The effect of moisture on the performance of woodwinds

Irish Flute
1

Hi all

Someone sent me the link below. I say someone, because I don’t yet know who it was! And perhaps you were all aware of it anyway. Wouldn’t be the first thing that I was last to know!

It’s a paper by the late Claire Soubeyran and now-retired conservation scientist, Tim Padfield on the topic “The effect of moisture on the performance of woodwind instruments”. Unfortunately, I can’t see a date to know when it was written, which is a bit strange, as publishers are usually very finicky about proper citation. Interestingly the paper covers a lot of the same ground and appears to come to a lot of the same conclusions as my two series on the “Effect of thread wrapping on flute tenons”. For those interested in such things, here are all the links:

Claire and Tim’s paper is at: https://www.conservationphysics.org/flute/flute.html

My first series starts: http://www.mcgee-flutes.com/wrap-survey.htm
My second series starts: http://www.mcgee-flutes.com/Wrap2-Intro.htm

2

That is an interesting read. Thanks for posting it! It is pretty consistent with what I have
observed. I was interested to see that the degree of deformation varied quite a lot based
on the material used to wrap the tenon, with cork and waxed silk thread showing the lowest,
and equal, deformation. The main conclusion in terms of tenon wrapping seemed to be that
the more hygroscopic the material, the worse the deformation, and that rewrapping or recorking
your tenons to tighten the fit when they become lose due to the repeated wetting and drying
cycles just accelerates the process.

None of this is too surprising, but it does make me wonder how tenons wrapped with a modern
hygrophobic material, such as polyester thread, would behave. Polyester thread is both more
elastic and more hygrophobic (less hygroscopic) than silk. So does this mean that it can do better
than cork, if silk and cork exhibited the same deformation?

It does make a convincing case for using stabilized wood and for flute designs that have fewer,
or ideally no, tenons.

Some of the measurements discussed were taken between 1979 and 1985, so the paper
must have been published after 1985, and at least some of the work was started in the late
1970s.

Oh, I just saw on her website that she obtained the grant to do this work and made
the journey to the US in 1985.

3

I am dealing with flutes that get sat on or worse, left on the highway where someone drives over them in a tractor-trailer. The issues discussed in that article are way too esoteric for me!

4

It’s an interesting read, especially the section on the microclimate measurements inside the bore over a period of time.

The one thing that stood out as a sort of red flag though, was that there was no mention of the type of wood being used for the tests, or how it had been prepared. We all know that wood changes as it ages from a freshly harvested billet, through air or kiln drying prior to machining. Antique flutes have an even longer time to settle down.

Some species of wood are much more prone to water uptake than others, which wasn’t mentioned. The paper treated “wood” as if all types had the same properties, with text like “The observed shrinkage is greatest at the very end of the tenon where the end grain wood is exposed. These cut cells absorb water very rapidly.” I wouldn’t describe the end grain of blackwood or cocus flutes I’ve owned as “absorbing water rapidly.” More rapidly than the edge grain, sure, but “rapidly”?

I’m suspicious of this photo in particular, of the test subject “tenon” that doesn’t look to me like the kind of wood we have in our flutes (although I could be wrong). It looks a bit rough and open-grained, possibly not even a hardwood?





Maybe a softer wood was used to show more dramatic changes, and I’m not disputing the overall points being made. It just seems sloppy not to mention differences between types of wood, and fail to specify the species of wood and its preparation used for the tests.

5

Very interesting, indeed.

It would be of interest to test a material such as roasted (“torrefied”) maple. According to the guy I get my roasted maple from, the process collapses the cell wall of the wood, thereby creating a stabilizing effect. Since it is the collapse of the cell walls within the wood (at the flute tenon) which contributes to the deformation of the tenon, would pre-collapsed wood avoid the problem?

Obviously stabilized woods would have an advantage, as paddler pointed out, and well-sealed endgrain is not difficult to achieve these days–there is a wide choice of epoxies, finishes, CA glue, etc..

6

I don’t think it would be all that difficult to omit the barrel from a flute that has a tuning slide, thereby eliminating the topmost wooden tenon (which seems to suffer the most). I did that with my previous iteration of my Pratten flutes. The flute simply came apart at the tuning slide instead. This made reaming the bore a little more complex, since working with longer sections makes reaming more difficult.

But if there were no need (from an ergonomic perspective) to have the body section of the flute in two sections, one could eliminate all of the tenons if they wanted to use an integrated foot. Or at the very least they could have only the foot joint, which seems to suffer the least from the moisture issues. It might complicate the design of a flute-case to suddenly have a flute with such different dimensions (shorter head section, longer body) but…

And a flute without a tuning slide (such as a Baroque flute) can certainly be made without anything but a foot joint. It’s not easy to do it, and there are other complications related to the length of the wooden sections that would be required, but I’ve actually done this on a few occasions. I experimented with a line of single piece Pratten flutes with integrated foot joints. One continuous piece of wood, so I know it is quite feasible.

The greatest difficulty lies in reaming such a long section. I make all of my reamers with the ability to attach to a long extension so that I have the option for making one-piece flutes with all of them. But when reaming one long section, the amount of torque you have to deal with becomes a factor.

7

Conical bore makes a good point about the wood types and unspecified treatment. The wood in the photograph
looks like an untreated fruit wood to me, maybe pear. I’ve been experimenting quite a bit recently with woods like
this, both untreated and resin stabilized. They can make good flutes, especially when stabilized, but unstabilized
they do move a lot more than a wood like blackwood.

It is not a crazy material to look at for experiments like this, though, especially in the context of baroque flutes,
which was Claire’s main interest at the time. I think the results are just more pronounced and over a shorter
time frame. The issue is reduced with woods like blackwood and cocuswood, but doesn’t go away. This is evident in
the bore compression we see in so many antiques, and which is documented quite clearly in Terry’s studies. We also
hear frequently from people who left their new blackwood flutes assembled after playing and then could not get them
apart. Water does enter through the end grain and tenons do swell and later over-contract. I think what Claire’s study
does show is that this effect happens very quickly with materials that are subject to this wetting-drying cycle.

8

Ah, well done. Wow, 1985, that was a long time ago. My work on the topic was started around 2011. I wonder if I would have done it had I been aware of Claire’s earlier work? Probably, since I was reacting to a couple of seriously impaired flutes (I remember using the term “strangled”) I had come across which I suspect were more damaged than the flutes she was investigating. But I might have saved myself quite a bit of effort!

I see in my introduction: “It’s probably worth pointing out that, while the topic of bore compression has been mentioned from time-to-time, nobody to my knowledge has previously delved into it and presented their findings.” Wrong! Makes you wonder what’s also out there we do not know about. Still, good to have the confirmation that her findings were repeatable. I think we can now take all this as given, unless someone can prove otherwise.

9

There are a number of odd aspects of the article, not the least being the whimsical illuminations. It seems to suggest that this wasn’t a peer-reviewed article in a learned journal. So we are wise to exercise caution in interpreting findings, although, as I’ve indicated, I seem to have found pretty much all the same stuff.

10

Agreed. From my observations, the situation is going to be much worse in the case of early flutes in fruitwoods or box, still significant in the case of 19th century boxwood and to a lesser extent cocus, and not so significant in the case of modern blackwood. Not sure about 19th century ebony.

What it all does remind us is that we have been just a bit barmy faithfully reproducing the twists and turns of 19th century bores when we can now see those bores may have been seriously mangled by serial strangulation. When we look back at my survey introducing my series, we see most flutes show some suspicious bore tightening in the first 27mm - the typical length of a top tenon. Some - the Strangled Boxwood in yellow, the Schuchart in the Bate in aqua, the Strangled Cocus in magenta - demonstrate serious collapse. The Richard Potter in orange is not looking good either. Only the “Geo Rudall, Willis Fecit” seems oblivious to the forces, and as I’ve said in the study, it has a very shallow thread trench and a very fat tenon. Too fat to be stylish, but fashion isn’t everything…

11

I found that series on your website really fascinating Terry, thank you.

It’s very non traditional, but delrin flutes use rubber (synthetic rubber) o-rings. I suppose an 0-ring channel would be too deep?

What about in place of cork, a band of silicone? It’s virtually indestructible and has a lot of elasticity.

I have a plastic Boehm flute, a Nuvo. It’s not a great flute–the keys are kind of slow–but it has a lot of clever features including o-rings at the joints and silicone key pads, which it seems to me would seal well and probably never wear out

12

Ah! We’ll get back into this banned topic again!

Terry - would you do a test in Boxwood with a corked joint, then the same but assembled with the corresponding socket? It would be interesting to see if the socket compresses the tenon similarly to thread wrapping.

Another parameter is the thread wrapping itself. I am finding I can get a cork-like feel if the thread is E gauge as opposed to EE. F is really coarse. D is too fine and takes too long to wrap. These are Gudebrod Thread gauges and the threads used were originally for wrapping hand-made fishing poles. I am thinking of making one or two for myself and competing with the J Pod of endangered Orca whales for the salmon that is swimming by currently. Less salmon this year and the whales are out there chasing them every day. I saw a Humpback Whale breaching just a few miles from my workshop last week. Annoying critters - they frequently keep me from my work! I’d rather be watching them which is similar to waiting for water to boil sometimes.

I digress - am in a manic period though a highly focused one. Today I head back to the workshop the first time in a week after last week’s mai out of some 40 flutes.

The Gudebrod thread is Nylon, 3 ply. Unlike all the other similar 3 ply threads on the market, this thread is unfused and so it packs flat with a fair bit of cushion. It may be that your thread compresson derives from this differende as I have never witnessed this in my flutes, including ones made from woods softer than Boxwood such as our local Madrona, Pear, Plum, Almond and Olive. I’;ve been using Gudebrod thread my entire career.

If you want Terry I can mail you a few spools for these tests. I’d be curious.

I have a client coming by sometime later. She got her flute corked. Was a total disaster. She is going back to thread and having me show her how it is done. When my Apprentice gets here in May I will have her assistance to make a video. Its not rocket science but I have found a few tricks that some may not have thought of.

Casey

13

I hope we can avoid this discussion falling into the polarizing, false dichotomy that “thread is bad, cork is good”.
I think this study by Claire Soubeyran, and Terry’s later studies, present ample evidence that wooden tenons are
a weak point in flute design and that when they degrade they can alter the bore shape and the overall performance
of the instrument. This is something we need to know about in order to protect our instruments, to make more
resilient instruments, and to make more accurate replicas of historic instruments (by factoring in changes that likely
occurred due to degradation).

Real value can come from delving more deeply into this issue. It is not just a thread vs cork issue. We have seen that
even a naked tenon can suffer significant degradation over a fairly short period of time, and that various lapping
techniques can make this worse, or much worse. It seems clear that the wetting and drying cycle is the underlying
cause of this, but it would be interesting to know which aspects of tenon design might help mitigate the effects. For
example, should tenons be thicker? I work on a lot of antique flutes and I do see quite a lot of variation in the tenon
thickness and the depth of the thread trench. Related to this, tenons often have combing cut into the thread trench.
This exposes a lot of end grain and must surely weaken the tenon further, especially with regard to its vulnerability
to the wetting/drying cycle. Is this combing really necessary? Given a thick enough tenon I would expect it to start
behaving more or less like the rest of the flute body, unless it is being very aggressively strangled by some external
mechanism.

The expansion due to wetting seems to happen very quickly, judging by the data presented, which is not surprising.
Basically, the wood in the tenon expands while, and after, we play. It makes a good case for not leaving your flute
assembled after playing! However, it is also clear that during the playing period the tenon is under compression from the
socket, and there seems to be no getting around this, regardless of how the lapping is done. By strengthening the socket
with a metal band, do we exacerbate the problem? Would it help if the socket and tenon could both move together, to
some extent? These socket effects may be why the results for cork and waxed silk thread were similar in Claire Soubeyran’s
study. Even though the cork likely allows more free movement of the tenon when the flute is disassembled, once snuggly
inside the socket there may not be that much difference between the two. In the case of cotton thread, though, the lapping
expands when wet and jams in the socket, causing even more compression (and probably tightening afterwards too).

In the context of this kind of study the type of wood used, and how it has been treated, is clearly relevant, but so is the
type of thread used and how it has been treated. And, in fact, how it is wrapped. I use thread quite a lot, but I’m not
religiously attached to it. I just want to make the best flute I can. When thread wrapping I follow an approach I learned
from the writings of Rod Cameron … but I can’t remember where I found it. I soak the thread (silk or polyester) in melted
beeswax, and I melt some beeswax on the tenon. The thread is then loosely wrapped around the tenon so that it sits in
this bed of wax. Periodically, I smooth out the layers using my thumb nail, and I keep doing this until I have a precise
fit. The thread is not under tension. It is neither stretched nor is it wrapped around taking the absolute shortest possible
distance, so there is scope for movement. It all mats together because of the wax. The thread itself also has some elasticity
and absorbs relatively little moisture, both due to the material it is made from and the waxing. So, I go to a lot of trouble,
but I still worry that the tenons may suffer compression and I’d like to explore improvements, both when restoring antiques
and in new flutes I make myself. This is why I value these studies, articles, and board discussions.

14

Why not go back to one or two piece flutes? Minimize the number of joints?

I like taking mine apart and putting them back together. it’s kind of a magic trick: Observe the parts: now they make an instrument! It’s nice to have a small case too.

But if you told me that the instrument would be less variable and more durable I’d be happy with one piece with tuning slide.

15

I think there are a couple of reasons for the 3 or 4 piece (or five) flutes. Depending upon the wood, it can be difficult to get long pieces that are in good condition (free of knots, etc.). Settling long pieces over time in order to prevent warping has complications. Long pieces will tend to warp a bit as they settle (depending upon the wood–some are worse than others) and this means you need to have pieces of greater girth to provide enough material to “true” the piece once it has settled. This might have to happen a couple of times in the process. So you need thicker pieces.

Apart from sourcing larger pieces and processing them to an acceptable stage for working them, you face the task of reaming them. Short pieces are easier to ream, and it’s easier to manufacture shorter reamers. You can make longer reamers and you can ream long pieces, but it can be much more difficult. When you are reaming a conical bore that is continuous (no joints) as you near completion, the reamer is in contact with the wood for most of it’s length and at that point the torque increases a lot. Figuring out how to hold the work and manage the mechanics of it becomes proportionally more challenging.

I have done such flutes (Pratten flutes, for example, that are one continuous piece of wood), so it can certainly be done. But my set up is a bit unusual. I can drill pilot holes to a great depth when needed, and because I don’t use my lathe for the reaming (using instead a dedicated gear motor that holds my reamers) I’m in the position of being able to ream virtually any length of wood providing I have a way to hold onto it (which I do). It takes a bit of muscle, however. And I’m obliged to develop a lot of alternative timbers to use for this task, including resin stabilized woods or ebonite. I could never get a hold of a piece of blackwood (for example) of the correct dimensions to allow for this approach.

16

There is also the issue of longer pieces requiring use of a lathe with a longer bed, and
difficulties of controlling chatter while turning long, unsupported sections, hence
requiring special mandrels, steady rests etc.

Working with smaller pieces has many benefits for the maker, but I think the split body
flutes (with separate left and right hand components) may have originally come in with
the idea of corps de rechange (interchangeable body sections) that allowed baroque flutes
to be played more easily at multiple different pitches. The idea seemed to catch on after
that and persist even after the invention of the tuning slide.

17

I remember back in the nineties seeing a wooden flute with three thin rubber o-rings set into grooves in the top tenon. Seemed like a good idea at the time, but I soon heard no. If you don’t keep the rings and the socket well greased, you can inadvertently roll an o-ring out of its groove and have it jam in the much thinner space between socket and tenon, whereupon it puts great pressure on both, and is really hard to get apart.

It does seem strange in this day and age that we depend on the bark of the cork tree, but I haven’t been able to think of any thing better! I think one of the plusses is that it is weak longitudinally, so it can’t exert any significant force (or resist swelling) like the thread pack can. It’s also compressive, so it can take up any spare room (within reason!) if say the tenon shrinks or swells. And it glues well, which a poly product might not.

But keep thinking!

18

Hopefully we can avoid fisticuffs this time! I think last time, it was a very new and bold revelation, one some were not prepared to countenance. This time we have correlation from Claire’s study. That’s not to say we fully understand all the implications yet.

Terry - would you do a test in Boxwood with a corked joint, then the same but assembled with the corresponding socket? It would be interesting to see if the socket compresses the tenon similarly to thread wrapping.

So you mean what would happen if a corked joint were to be left plugged together longterm through multiple drying and wetting cycles? Is the cork’s compressibility enough to avoid the problem occuring? My guess is that you’d see some compression, but not as much as thread. But it’s not really a meaningful test is it, as we tell our customers to pull the flute apart and dry it after playing?

Another parameter is the thread wrapping itself. I am finding I can get a cork-like feel if the thread is E gauge as opposed to EE. F is really coarse. D is too fine and takes too long to wrap. These are Gudebrod Thread gauges and the threads used were originally for wrapping hand-made fishing poles. I am thinking of making one or two for myself and competing with the J Pod of endangered Orca whales for the salmon that is swimming by currently. Less salmon this year and the whales are out there chasing them every day. I saw a Humpback Whale breaching just a few miles from my workshop last week. Annoying critters - they frequently keep me from my work! I’d rather be watching them which is similar to waiting for water to boil sometimes.

I digress - am in a manic period though a highly focused one. Today I head back to the workshop the first time in a week after last week’s mai out of some 40 flutes.

The Gudebrod thread is Nylon, 3 ply. Unlike all the other similar 3 ply threads on the market, this thread is unfused and so it packs flat with a fair bit of cushion. It may be that your thread compresson derives from this differende as I have never witnessed this in my flutes, including ones made from woods softer than Boxwood such as our local Madrona, Pear, Plum, Almond and Olive. I’;ve been using Gudebrod thread my entire career.

If you want Terry I can mail you a few spools for these tests. I’d be curious.

At some stage perhaps, Casey. I’m up to my ears in stuff too at the moment, so I’m not in a position to do anything now. But yes, I think there would be threads that make the problem worse, and threads that reduce the effect. And the thickness of the tenon is going to be really significant, so it goes to flute design too.

I have a client coming by sometime later. She got her flute corked. Was a total disaster. She is going back to thread and having me show her how it is done. When my Apprentice gets here in May I will have her assistance to make a video. Its not rocket science but I have found a few tricks that some may not have thought of.

What was the problem with the corking? Ineptly done or unwisely done?

I’ve come across a few flutes that I think just are not up to corking. Typically they have very thin tenons and very thinly walled sockets, so there just isn’t enough room for a reasonably thick piece of cork. I like to use 0.8mm (1/32") sheet cork. Sadly it’s these flutes with very thin walls that are at greatest risk of strangulation!

19

I’m not sure that tenons are the problem, as I’ve seen no evidence that cork-lapped tenons suffer from any problems. Claire seems to have come to a slightly different conclusion, but I haven’t had time to read her work thoroughly and consider the implications.

I guess the question is what alternative do we have for tenons if we want the flute to be compact when travelling, and to permit fine adjustment of rotation. Rose (as in Rudall and Rose) patented a flute that appears to have tuning slides embedded at every joint. But imagine the splitting issues this might introduce!

20

About now, we might be starting to think, OK, I get it, tightly bound thread wrapped around thin tenons particularly in relatively soft and thin timbers is going to end up with cellular collapse and compressed or even strangled bores. But, just in case we’re getting a bit smug, I am aware that that doesn’t appear to be the end of it. As I mentioned above, I haven’t even had time to consider Claire’s work to see if she references the same issue. Later…

Check this set of bore curves, from my “Classical Bores I have known” page.

If you look at the bore traces over the first 30mm from the left, you can see signs of tenon compression in most of them. Nothing new there, just confirmation that this is a commonplace issue. The varying shapes of the compression is interesting…

Coincidentally, most of these flutes are of the “long body” type, so there is no LH/RH tenon-socket joint, except for Rudall & Rose 5501 and Rudall Carte 7174 (navy and aqua traces). Their middle joint occurs at around 210mm. Note that there is evidence of tenon compression at the bottom of the LH section on both. But there is also appearance of bore expansion just below the joint. What?

All of the flutes have a separate foot joint, circa 320mm in from left. Again we now expect to see some tenon compression leading up to that, and we often do. But again, we sometimes see what appears to be bore expansion just into the foot bore.

Now, in the light of my and Claire’s work, it’s easy to understand the tenon compressions. The thin wood of the tenon gets wet and tries to expand, the tight bands of thread wrapped around it prevent that. Cells are crushed, and when the moisture evaporates, the wood shrinks to a smaller size than it used to be. This loosens the thread wrap, and the diligent owner compounds the problem by rewrapping the thread tighter. History repeats itself, maybe many times.

But what explains the bore expansions just below the sockets these tenons plug into? If the bands of tight thread are busy stopping the tenons expand, they can’t be also expanding the insides of the sockets, can they? Or are we seeing some kind of general collapse except in the areas that that these bands of tight thread are?

And what is the role of the (usually) metal rings that support these sockets? If a band of tight thread can cause mayhem, surely a metal band has some powers too?

I don’t have answers at this point, only questions. Just alerting you to them, to demonstrate our work here is not yet done…