I made a small batch (1/2 oz glue to 0.9 oz water) of Milligan & Higgins 192 gram strength glue. I followed Frank's quick method for making the glue. I stirred the granules and water together, let it sit for about an hour, at which point it looked like fish eggs. Then I cooked it in "double boiler" to 140 degrees. It looked pretty good and it had the consistency it should. It's thin but not too thin, it's somewhere between what and good maple syrup.

I decided to test it on two small scrap pieces of mahogany. They both have been sanded, so you don't really feel the grain so much when touching them. Anyway, glued them up (grain perpendicular) and clamped them overnight. I couldn't break the pieces apart with my hands the next day, but I was able to when I held one piece with pliers and tried to pry the other piece with another set of pliers. I should note that the piece broke not because I pried it, but because I put pressure perpendicularly to the grain, which made the piece "fold" and break. The break was very clean with no splinters or fibers sticking out.

Because I'm new to hot hide glue, this test seems inconclusive and I'm not sure what to make of it. So I wonder what are some ways people here use to test their glue?

Views: 1300

Reply to This

Replies to This Discussion

Thanks both of you. Rusty for the advice and Paul for putting into words what was bouncing around in my head. If you wan t something to stand up to pliers I think you need to to go to nuts and bolts.

Good article - I love tests with p-values! One unmentioned factor is shelf life, both for bottled hide glue and pre-prepared HHG (refrigerated cubes, etc). I assume that, if Winterthur is this methodical, they also have protocols for labeling, shelf-life, and disposal.

I've been thinking about the technique of applying HHG to each side of a joint before reapplying it and clamping up. 

 I haven't done this intentionally but I have repaired several open joints by loosely cleaning out the old glue, but not really worrying much about getting it all, then applying HHG and clamping. So far they have all worked just fine.

That said, I wonder if this is more about violin building than guitar building. In my VERY limited experience with violin family instrument, it seems to me that removing the top to make adjustments or repairs is not only more acceptable but actually expected in these instruments. Applying a layer of glue that sets up before the final glue up would probably make disassembly easier. The fact that hide glue sticks very well to hide glue would enhance this technique. From my (again) very limited experience it seems that removing the top of a violin made using this technique is much less likely to pull fibers of wood from either side of the joint thus making it MUCH easier to remove and replace which may be the point.  

In guitar building it's usually assumed that opening the "box" after final assembly isn't really desirable so we want a stronger glue joint. I think this is why a lot of guitar builders and repair person's use different glues for different purposes but, often tend to use HHG for the things that they know will are likely to need to be removed in the future, mainly the bridge and neck, even if they don't use it anywhere else. 

I'm not saying that pre-coating with HHG is a bad thing. I  haven't experimented with it enough to form any sort of functional opinion and I really have not idea what difference, if any, it makes to the strength of the joint.  I'm just pointing out that a violin and an guitar are not subject to the same stresses or the same expectations in durability so it may not be wise to assume that what works well on one will automatically work well on the other.   ( I might be talking through my hat too so take this with a grain of salt.)

Hi Ned, all,

Firstly, coating both sides of a glue joint evenly is good practice.   Only coating one side, especially with a fast set glue like HHG will give an uneven adhesion as the glue will bond differently to each piece in contact due to immediated cooling that takes place and the lack of penetration or wetting of the unglued piece.   

The article I used to demonstrate just "what is known" also iterates the need to evenly heat each piece to be glued prior to setting the joint - this further assists the even adhesion (normalising) of the joint and the ultimate strength.

But, it's not so much this particular article that is important here, it's more a case of luthiers standardising their practices and procedure in relationship to the things we do.   The glue manufacturers provide good information that largely goes unread as does the countless articles in the professional domain about factors affecting performance, prepping and actual controlled tests.  

There is no voodoo in the process of making a good glue joint  with any glue and accordingly, if the tech spec is followed and all the factors (selection of glue type, environmentals, cleanliness, accuracy of joint prep and clamping pressure/time etc) are taken into account, the joint will be good.

Bottom line stuff here is that making a glue joint the best it can be from a technical standpoint is good practice, especially for critical application we experience in luthiery.  The process of doing this requires knowledge and experience that can only be gained by reading and research and technique - simply getting two sticks to stay stuck together and using that as a validation is a poor way to approach the high quality, long term results we strive for.


Thanks guys, for your insights.  All well taken.


I understand and agree Rusty. My reference to glue on both sides of the joint was in conjunction with the idea of letting it dry before applying more glue and clamping up the joint.  When ever possible, I like to put glue on both sides of a joint too. I usually just do it once and clamp it up.

 I also agree with the idea of using the information available to make our decisions rather than trying to reinvent the wheel. Today, it's easier than it's ever been to find it.  It's often not the most entertaining reading but I usually learn something worth knowing (if I can stay awake, that is.) Unfortunately, my history is littered with half read technical papers and books. I now "sort of" know a bit about a very large number of things... most of which is pretty much useless to me   

Russell, thank you for attaching that paper. It was an interesting and an informative read. I've never looked up how much pressure hot hide glue can take. I did, however, spend time reading on hot hide glue in the books that I have, on, and several other websites I stumbled upon (I know how to use google). I also went to Milligan & Higgins' website, which is quite bare, to read on their hot hide glue, but there's no technical documents there. I had a similar experience on Behlen's website. I didn't go to Franklin's website, because to the best of my knowledge, they don't manufacture hot hide glue, only liquid hide glue.

I don't have any intentions to reinvent the wheel, and I didn't test my batch of glue because I don't trust hot hide glue to be strong enough for my applications. I trust the science, research, and literally thousands of years of experience that stand behind that glue. I just don't trust myself. I followed the instructions to making and applying hot hide glue to the best of my abilities, but I figured it's worth checking the strength of the glue I made. I mean, why not? It is my belief that if anything went wrong, it won't be something obvious, like the glue not adhering at all - it'll be reduced strength.

I will admit that I was a little haphazard in one of my steps - I grabbed two scrap pieces that were nearby and glued them together. My testing was inconclusive, because there was no way for me to quantify how much force I was applying to those two pieces with the pliers. Not to mention figuring out the vector sum of the forces applied. That's why I came here. I didn't know of a good way to test and quantify the glue I made. So yes, I wanted to hear if trying to pry the joint with pliers is dumb (and it is, so thank you for answering that), and I wanted advice on what are some other tests I can conduct. Doug provided me with one (thanks again, Doug!).

In fact, this thread is exactly what you're talking about, Russell. It's trying to find a standard for testing a new batch of glue. What's a good way (meaning it can be reproduced by others and quantified) for testing it, and what is an acceptable range of force (or weight) a joint should be able to handle? In other words, how do you translate an average of 4000 shear point PSI applied to a hot hide glue joint that was in an environment with 50% RH to terms the average luthier can understand and tests he or she can perform? I suspect no all luthiers know enough mechanics and math to fully understand these terms, nor to conduct simpler tests that verify these findings. I most definitely doubt most luthiers have access to lab equipment where they can repeat the tests described in the paper. Not to mention the time to wait to test a new batch.

P.S. Russell and Paul, while I appreciate the technical insight you brought to this thread, I do not appreciate you being presumptuous and dismissive towards me. It's offensive to assume I've done no research and to suggest I "Google 'Hot hide glue preparation and application'". It is also quite dismissive to say "The thing I'm 'not getting' is the OP's ultimate goal. Is this just aimless experiences shared or outcome based experimentation? A plan is easier to execute with a definitive goal in mind." This isn't the first time that you've showed this attitude towards me (see the end block thread), and I ask that you please stop. Perhaps you confuse my asking a lot of questions with a refusal to do my own research. Well, I do quite a bit of reading and research on my own. Have I read everything about woodworking or luthiery? Absolutely not, but I'm getting there. I ask a lot of questions because I want to verify the things I've read and learned. I want to make sure that the methods I use are accepted not just by any average joe on the internet, but by professionals too (i.e. this forum).


It is hard to determine an individuals ability and level of expertise on a forum.  I go by what I see and what I hear from the post.  If I see the level of knowledge to be below basic I will suggest that the poster go back and do a bit of study.  

It's not dismissive and you should not view it as such:  it is good advice and saves a lot of time being wasted conducting ab initio training here .   Cabinet makers and professional luthiers live and die by their skills in (glue) joinery and there is a lot of free, validated advice on the web that will enable you to understand the basic procedures, products, terms and applications for the various types of general use glues.

The inclusion of a study on some aspects of HHG was to demonstrate that the information required is readily available and importantly, to answer your question:  you do not need to do the tests or understand the parameters - you just need to be able to understand the results.

To test your glue joints is simple: glue them up in accordance with the glue makers specification/preparation advice  in a suitable environment with appropriate clamping pressure.   Hint, you can never over clamp a glue joint with a speed clamp.  

Allow the glue to cure/dry/set for the recommended length of time in a controlled temperature place and then break the joint in an orientation that the glue joint will be stressed in when it is in service.  (Note that HHG has poor shock resistance in the shear axis as per the study provided) .

Then stress the joint until it separates.  If it breaks without tearing out a stipulated percentage of the wood (ie a clean break with two smooth surfaces) you probably have a faulty joint.   If you get a percentage of tear out of the wood (which depends on the wood being glued) which is within specification , you have a good joint.  Which is good if you are a luthier or an aging Hippy.

This is advice which is general in nature but serves to get you onto the first rung of the ladder.

Eliya don't take this so  hard, you are not under attack you are under guidance and many will step up to reassure you and assist you at length as they have done previously.    I recently got a master class on <youtube>  from David Collins on guitar electronics that made my eyes water - despite myself having a qualification in military avionics and 20 years of making Electrics, he cleaned my clock with his understanding and depth of knowledge.  I am grateful to David for sharing this remarkable information.  Sometime we just don't know what we don't know.

Take it easy,


Testing glue joints like this is a noble effort, but when looking at the big picture of how it will perform in application, should be taken with a grain of salt.

It's an inconvenient fact that forcing a test joint to complete failure often does a very little to appraise properties over long term use. Testing to failure under temperature or humidity extremes may offer a bit more insight, but still neglect the critical real world factors of cyclical stresses over time (time is a critical factor with direct influence which is too often forgotten).

I can glue up several identical test pieces with hide glue, PVA, epoxies, superglues, and all sorts of others, and when forced to immediate rupture get surprisingly similar failures depending on the type of test. Glue up a piece of ebony cross grain to spruce, then put it under near continuous high stress for 20-30 years of cyclical temperature changes with occasional extremes, repeatedly cycling through different expansions and contractions as it goes through constant humidity fluctuations. Subject it to millions of vibrations and flexing of the plates, cycle the load up and down over and over, give the occasional moderate shock now and then, and in the real world the strengths and weaknesses can be appraised in ways no simple test can consider.

The adhesives will weather these conditions and cyclical changes much differently. Some that are too rigid and brittle may fatigue under cyclical stresses and fail much more quickly. Others with just a bit of elasticity may see less fatigue to a point, by too much and they could exhibit creep over the decades (even though this would never show up accurately in immediate failure tests where time was not tested equally to force). Some may fair well in brief periods of temperature or humidity extremes but suffer in repeated cycles of shifting between more moderate changes. The list of course goes on and on as to what is beyond the scope of simple abrupt failure tests, so while insightful to a point, they should not be taken as any final word.

Then of course when you take the view of long term preservation, it must be accepted that all joints will eventually fail. Some may take a few decades, others a few generations, but sooner or later every joint under stress will give. If you are concerned with preservation of what could become heirloom artifacts, then it is worth considering the serviceability of these joints when they do.

This (along with many other properties) is where hide glue has a distinct advantage. When a non- natural adhesive such as Titebond fails, the saturation layers left on each face of the joint are spent - all the hydrogen bonds the glue uses to create a proper joint are taken up with an essentially inert dried polymer, and new glue applied to these contaminated surfaces will never form a bond comparable to fresh clean wood. When reglued then, material removal is inevitably required to get past the contaminated saturated layer down to fresh fibers. How deep the saturation goes, and how frequently the joint needs to be serviced (bridge joints reasonably expected a few times each century) will factor in to the usable life of the instrument.

Protein glues like hide glue however, are more friendly to less invasive service. Although the saturated areas will contain a lot of broken strands of collagens (kind of like a lower gram strength hide glue), they will still retain the ability to rebond with new glue in a serviced joint. Therefore unlike PVA glue, you can still achieve a very good bond without having to fully remove the saturated layer of fibers, and prolong the life of the instrument by allowing periodic service to be performed without removing significant material each time.

Glues are indeed complicated, and testing them perhaps even more so. Not trying to persuade what you should use or even to explain the pros and cons of the varieties. Just saying it's more complicated than what a simple catasrophic failure test can reveal, and they don't tell the whole story of how performance will compare in the field.

Thanks David,

I flew test and evaluation in the military for part of my service (30 years, 5000 hours total P3B/C) and have experience in failure mechanism both in mechanical bonding and adhesive bonding - particularly long term high cycle/high stress fatigue fails (accumulated "micro fails"are part of this, and you refer to them specifically).  This is not to show-off knowledge - it's just to establish my rights to speak to the subject. 

Thanks for words on the subject to all concerned:  I was copping a hiding for trying to explain just how complex this subject is or can be.   My glue testing regime presented was a 101 attempt to at least get the the subject somewhere near useful for a inexperienced starter.

We build guitars for the commercial marketplace  and our reputation depends on our  performance and build quality.  We do not test the glues we use - we rely on outside agencies or the maker to do this.  

Thanks bloke,


Thanks David, that's a really informative reply. However, this isn't what I'm asking. I'm not trying to test the strength of hhg to determine whether or not it's good enough for my application. No, I trust the research and professional experience.

See, the difference between my Titebond original and freshly made Titebond that's still in the factory is very small. I trust that they have high quality control and test batches of glues before they pack them in bottle to ship out. The glue in the bottle of Titebond that I have is nearly identical to what's in the factory right now. The only significant difference is that mine is older/spent more time on a shelf. Moreover, Titebond (and superglue and epoxy) is very easy to apply. It's idiot-proof - you brush it on, you clamp it up, and you go about your day. There's no mystery to it.

Neither of these things are the case with hot hide glue. The glue I'm about to brush on a piece of wood is very different from the glue that's currently being manufactured. I'm sure they have their own strict QC and they test batches, but the fact still remains that I introduce a lot of errors. I cook the granules, and while cooking it isn't rocket science, it still introduces possible errors, whether they're human errors or equipment errors (like a bad thermometer). But not only that, the application of the glue isn't as easy as with Titebond. It gels quickly, and if the clamps don't go on before it gels, or if the piece move around as you clamp, then the joint could be compromised. These are the things I'm testing. My cooking of the glue and the way I apply it. What I'm not testing is whether or not the glue is suitable for gluing guitars.

Russell answered my question. If I glue things correctly, when I break the joint, the % of wood tear should correspond to what's given in the paper he attached (%59.82). This test is good enough for me.

Just some  thoughts.

Hide glue in dry form is very stable. As long as it's been kept dry and relatively cool, it's likely just as it was when it was made. Probably more so than Titebond. If it's not clumped together, it is almost certainly fine.

I think the ,"testing" you need, is for your methods of using it.  Experiments that will give you experience and a feel for how to work with it as quickly as possible. To that end, I would try some scrap tests of the extremes of use.

Make a bunch of pieces. Glue some together with the glue deliberately too thin, too thick, too hot, to cold, clamped together quickly, after 30 seconds, a minute, two, three, surfaces very rough, Just planed, etc.

These tests will give you more of what I think you need for confidence. Not confidence in Hide glue's strength, but in your ability to use it well. Doing things wrong deliberately, is often more illuminating.

One last thing. Don't be seduced by a thermometers digital display or lots of decimal places on it. If it is a cheep one, get another, or two, and see if they agree. Boiling water is 212 degrees at sea level, about -2 degrees for every 1000' of elevation above that. It's a quick test to make sure a thermometer is not way out of whack.


© 2024   Created by Frank Ford.   Powered by

Badges  |  Report an Issue  |  Terms of Service