indeed a subtle but perceptible change in resonance somewhere in the 12 to 14 day range.

In summary, my personal favorite PVA for guitar assembly (and all general purpose woodworking for that matter) is Titebond Extend. I do not believe it is likely that there is a huge amount of tonal variation between quality woodworking glues. With the subjective tools for analysis I have provided, it is possible to evaluate the adhesives for yourselves and hopefully make informed decisions in the absence of empirical data.

Products such as Hide Glue, Titebond Extend, and Titebond Original all have excellent track records when used properly. Titebond II Extend, Titebond II, and Gorilla Wood Glue can also produce excellent results but can only be considered advantageous for large scale manufacturing in which RF curing is utilized. As a side note (and further evidence of my current lack of connection to any adhesive manufacturer) Gorilla Wood Glue is an exceptionally nice type II water resistant PVA. Not only does it have performance comparable to Titebond II, it also does not contain dyes and dries to a nearly clear film. I can't stand the liquid polyurethanes they make, or their marketing claims, but I always like seeing high performance PVAs that do not contain dyes... All "yellow" glues could be colorless and produce clear films when dry.

theguitarwhisperer--

Titebond Extend is part of the cabinet shop line of products and as a result is most commonly found in specialty stores (woodcraft and rockler), cabinet shop suppliers, and amazon. You didn't hear it from me: but everyone in the tech group at Franklin can enter orders for free samples at their own discretion.

viscosity....hmmmmmmmm...

was looking at the spec sheets of the Tit(e)bond products (and that is said affectionately as I've been using Franklin products for decades) and notice I am in total error on my statement about TII and TIII being thinner than TI....I swear they act thinner when applied and worked with, but obviously I am in error...perhaps its the fact that TI will show a quicker initial tack that makes it seem so...now that I think about it TIII certainly pours out of a gallon jug far slower than TI (have to admit I haven't used TII in over 4 years so I don't have a real good memory of that product)...either way I was obviously erroneous with my statement

The viscosity is pretty close between those products. TB III has a strong tendency to settle and separate, resulting in a very thin consistency. With that in mind it is always a good idea to agitate bottles of adhesive before each use. Shaking requires too much effort: turn the bottle upside down and rotate it as you tap it on the corner of a bench or countertop.

whilst it is edifying to get the details for the glues upper limits of adhesion,, etc., that is not the be all and end all as far as what needs to be considered when selecting a glue for instrument building.

without assigning an order of precedence, the following sort of summarize my primary considerations:

- does it detract from the transmission of vibratory energy, i.e. sound,
- does it keep things in place under stress, both physical and environmental,
- is it readily and non-destructively able to be un-glued for maintenance and repaIr, do want to be cursed by future luthiers for using amg like products,
- convenience and safety(all aspects) of its use,
- will it adhere to itself when regluing,... etc.

the list could be further extended with minor considerations,, e.g., glue line color.

there is no single glue which tops every consideration. there are specific uses where being best in only one may be enough to determine your selection.

anyone who has done much repair work knows that some glues can creep catastrophically under heat and tension whilst under the control of forgetful, careless or ignorant owners.

like so much in this business, one size dorsn't fit all, so we have to make compromises in glue selection as in most other thimgs.

all that being said, for me the only tb i would use is tb1, though i prefer to use lmi white. favorite is still hhg, evem though it doesn't score well on converience or ease of use.
-

I agree with you for the most part. Many of your concerns were addressed in another thread in which I discussed adhesive selection. If you have the impression that I consider any single adhesive can be used for all applications, then my opinions have been sorely misunderstood.

My professional opinion is that there is generally an optimal adhesive for any given application, and determining the most appropriate product was once a major part of how I made a living. I'm not going to argue about my level of technical expertise and experience in areas such as failure analysis, troubleshooting, repair, and restoration. I provided a simple test that is capable of proving the actual failure mode of a bridge that detached. If thermal plasticity is indeed the cause then an truly inferior adhesive was used, which would likely constitute a violation of several of your primary considerations. I have analyzed bridge failures in my personal life, and in a lab setting, which is why I stand by my observations of tensile substrate failure.

Compromise is generally not necessary, and never to a great degree. As stated before: I do not have an affiliation with any adhesive manufacturers aside from contractual obligations related to extensive knowledge of trade secrets. In no way does this affect my ability to provide accurate information and recommend the best products I am aware of for a specific application regardless of who makes them.

I still recommend that if you only use one variety of Titebond, it should be Titebond Extend Original. I have not used LMI white, although I am wary of it based on what I do know about it (most importantly that it is expected to spoil.) As soon as I get around to picking up a couple of bottles I will send one off for competitive analysis. With absolute certainty, I have stated in the past that hot hide glue's only true disadvantage is lack of convenience. Its reversibility due to direct contact with water can be viewed as a strength or weakness, depending on the situation.

If anybody does any of these tests, I would suggest including Elmer's Carpenters Wood Glue. My perception from using it is that it dries to a similar degree of being "crystaline" as hide glue. Dried glue chips off rather than bends, and does not seem to be at all flexible. Again, nothing scientific, just observation FWIW

Grant

I know I've tested Elmer's Carpenters Glue and Elmer's Max. The part I always had mixed feelings about, as a scientist, is that during competitive testing: fresh product from the warehouse is what was always used for the in house products. My personal feeling is that products should be acquired from store shelves, so everyone's product has had a chance to perform differently from exposure to various environmental conditions encountered during shipping.

I'll see if I have the results archived anywhere. Off the top of my head Elmer's Carpenters Glue was similar to Titbeond Original at room temperature. But I don't recall the 150 shears well enough to call it either way. In the worst case scenario some of us could band together and call manufacturers to pester them for copies of test results. The real trick is getting raw data. I always reported my results with 90% confidence intervals, which makes insignificant differences pop right out. Marketing people, and others who should know better, hated it because it made many of the results ambiguous. Unfortunately, the reality is that a lot if high quality wood glues perform very well and in a very similar manner.