Effects of thread wrapping, Series 2:

Plans A & C


Introduction, Plan A

After formulating some plans which we hoped might help restore our strangled boxwood flute to better health, we found ourselves unable to put the simplest plan, Plan A, into operation.  The plan called for just playing the flute, but the flute was not yet playable.  It was missing keys, it was cracked through the barrel and into some of the sockets, it had dodgy brass rings that are probably crushing the bore near the sockets - this thing is a basket case!  So, we put Plan B into operation first, and learnt some rather startling things.  While that was happening, I did some hasty repairs to the flute so we could revert to Plan A.

The Repairs

The repairs were minimalist.  Not because I was lazy or impatient, but because I didn't want them to impact on the experiment in hand.  For example, if I replaced the dodgy brass rings that appeared to be crushing the RH and Foot sockets with something permanent and nicer, their presence would prevent any hope of the sockets being restored to their original size.  Better to try to limp by without the rings, than compromise our quest.  So what did I do?

I elected to just leave the dodgy brass rings off, and support the weak socket wood by wrapping them firmly with string when I wanted to play the flute.  I now have a lot of faith in the constraining power of string!

The head was easy.  All it needed was a new stopper, so it got one.

The barrel was also straightforward.  Because of the crack through it, it was leaking like a sieve and certainly rendered the flute unplayable all by itself.  But the barrel, being lined, is not going to play a big role in this experiment, as its bore dimensions are defined by metal, not wood.  So, I felt free to do the usual repair - remove the slide, glue the crack, rebore and reinstall the slide. 

I had harboured hopes that I could get a more accurate reading of how much the barrel wood had shrunk, after I glued the crack shut but before reboring.  But the old bore was so distorted, I could have had my pick of readings.  Not unusual - I've tried that before with the same results.

The cracks in sockets I dealt with simply by injecting superglue.  I figured I could always get that off if necessary, and that I could review the success of the glue when I came to putting nicer rings on later.  If the flute survived that long!

I got around the leaky and missing keys by simply leaving all the keys off, and plugging the holes with poster putty when I wanted to play.  That limited me to playing diatonically, but that's fine - all we want is my damp breath coursing through the instrument.  But I felt it was important that I should be able to play up and down the scale - to close all the holes and just play D wouldn't be realistic.

I did find that I had to make sure the putty covered right over the pewter plug inserts, as more than one of them were leaking around the insert.  Take note, restorers!

I also found I had to plug a small knot on the back of the LH section that had been filled with resin back when the flute was made.  Seems the resin had shrunk more than the wood, or perhaps Plan B's excessive re-hydration had loosened it.  It took me some time to find it - the Magnahelic Flute Leakage Detector told me I had a problem, but not where to look.

Finally, I had to deal with the question of tenon lapping.  Because I had hopes that the tenons might be able to be un-strangled, I didn't want to do anything semi-permanent, like cork, which might then have to be undone.  But I clearly couldn't properly thread-wrap them either - that's how the poor flute got strangled in the first place.  So I opted for temporary thread wrapping.  I used some quite thick thread which went on quickly, and came off equally fast.  The thread would only be on for the playing time, and even then I made sure to wind it on loosely.  Being thick and squishy, a single layer was all that was needed.  When off the flute, I wound the three lengths of threads on to a short length of dowel, marked H and F at the ends, to make it easy to know which thread went where.

After all that, I had a flute that worked, well sort of.  It was rather sullen - more of a dead stick than a vibrant flute. And it still looked pretty miserable.  This was a patch-job, not a restoration.  That might come later.

What, no oil?

One thing I did decide against was oiling the flute.  A bit cruel, given it probably hadn't been played or oiled for 100 years or more.  But I'm trying to produce change, not trying to avoid it, and oiling the flute would have rendered it harder to influence.  Later, I promised.  I also promised to take it slow and easy. 

Indeed, it occurred to me, this is where we can actually test our advice on "blowing flutes in".

Plan A

You'll remember Plan A was:

subject it to some normal playing, to see if, freed from its tenon wrapping constrictions, we see it starting to move spontaneously towards one of the shapes we predicted in the previous articles.

More of a mission statement than a plan, isn't it.  It needs flesh.  What are we actually going to do?

Moved by the notion that we can also use this experiment to test some of our advice to players about "blowing in" a new flute, I decided to make the experiment as real as practical.  This meant:

  • the instrument would live in a hardwood wooden box, lined with cloth, pretty much like the 19th century mahogany cases. 

  • the instrument would only come out to be played.  After playing it would be mopped out, and the flute and the mop would go back into the closed box until next time.

  • I would measure the weight of each section of the flute before playing, after playing and after mopping.  I'd also monitor the weight of the rag. I would use scales with a resolution of 0.01gms, and I'd zero and calibrate the scales against its calibration weight each time,

  • I would use a datalogger, set to take a measurement every 5 minutes, to monitor temperature and relative humidity in the case,

  • I'd aim to play 10 minutes the first day, 20 minutes the second, and so on, until I reached some obvious point of conclusion,

  • I would graph the results, subtracting the original weight of each section, so we could see how much water it took on, where the water went, and how long it stayed around.

  • I gave the flute a few days in the box to stabilise before starting the experiment.  The same measurements detailed above confirmed that the flute and box had equilibrated.

And the results?

The results are shown in the graph below, and I'll walk you through them to make sure you get the full picture.

Firstly, the vertical scale shows the weight gain, in grams.  Good to remember that 1 gm of water is also a millilitre, or a cubic centimetre.  So, visualise the 1 gm line as an ice playing dice.

The horizontal scale shows the status, plus the date and time.

Each section of the flute has its own coloured trace, in solid.  The colour code is given in the legend in the graph.  The overall weight of the flute is shown in dark green dashed, and the rag in brown dotted.  Let's follow the action....


Day 1 - 5 May

"Before", 5 May, is our reference point, so everything is defined as zero.

I played for 10 minutes and took readings.  You can see that the whole flute gained just over 0.7gms, or mL.  That's not a lot of water!  Most of it is in the head, and most of the rest is in the LH section.  Negligible amounts made it to the RH and foot sections.

Then I mopped out, starting at the head, then barrel, LH, RH and Foot.  The rag was a bit of T-shirt material, chosen so that it would just pass through the foot, with the stick reversed, and used as a pull-through.  The stick was used as a push rod for the head and barrel.

Note that logic actually suggests we should mop out the foot first, and then make our way up the body, and do the head last.  This would reduce the likelihood of distributing the head moisture throughout the rest of the body.  But I wonder who thinks to do that?  Don't we all go for the wet bit first?  Because I'm looking for change, I didn't do it either, but I should test that theory too sometime!

(Hmmm, notice that every time we embark on one of these experiments, we end up flagging a few more experiments that are needed!)

Note that the mopping-out slightly more than halved the moisture in the flute, leaving the flute and the rag about equally moist.  It dramatically lowered the moisture in the head , but only reduced the LH moisture a little, possibly because of redistribution of head moisture, or possibly because it's easier to dry metal than wood, or maybe just because a wet rag now won't dry anything else.  Alternatively, maybe that moisture had already largely soaked into the wood of the unoiled LH section.

Day 2 - 6 May

As you can see, when I opened the box the next day and ran the measurements, pretty much all yesterday's moisture had disappeared, probably absorbed into the wood and fabric lining of the box.

This day, I played for 20 minutes.  I noticed though that in that time some moisture had run down the flute and dripped from the end, and the temporary thread wrapping around the upper tenon was also wet.  Sure enough, we see that borne out in the next set of measurements, "20 min play".  The measurements record more condensation than the 10 minute play, but not double that amount.  The various measurements more-or-less reflect the same pattern we saw after the earlier playing, although more of the moisture can be seen to have made it into the body and foot.  The head actually retained less moisture, probably because the moisture had become mobile. So, clearly, I had run into a turning point already, one where the moisture is no longer retained in the bore.  If it's not there, I can't measure it.  And if it's not there, it can't influence the moisture content of the flute wood.

Day 3 - 7 May

Once more, the next day, almost all the moisture gained had disappeared.  Playing flute might be fun, but it isn't the most efficient way to pump water.

I had planned for a 30 minute playing time today, but doubts were rising in my mind.  Even at 20 minutes playing time, the water had started to escape the flute, making it 1) ineffectual in increasing the flute's moisture content, and 2) unavailable for measuring.  But as well, I have to say I wasn't enjoying playing.  I'd picked up a ticklish cough somewhere along the line, and playing triggered it regularly.  I started to wonder about that mould we had to wipe off in the last article!  But as well, this flute, in its current condition, was simply no fun to play.  I staggered on to 10 minutes, and decided enough was enough.  As you can see from the graph, the third playing cycle produced results lying between the first and second.  Back in the box!

Days 4 and 5.

Next day, I did the usual readings, and found again that the flute had pretty much reverted to original weight and measurements.  To be absolutely sure, I gave it another day and found the same.  A total of 40 minutes playing time, and the moisture content was right back to where we started!

So, what did we learn?

Quite a bit, as it turns out.  Some of it is new, some is confirmation of what we probably all have noticed.  But of course it's good to get the confirmation.

I'll replicate the graph here so you can easily check any points you wish to.

  •  a metal lined head will quickly attract condensation, but, after about 10 minutes playing, that condensation will start to run down the flute.

  • the LH section of the flute ends up most moisture affected and is probably the only section at any risk (in a flute with metal lined head)

  • lower sections are more affected with longer playing times, but probably mostly because of run-off, rather than condensation

  • mopping a flute, at least with old T-shirt material, removes about 1/2 the water that was present after playing, but is probably helpful in redistributing the remainder, reducing local swelling and speeding evaporation

  • our measurement methodology holds up well until the point where water starts leaking from the flute. (But who cares after that?)

  • this methodology could be used to compare the efficacy of different mopping materials and tools

  • it may also offer a means by which the protective effect of different oils could be tested

  • the small amounts of moisture acquired have been dissipated overnight in a closed wooden box

  • a wooden flute box is shown to be very effective at buffering temperature and humidity.  The data-logger record varied only a few degrees and a few percentage points of RH over the test period.  As soon as the moisture evaporated from the flute and rag, it was assimilated by the case and/or lining

  • the metal-lined barrel appears to absorbs no moisture, even though some had leaked into the socket

  • it seems very unlikely that just playing for short periods is going to return our flute to its original shape in any reasonable amount of time. 

  • playing for long periods will subject the flute to very patchy moisture distribution which may or may not be helpful in restoring the original dimensions

Plan C - Simulated playing

So, if real playing is not very effective in wetting a flute uniformly and easily, can we find another way?  What about putting a wet rag inside for a while.  Surely that will work!

So, I found a scrap of rag (T-shirt material - by now you're probably getting a fairly accurate impression of summer attire in our little coastal village) that fitted easily in the RH section of the Potter.  I chose the RH section as the LH section was off assisting authorities in another line of investigation.  Anyway, it was still oversized from our humidification experiment, and I thought it would be interesting to see if something similar happened to our RH section, assuming of course we can convince it to take a drink.

I soaked the rag in water, then squeezed it out a little to the wet-but-not-dripping stage.  I threaded it on a cleaning rod, ran it just through the joint, detached the rod and tamped the loose ends into the ends of the section.  Then, every now and then, I pulled out the rag and weighed the section, comparing it to its dry weight in the graph below:

You'll see that, as soon as wetted by the wet rag inside it, it gained about 0.25% its own weight in water.  That water was visible, glistening on the surface of the bore when the rag was removed.  But even over 2 hours later, it had only gained a further 0.4%, making a total of 0.65%.  At 144 minutes, I terminated the experiment and mopped the section out.  Interesting that I was only able to remove about 0.1% of the water.

Interesting to compare the efficiencies of the various ways to inject water:

Exposure to saturated air 0.08%/hr
Wet rag method 0.16%/hr
Playing Depends on which section of the flute

So, the wet rag method works better for short periods, but would require to be refilled from time to time.  The saturated air method has the benefit of unattended operation.  Playing is just too haphazard, unless that's the purpose of the test.

What we haven't tested...

It's just as important to note what we haven't tested here, but really should at some time:

  • longer playing times (but needs a flute and flute player in better condition!)

  • flutes without metal liners, or with partial liners

  • timbers other than boxwood

  • fresh boxwood

  • the effect of various oils in reducing the rate of water uptake,

  • other mopping materials and tools

  • the difference between mopping the head out first and mopping the dry end out first

  • when should you keep the moist rag in with the flute, and when should they travel separately

How come we don't already know all this?

The wry thought springs to mind - how come I'm having to "discover" all this stuff, why don't we know it all already?  The flutes we play have been around for hundreds of years, yet where do we find anything about the effects of moisture on them?  We've had the technology - the beam balance is known to have existed for 4000 years or more.  Certainly, if you can make a flute, making a beam balance is a pushover.  Surely I can't be the first flute maker/researcher in hundreds of years to wonder just how quickly the water condenses and just where it goes?  Instructions for blowing-in flutes exist, but what informs those instructions - science, experience or guesswork?  Or have we relied on remaining just this side of disaster, like the reported instruction for tuning a lute:

Tune up thy minnikin (thinnest string) until it breaketh, then, tuning not quite so high ...

Who should study this stuff?

It also begs the question - who should be studying this stuff?  Most scientific interest in flutes has come from the acousticians - Benade, Coltman, Fletcher, Nederveen, etc.  They can tell you what the changed dimensions do to the flute, but not how or why it changed dimensions.  It's just not their field.  The jet raises some interest among the aerodynamicists, but again, this has nothing to do with jets.  Organologists, the name for those who study musical instruments, tend to be more interested in classifying instruments than monitoring their dimensions in use.  We're really looking for wood technologists, but, where they exist among students of musical instruments, they tend to be seduced by the vibrant instruments such as the fiddle.  That's where the action is; everyone knows that the wood in flutes is just a box for the air inside.  A living, moving box, as it turns out.

So, what about flutemakers, shouldn't they be across the technical dynamics of the materials they work in.  Sure, that would be great, but is it practical?   Should we all have studied moisture migration and wood movement in Flutemaking 101?  But, even if we had had the luxury of being formally trained in our profession, who would have informed the course presenters of what we are measuring here?

So, the writing's on the wall (sorry, Mr Belshazzar, sir, I promise it won't happen again, sir.)  If we want to understand, really understand what's going on inside our flutes, we'll have to take a look ourselves.  Which is why we are here today.  So, get over it and get on with it!


Well, we certainly learned some interesting stuff, and we have subjects aplenty for a stack of other studies.  But I think we can ditch Plan A in terms of getting Mr Potter's poor flute back into condition.

Where to next?  I'd have to say that the question of Hysteresis is weighing on my mind.  Did the LH section move to a higher moisture content and larger dimensions during Plan B just because of some natural hysteretic effect in wood?  Or did some other change take place during Plan B?  I feel an experiment coming on.  Wrong, I feel two experiments coming on...

On to Hysteresis?

or Back to McGee-flutes Index page...

  Created 28 May, 2011