FLUTE MAKING for “Model Engineers Workshop” magazine
I am one of a small number of artisan makers of modern musical instruments. Collectively we keep alive the traditional hand-working methods that were common until the 1960s but which are now in serious decline. We offer a truly bespoke service, a fastidious attention to excellence and listen closely to our customers. We share the same enthusiasm for our work as do the craft makers of everything from bicycles to beer. It is probably true to say that many of the very best instruments are still made entirely by hand, often, as in my case, by one person.
In this three part article I am going to outline some of the processes involved in making a flute and while not forgetting that my readers may have more interest in tools and techniques I shall also have to explain my design philosophy as determined by musical requirements and how it may differ from that which governs the mass production of the flute in factories around the world. The processes and materials are not necessarily those used by other makers.
Part 1 will form an introduction to the subject and go on to describe the tube making.
Part 2 will cover the key making or “engineering”.
Part 3 will talk about finishing and maintenance.
The word ‘flute’ describes a large family of instruments but not withstanding the accomplishments of your Editor on several of the more primitive sort, I am limiting the discussion to the modern orchestral flute made nowadays mainly in metal.
Schematic of the modern Boehm flute
I did not take a course in “flute making” for back in the 1980s there was no means of doing so, at least not in the UK. I was however married to an accomplished professional flute player, had a degree in engineering, played the clarinet a good deal and was currently self-employed as a harpsichord builder. Model engineering was in the mix too. When she announced that she needed a better instrument we consulted a friend, the well-known silversmith John Webb, and to cut a long story short, he and I set up to make her a new flute! At this point a technical clarification is needed. The flute comprises three main parts, or ‘joints’: the middle and foot joints that carry all the tone holes and keys, and the head joint. The latter could be described broadly as the mouthpiece. It is a short tapered tube carrying only the ‘lip plate’ in which is cut the embouchure hole. [Photo 1] Naturally it is of crucial acoustic importance but is usually thought of as a separate entity, so much so that there are specialist makers of them around the world who do nothing else. John, at the time I met him, was one of these and had begun to find a market.
Photo 2 shows a complete flute with all three joints connected. Photo 2a gives a closer view of the body. Although a perfectly standard Boehm flute (the fundamental layout having changed little since Boehm’s day, c.1847) this one has several new features explained in the text. Photo 3 shows the bare tubes.
Wessel flute fitted with Sheridan head joint
Another flute with an alternative key arrangement for G &G# at centre
The three joints shown separately
The Early Business
Flute production at that time was, and still is, dominated by American and Japanese makers. In earlier times the better instruments came from Germany or France while English makers and players followed a long tradition of using wood rather than metal. Fashion for the all-silver “French style” swept all before it during the 1960s and the English makers began to disappear.
Meanwhile, many of the leading London players had developed a passion for rebuilt 19thC Parisian flutes, claiming them to have a far more beautiful sound than anything coming out of American or Asian factories. The trouble was that these old flutes had been around a bit and were very delicate, having thin tubes and notoriously unreliable mechanism. So it seemed to us neophytes that here was our starting point: we would try to emulate these expressive instruments but improve their mechanicals with the help of some modern materials. Lots of necks sticking out here for the simple reason that you don’t “copy” something by altering its fundamental composition! But unless you do, nothing much will be learnt.
To avoid having to delve into the history of the flute, for that is well beyond the scope of this article, I will simply state that we decided to keep to a relatively thin silver tube, around 0.013” wall, and try stainless steel for the keywork instead of the usual silver or nickel silver. Steel would be lighter and stiffer than the traditional metals and we planned to inlay some of it with polished black plastic to lighten it further. We had several other ideas, some of which were tried out in prototypes and test rigs but the design soon settled down to a broadly conventional flute with a highly responsive action that did in fact sound quite like those Paris flutes.
My role in all this was to make the keywork and fit it to the handmade tubes supplied somewhat erratically by John. I also did all the finishing work and got them to play properly under my wife’s guidance. Thirty years later I am doing the whole thing with the exception of head joints, which as mentioned earlier are better bought from specialists. Most of my customers already own a good head joint that they want to carry on using in a new flute so it makes good commercial sense to do it this way.
What makes a flute good enough for a fine player? This question has of course dogged the lives of every instrument builder since the beginning of written music. Musical performance of every sort is a highly competitive and emotionally charged business. There are always many more musicians than available jobs and only the best will earn a living by playing alone. So their instruments are of vital importance in giving them the edge. Here is a list of the basic qualities of any wind instrument:
- play in tune
- have a highly reliable mechanism (usually referred to as “keywork”)
- be responsive to the tonal needs of the player
- have a focussed, resonant sound that projects well
- play evenly from bottom to top
Taking these points in turn,
1. The tuning of any wind instrument is a compromise. You are trying to get all 12 notes in each of three octaves as near to “equal temperament” (as a piano keyboard) as possible. This is theoretically impossible but every generation of makers and players sees an improvement. The exact position and diameter of every tone hole is determined by a mixture of maths and playing experience. The idea is to make playing “in tune” as easy as possible with the minimum of embouchure change. Nowadays we have got pretty near.
2. Properly engineered keywork is vital. All the holes on a flute are covered by thin cups or keys, each containing a soft pad to make an airtight seal. With only 9 active fingers (the right hand thumb is used for support only) some of the keys are closed automatically by others, involving a number of clutch mechanisms. Each pad must close off its tone hole completely and positively with the absolute minimum of finger pressure. If there is a small leak the flute will not play properly; it will lack power and resonance. So keywork has to be precise, light in weight and hard-wearing enough to withstand the rigours of daily practice, travel and changing atmospheres. And, of course, comfortable. Rather like a satellite, once it leaves the workshop, the flute is on its own out there, receiving little attention and sometimes no maintenance for years.
3,4 & 5. Flute sound should be anything but bland and monotonic. A good player will expect to produce a whole palette of colour and dynamic according to the demands of the music. Nobody knows exactly how this is achieved, some instruments enabling these variations more easily than others. American flutes tend to be heavy and thick-walled. This gives them raw power but often at the expense of expressiveness. They have a tendency to limit the colour spectrum and impose their own idea of flute sound upon the player. I am convinced that too much weight in the keywork has a bad effect: a sort of deadening. Different players can get a different response from the same instrument, partly because the acoustic system includes the oral cavity; everyone’s embouchure, and of course skill level, is unique too. Players know instinctively what they want from their instrument and it is the maker’s job to interpret this as best he can; he is a “tool maker” and intuition is often a better guide than physics.
Another major consideration is that most flutes will form part of the orchestral woodwind section where they have to blend. The standard repertoire comprises a vast amount written in previous centuries as well as our own by composers familiar with earlier, sometimes more primitive, versions of the modern instruments we know. You cannot therefore introduce a radically new sound into the mix. The musical profession is a conservative world that tolerates only minute incremental changes. Anything the maker can introduce that improves the ease and flexibility of the playing experience is good because the player is then able to concentrate solely on the music without worrying about the instrument. He wants the best tool for the job. This should be the maker’s goal, pure and simple.
This is the acoustic heart of every wind instrument, containing the air column which is set in vibration by a rapid alternation of eddy currents around the embouchure hole. (Or in the case of the oboe, clarinet and bassoon by the reed vibration) The body tube of a flute is cylindrical and has a bore of 19mm and a wall thickness (varying between makers) of 0.012” to 0.018”. (The two units of measurement may be characteristic of our times, the original work on the “modern” flute having been done chiefly by Theobald Boehm in Germany during the mid 19thC while later developments were mostly English or American). The default material is silver. This is far too expensive for the millions of factory produced instruments so nickel silver (a copper/nickel alloy) is used instead for the cheaper instruments and then silver plated. Gold is also used for some high end flutes and occasionally platinum. Other exotic materials such as pure tin, glass, even rock crystal, have been tried too but usually end up in museums. The tube material does have an important secondary effect on the sound but practicality and hygiene considerations have a part to play – precious metals are less reactive to the kind of stuff likely to be deposited on them.
I use silver exclusively and the seamless tubing is manufactured especially for me in Birmingham. I discovered recently on a visit to the works that, in place of the sophisticated CNC plant that I had always imagined necessary for the high quality production of such tube, it is actually made more or less by hand by the same fellow who has been doing it for nearly 30 years! He has become better and better at getting it perfectly round and within tolerance and it was nice to be able to compliment him personally. The tube starts life as a thick-walled tubular casting. It is then drawn down progressively through steel die plates on to a series of steel mandrels until it is several metres long. Constant annealing is required and scrupulous cleanliness to prevent pits and scars. Various other diameters are also needed for tone hole saddles and sockets (see fig 1 for explanation of terms).
Fig 1. Stages in footjoint construction
Silver is a delightful metal to work with, somewhat like copper in that it rapidly work hardens; very sharp tools are needed, particularly for turned work. In the annealed state it is fairly soft so great care is needed to prevent damage such as dents. These can however be removed with almost no trace, a job often needed during servicing or repair work.
In the early years we made our own tubes individually by wrapping a silver sheet into a tube, soldering a butt joint and then drawing it down on a hand operated draw bench. The whole process was extremely difficult; every tube would come out slightly different from the last and the scrap rate was about 50%. Thinking about all that now makes my hair stand on end.
I deliberately order the tube slightly undersize on bore with a wall about 0.001” too thick. This gives me the freedom to get both exactly right on my own mandrels as well as straighten it when necessary. This work is done chiefly by very light burnishing using a piece of polished ¾” steel rod. Uniform strokes up and down all the way will gradually enlarge the bore while if they are confined to the concave side only, the tube will curve in the opposite direction.
The tube is then annealed and checked again for straightness. Laying out the hole positions is done on the lathe using the lead screw. [photo 4]. A small cross is scribed at each centre and a tiny mark representing the edge of each hole. This work needs to be done with great accuracy working to about 0.002”. Some of the holes are in line with each other but several are placed at various angles around the tube. This information is carried on a large disc temporarily attached to the mandrel supporting the tube.
Fig 2 Saddle positioning
Marking out tone hole positions. The disc shows all the angular data and the piece of angle clamped to the lathe bed carries an index mark.
The distances between each hole and from the end of the tube form what is known as the “scale”. Many different scales have been devised since Boehm’s day and overall pitch has also varied. This has now settled at about A-442 in most countries. The scale I use is best described as “work in progress” by William Bennett (known to all as Wibb), one of our top players and professors of music teaching who has done more to further the modern flute than anyone. I am fortunate to have received his support throughout my career. His latest scale is as good as it gets.
When all hole positions have been checked, a pair of fine dividers is set on each cross and used to scribe arcs representing the outer edge of the saddle. See fig 2. These marks are vital because a pilot hole will be drilled through the cross and if that drill decides to wander off course, which it usually will, the centre is immediately lost. The purpose of this hole can be seen in Fig 1 and Photo 5. A small clamp bar is needed to hold each saddle secure during soldering. A 10BA screw passes through it into a similar bar placed inside the tube. This photo shows two alternative foot joints of different lengths, the longer one provides not only an extra low note B but also a different colour or timbre to the sound of the entire compass. Players can choose which they prefer and often have both.
All saddles clamped in place ready for soldering. Two foot joints are shown (see text) and a head socket.
My flutes are unique in having silver-soldered saddles. Nearly all production instruments have so-called “drawn” tone holes whereby some of the tube material is drawn and stretched up to form the “saddle”, no additional material being required. This is a process that can be done by hand using a special tool but the tube wall around each hole is then inevitably thinned and highly stressed. This locked- in stress runs counter to my philosophy that the tube should be stress- free and homogenous; it should leave the workshop in the fully annealed state. It will gradually harden with age and playing but in a way defined only by the acoustic vibrational patterns set up within it. Unrelieved stress caused by manufacturing processes could be one reason why many mass produced flutes play unevenly. Some upmarket flutes do have “soldered” tone holes but invariably this is carried out with soft solder which, in the old days, would have contained lead. Now lead is the great enemy of silver: at room temperature any lead solder on silver will gradually become porous as the two metals eat one another; at red heat an instantaneous reaction takes place resulting in a hole in the silver! I learnt this to my cost many years ago when a tiny spot of lead solder, unseen on the hearth, somehow got itself on to the piece of silver I was heating. I didn’t realise until later I found a mysterious little hole.
The saddles are made from five different diameters of seamless tubing and are profiled accurately to fit the body tube. This is done by offering up a length of tube to a drum sander of the same diameter as the body tube running in an old lathe. After sanding, the saddle is parted off to the correct length plus a finishing allowance of a few thou.
Soldering on the saddles is a job to be done on a quiet day – no radio, no storms and floods, phones off the hook and no likelihood of unannounced visitors or postmen wanting signatures. Utmost concentration with a completely relaxed state of mind is what’s needed here; the tube is supported on an old bow saw blade with the teeth ground off. This keeps it straight without taking away heat. Just enough solder is introduced on the inside of each saddle and is seen to flash around making a neat and strong bond. I use only the silversmiths hall-marking solders, Easy, Extra Easy, Medium and Hard. Each contains a minimum of 66.7% silver and all are free flowing and delightful to use, melting at different temperatures. The work is liberally fluxed with Argotect rather than Easyflo; this prevents the formation of “fire stain” leading to less cleaning up. Standard silver is 92.5%, the remainder being mostly copper. At high temperatures silver is porous to oxygen allowing the copper to become oxidised to cuprous oxide which is pinkish. The longer it remains in the fire the deeper is this oxidation and it cannot be removed by the pickle. You are left with faint pink blotches that can only be removed by abrading. Curiously, if the staining is uniform all over the piece it is quite pleasing but you don’t want it on a flute.
Mouldings and sockets
These are the stiffening rings placed at the end of the main tube sections and also form part of the sockets that connect the body tube to the head joint and foot joint. The rings have a decorative section based on the classical moulding shapes such as the cyma, torus, etc and are made from rectangular silver strip that has been formed into a circle, hard soldered and turned using a form tool. Between soldering and turning the bare ring is jammed on to a tapered mandrel and hammered all round until it perfectly fits over the relevant tube.
After turning the ring is annealed and pickled before fitting to the tube. If it is a little loose on the tube, a graver is used to raise a number of tabs on the tube surface that will grip the inside of the ring during soldering.
The sockets themselves start out as short lengths of a larger size tubing that closely fits over the body tube. Socket tube is burnished as described above until it fits smoothly. Silver is a slippery metal so sliding one tube into another relies on a good fit, absolute roundness and no trace of oil or dirt.
Cutting the tone holes
Once all the soldering is complete the material inside each saddle must be cut out to form the tone holes. This is undoubtedly the most hazardous process involved in making a flute by hand.
Cutting the tone holes out.
The problem is the tube is very fragile and cannot be supported on any machine other than by a mandrel passing down the inside. So milling them out is really out of the question and in any case would take far too long; I use a tiny burr running at top speed in a hand held drive unit to rough out each hole, continue with a larger burr followed by fine sanding drums. Once again, complete concentration, a relaxed but very firm grip and total confidence in the tool is the order of the day. A tentative approach will lead to disaster when the tool kicks violently, jumps out of its path probably nicking either the saddle or the tube outside. The scrap-box may then receive a large donation while misery reigns. The main cause of this trouble is either a blunt burr or more likely a worn nose bearing in the appliance. Given that accidents and errors do occasionally occur I learned a long time ago that almost never is it worth the time spent trying to correct large ones. Much better start again, despite the loss of precious metal. This is Rule 1. Rule 2 is that every customer is equal. Never ever should one be tempted into thinking that a young student or beginner is less important than a top professional and might not notice some tiny defect. Next thing you know she is having a public master class with that same pro!
The final job is to clean up. A scraper run around the bottom of each hole will ensure a smooth rounded junction. The traditional undercutting in this area applied to wooden instruments such as clarinets and oboes isn’t necessary or possible on a metal flute that is so thin walled. Assuming the tube has been pickled to remove most of the oxides, there may be small and very thin patches of fire stain despite the special flux. Very fine Scotchbrite mops can be used around the holes but generally a wet & dry paper of 1500 grit is sufficient followed by 2500 and finally 4000 cloth. Ultimately it will be polished with rouge but that can now wait until the instrument is almost finished.
Part 2 – Keywork
So we come to the engineering for a modern flute. I have already described the broad requirements but before describing how the parts are made let’s drill down further into the characteristics of good keywork. All musical instruments have traditionally been built with an eye to aesthetics: apart from needing to sound well they also try to satisfy the eye and the touch. Unlike early keyboard instruments that might have been lavishly decorated with either fine veneer work or paint, the flute is relatively bare in this respect, but the keywork can be made to have an understated beauty of its own. Form, function and comfort are intricately linked. The old French makers were masters of this but their example has rarely been followed by 20thC mass producers for such subtleties of design are usually at odds with economic production. One route towards this refinement is to remove all material unnecessary to strength and stiffness, a rigour somewhat reminiscent of aircraft or bridge design.
Keys must move up and down many millions of times in just one year of playing. Each one must do this without any hesitation or binding and as silently as possible. Most keys normally stand open, are closed by the fingers and return by wire springs. Each closure, which might only last a few milliseconds, must result in a perfect airtight seal over the tone hole. The player will have no time to increase finger pressure to cure a slight leak. Instead, the note will simply not appear, or only at half cock, resulting in a passage of music that sounds “fluffed”. In engineering terms the key must therefore be light in weight so its inertia is low, the spring should act instantly but have a low rate so that the finger does not feel it “winding up” as the key closes. Long thin springs are better than short fat ones. Friction must also be minimised by good pivot design and absolute minimum oil.
Every key will have a tail that limits its travel back towards the open position; the tails contact the body tube so must carry a tiny cushion to keep them quiet. These cushions are usually made of felt or cork.
Mounting the keywork
The bare body tube joints are now fitted with a support structure for all the keys. This comprises a number of posts soldered to a silver strap that is in turn soldered to the tube, usually in sections. The stainless steel posts are cross-drilled at right angles to carry the rodding which supports the keys; some are threaded, others are plain. They are also drilled for the spring anchorage, mostly about 0.025” diameter. The various rods that pass through these posts must all be parallel to the body tube and to each other so great care is needed to get the posts all exactly right. In the photo the foot joint is seen to carry all five posts in a straight line such that the rod can pass easily through four of them and screw into the fifth without any deviation from straightness. Some of the rods are fixed like this one while others can oscillate, between conical “point screws”. See Fig 1 below.
1 Stainless key posts silver soldered to straps which are then soft soldered to body tube. Good alignment is vital.
The cone allows subsequent adjustment to take up wear. Cheaper instruments have cylindrical screws that do not allow this, so once wear appears the resulting rattle cannot be eliminated so easily.
All the rods and screws in my flutes are stainless steel and case-hardened. This process is generally thought impossible with stainless but in practice it does work, albeit with a very slight loss of corrosion protection. Since the hinge tubes that rotate about these rods are also stainless the hardening gets around the problem of two similar materials in relative motion.
The straps are wired on to the tube for soft soldering. I use a tin/silver alloy for this which runs freely underneath leaving very little to clean off. If a blob does get in the wrong place the best way to remove it is with a tiny brass chisel that is slightly softer than silver, so won’t mark it.
The use of stainless steel seems to be unique to my instruments. All other woodwinds use more easily cast or forged metals which, if not silver or gold for high end work, will most probably be either nickel silver or silicon brass, both of which then have to be electro-plated to prevent tarnish. In all cases there are very many soldered joints, each section of keywork being built up from several components.
Stainless does have several advantages but the one big drawback is that it cannot economically be cast or forged in such small scale; instead many of the parts have to be made by hand, which is why the large manufacturers cannot consider it. In the early days I had to hacksaw each part out by hand as no bandsaw blade will touch stainless. I improved on the situation a bit by building a vertical hacksawing machine but by then laser cutting was just arriving on the scene and seemed the obvious way to profile all the many little parts. Later on the machines became so powerful that burning around the tight radii became a problem; for many years therefore I have had them cut out by water jet.
2 Water cut plates of various thicknesses, 2 – 6mm.
The tone holes of a modern flute are too large to be covered by a finger alone. Each one therefore is covered by a lid, known as a “key” (to be nicely confused with either a piano key or the key in which a piece of music is written, like G major for example, or indeed a house key). Each key is fitted on its underside with a special pad that makes contact with the tone hole and seals it when the key is pressed. The seal must be 100% airtight under almost zero finger pressure. The pad is composed of a cardboard washer, a felt washer and a double layer of thin skin taken from the lining of a cow’s intestine, known in the trade as fishskin (don’t ask). They are nearly all made in Italy and are easily bought in a variety of sizes, thickness and hardness.
Each key is attached to an arm that is hinged by a length of tubing. Fig 1 shows a typical key in profile. There are five fingered keys that usually have a central perforation about 9mm in diameter which requires the finger to cover it while closing the key. These holes allow better “venting” when the keys are open and also encourage a better hand position in young players. The remaining solid keys are not closed directly by a finger but by a simple mechanism, lever or clutch.
3. Typical piece of keywork operated by the left hand second finger. The side lever at right is worked by the right hand.
In Photo… the key on the left is actually closed by the other (fingered) one via a clutch plate, just visible. It can also be closed by the side lever on the right independently via the “bridge” piece, or by the second clutch worked by a key on the next section. This piece of keywork is suspended by a rod that simply slides through all the hinge tubes and is pivoted between two posts. It consists of 13 separate components.
A note on silver brazing stainless steel. For the ordinary austenitic steels like 303,304,316 etc used by model engineers there is no problem at all. I use Easyflo flux and JM Extra Easy solder. Proper cleaning is vital because it’s only too easy to think that a piece of bright stainless looks clean when it isn’t. Abrasion or scraping immediately before fluxing is best. The correct pickle is a 10% mixture of nitric and hydrofluoric acids (yes HF, not HCL) used warm but quickly. Nitric will remove excess solder but can leave it porous. The HF will clean off the black oxides but is a highly unpleasant and dangerous material that I ceased using years ago. Nitric on its own is good enough. Note that the normal sulphuric acid pickle found in most workshops will have no immediate effect other than to remove flux but it will severely corrode stainless steel given more time.
The silversmiths solders are so free flowing that a joint such as between hinge tube and key arm can be assembled tightly before fluxing to allow accurate positioning; flux is then applied all around and will find its way into the joint followed by the solder even though joint clearance may be almost nil.
Stainless is a very poor conductor of heat which means that parts can be soldered in close proximity to previous joints without danger of remelt. This is particularly useful during repair work as will be seen in Part 3 of this article. Silver by contrast is an excellent conductor making soldering work more difficult.
The need for precision
I once got a friend to play through a Bach flute sonata very slowly so that I could count the number of times a particular key was moved. Multiplying the result by the number of times she might realistically practise or perform the piece in one year produced the astonishing fact that this key alone would move many millions of times in that short time. Given that the instrument will often be wet from condensation, dirty and full of fluff if used in stage or theatre environments and that service intervals are often well over three years it is clear that the demands on the mechanism are severe. During that time the player expects flawless mechanical action and no rattles!
A common problem is the sticking key, ie it stays shut when the spring ought to open it. This is truly maddening for the player and could easily wreck a performance. Another less serious defect is the sound of a metal to metal clash often caused by too little clearance between moving parts. Photo… shows the potential for this. Which of our famous steam locomotive designers was it who said in relation to valve gear that the art lay in the prevention of two large chunks of metal occupying the same space at the same time? He was right of course and it applies equally to woodwind keywork.
The trade-off is a classic one: freedom of movement as against quietness. A free action can be a noisy one but a quiet action may well err on the tight side and have a tendency to seize. We therefore balance on a knife edge. The flute is particularly difficult in this respect. No one minds a rattling bassoon or saxophone but flute “noise” is objectionable. It mostly depends on having the right clearances between moving and fixed components while always remembering that imperceptible changes in geometry can be caused by a change of temperature. A frequent cause of the sticking key is lack of end clearance at the hinge. Bearing in mind that this would be measured in “tenths of a thou” the tiniest bit of grit or greasy grime can block it up. I shall have more to say about maintenance in Part 3 of this article.
Many of the water cut parts have to be reduced in thickness since stainless plate is only available in metric whole numbers, such as 3,4,5,6mm etc. My design started life in the days when SWGsizes were available and 10G was perfect for finishing at exactly 1/8”. So now I have to start with 4mm and grind the parts down further. This is done on a big sanding wheel using various holders and two grades of paper. The edges are finished with various little homemade sanding drums that can be fitted with cloths ranging in grit size from 240 to 600, ie much finer than anything on the market. My design for these drums is shown in Fig… They are very quick to make and I strongly recommend their use for all precision metalwork finishing.
Fortunately imperial size hinge tubing is available and arrives semi-polished. The bore does need a small correction to bring it up to size so that it becomes a sliding fit over my ground rodding. I have a collection of drills all hovering around 3/32” to 2.4mm and choose one to give the best size. Reaming in this scale is out of the question so finishing is done by wrapping a short narrow length of 240 grit paper in a spiral around a piece of 1/16 rod. This is worked in and out of the rapidly rotating tube until the correct fit is obtained with the rod. This all sounds very crude and inefficient but I see no economic alternative and it actually works very well.
While on the subject of tube, readers may be interested in a method of straightening tubes (and rods) for it goes without saying that unless a hinge tube and its rod are perfectly straight, smooth working isn’t possible. Stock material is always slightly curved. I haven’t tried this in anything above 3/16” diameter and am only concerned with 1/8” or below, but I see no reason why the principle shouldn’t work for larger sizes. Photo.. shows what to do.
4. Tube straightening. The rotating tube is firmly stroked from left to right while maintaining a marked curve just within its elastic limit.
One end of the tube is gripped tightly by the lathe collet or chuck while the other is free. A piece of wood with a hole in the end is all that’s needed. I use a small hammer handle and the hole is about ¼” and bell-mouthed. If the tubing is only slightly bent the procedure is as follows: pass the handle over the tube and start the lathe running at about 1000rpm. Now angle the handle so that both ends of the hole start rubbing hard against the tubing while at the same time sweeping it slowly towards the other end. This will seem to put a strong curve in the tube right at the limit of its elastic range. At the end of the sweep slowly bring the handle hole back in line with the lathe centre line, without slipping off the end. You can almost feel the original kinks being ironed out and it should now look straighter. Two or three passes may be needed. Then reverse the tubing in the chuck and do it again. If the tubing was severely bent at the outset be very careful – it is sometimes necessary to support the outer end in the fingers to stop it whipping about. Straightening rod generally needs more determination and more angling of the handle, depending on the metal status, ie if it has been annealed it will respond better. It is possible with practice to get rid of the slightest kinks that may be almost invisible but can often be felt in the fingers. The method also works for quite short lengths down to about 1”. Use a narrower handle and a collet rather than the 3 jaw chuck. With long lengths be very careful to keep the handle on the tube while it is rotating, in order to restrain it.
The reason for inlaying a black plastic into all the main keys is twofold:
- to reduce weight and inertia
- to provide a warmer and more secure touch than metal
Plastic, being a poor conductor of heat, feels relatively warm to the touch in cold conditions and doesn’t become wet and slippery in hot climates.
I use a jet black acrylic that takes a high polish and is fairly hard. Inlaying the round keys is straightforward; the plastic discs are turned with a form tool and set into the key recess when all soldering operations are finished. An industrial cyano adhesive fixes them.
5. Keys are made by turning. Other manufacturers invariably press them from sheet metal.
The odd shaped pieces (see photo..) are cut out of thick sheet material using the hole in the key as a guide. After filing to the exact shape, the inlay is squeezed into the key until it makes contact all around. Thin adhesive is then applied at the back and finds its way into all the little hairline gaps before setting hard. Finally the excess plastic is sanded off to produce a gentle rounded surface that blends smoothly into the metal surround.
8. The construction and installation of the left hand thumb key cluster. The key is kept open by 2 flat springs screwed to the underside of each lever.
8a. Right hand section
I shall describe the finishing of the flute in Part 3 of this article. This will cover the critical subject of padding a flute such that every key can be made airtight and some of the factors affecting the life and service intervals of the instrument. Before concluding this part however I must mention an extraordinary commission that came my way a few years ago.
The Pronomos Flute
9. Compare this picture to a normal flute shown in Part 1. It has the same basic layout of a Boehm flute but with many extra keys and levers. The black rectangle is a hand rest. Overall weight was a major consideration so keywork metal was reduced to the minimum consistent with strength and comfort.
The modern flute as we know it hasn’t changed much since the middle 19th century except in detail design and improved tuning. The fingering system laid out and simplified by Boehm is sacrosanct and in no need of fundamental change. It will play a chromatic scale in just over 3 octaves. However during the 20th century an increasing interest in “quarter tone” music has caused woodwind instrument makers to consider how all these extra notes might be sounded. To those unfamiliar with the jargon, ¼ tones are “the notes in between the semitones” such that instead of the normal 12 notes to the octave there would be 24. To most of us these notes would just sound out of tune and impossible to associate with any Western music from the last 1000 years. Music from other parts of the world however has always made much use of these sounds, so cross-cultural contemporary composers often like to blend them with more conventional Western music. In fact a whole new contemporary music repertoire has grown to deliberately challenge our traditional attitudes as to what music really is.
Flutes able to play some of the quarter tones have been around for a long time but the Pronomos takes the ideas much further. Not only is it able to play a full 24 note octave but with non-conventional fingering it can be made to play many more. This means the player can imitate many different types of simpler flutes from around the world that will not play the Western scale. It was designed by the Hungarian player Istvan Matuz in the 1970s, then rediscovered and much developed by my client more recently. The first fully engineered version is illustrated in photo 9. In the hands of this remarkable Spanish musician the Pronomos is quite extraordinary. Yet it can still play ordinary Western music when required. It was very difficult to build but after 5 years of constant use around the world seems to be holding up well. It has been calculated to offer over 23 million different finger combinations, the vast majority of which, to our ears, produce very weird sounds indeed!
10. Pronomos thumb keys.
11. Pronomos foot joint keys showing several extra levers known as ‘gizmos’. These are arranged to be pressed either alone or together with the main keys. A lot of work for just one little finger!
12. A normal foot joint arrangement for comparison.
Part 3 – Finishing and maintenance
In the previous articles I described the basic processes of manufacture of my modern concert flutes. In this last part of the series I shall talk about the finishing of the instrument and keeping it in good health thereafter, for like all mechanical things it needs proper maintenance and sometimes repair.
Once all the keywork is made and fitted, the main outstanding jobs are to fit the return springs, install all the pads and provide the necessary cork and felt cushions to limit key openings. The flute will also receive its final clean and polish.
The springs are simple lengths of stainless spring wire varying in diameter between 0.022 and 0.032”, the actual size chosen depending on length. One end is anchored in a key post near to the point of action while the other end bears upon a tiny lug integral with the key arm or soldered to the hinge tube. The anchoring is achieved by hammering a flat on to the wire end; the other end can either be left as cut or preferably ground to a gentle taper and polished. I do this by holding the spring in a pin chuck and twiddling it round against a very fine abrasive wheel. There is no need for any accuracy here but the result looks more purposeful and is useful if the spring is short, improving its characteristic in bending. It is then poked into the hole in the key post and forced home by specially adapted pliers. The flat stops it rotating or coming loose. With the relevant key in place the spring can be tweaked to give exactly the right amount of force such that the key can be easily closed but will spring up again with no hesitation. Keys that are normally sprung shut like the G# generally require a slightly heavier spring but players have widely different tastes, some preferring the keys as light as possible, others wanting more resistance. The main thing is to get the whole flute sprung evenly so no finger will encounter any discernible variation from one key to the next.
Photo 1. Three of these springs keep their keys normally closed but at lower left the small key is sprung open.
Given that the body tube is the acoustic heart of the flute, determining its general timbre, its personality, response and tuning, all this would be hidden were each and every tone hole not completely sealed when required. Imagine a plain length of tube having no holes connected to a head joint: it will sound just one note, known in acoustic terms as the “fundamental”. More notes are available as natural harmonics and can be sounded by blowing harder or slightly differently but few of them apart from the octave are of any practical use, being in disagreement with the so-called tempered scale that musicians require. So holes are cut into the tube in order to shorten its acoustic length. This length is the distance from the mouth hole to the nearest edge of the hole. If all the holes are closed, the tube will sound its fundamental: middle C in the case of the normal flute. Opening each hole in turn, from the end, results in a semitone increase in pitch. However, if just one of those holes is not quite closed, ie leaking, the sound will be less powerful, less resonant and to any serious player quite unusable.
The pad, as explained previously, takes the place of a finger. It is the point of contact between player and her instrument. She expects it to shut off that hole in an instant with just the lightest touch. If she is faced with a rapid passage of music, each key must touch down and seal completely within a very small fraction of a second. She doesn’t have time to lean a bit more heavily on a key to make it close. If it has a leak the note simply won’t speak: it will be missing from the sequence and she may think she has fluffed the passage.
Pads are made using a combination of simple, natural materials: cardboard, wool and skin. No synthetic substitute has ever been found, despite numerous attempts, that properly replicates the feel and complex requirements that the traditional construction offers. You can think of the cardboard, in the shape of a washer, as the baseboard. Upon that lies a felt washer of varying thickness and hardness usually about 1 – 2mm thick. Overlying this are two or three layers of very thin skin which are drawn over the felt and around the back of the card to which they are glued (see photos 2 & 3). The result is a little white or yellow cushion which, once installed correctly into the key, will be air and watertight. The felt has enough “give” to withstand either a heavy fingered player or a light fingered one. The skin, taken from cow intestine is similar to gold beater’s skin and has the property that when moistened and briefly heated it will shrink (see below).
Flute pads are nearly always held in place mechanically, either by screw and washer or some kind of retainer. Very small pads like those found on piccolos or oboes, and three on the flute, are usually glued in with shellac.
The technique is as follows: given that the overall key geometry is correct, the installed pad should just touch the tone hole all around. When the screw is tightened the centre of the pad is pushed inwards and this creates a series of radial ripples in the skin. The pad in this state is therefore quite unable to seal the hole; it has to be “ironed” flat. There are a number of ways of doing this but my own method is (of course!) much the best. All that’s needed is a fine jet of steam directed on to the pad for 3 seconds. It is then held down lightly on the tone hole using a spring clamp for about 2 hours. The skin shrinks, the ripples disappear and the pad takes up a very light outline of the hole. Once dry it should be fairly stable and can be tested for leakage. I usually put in 4 to 6 pads at once.
In the photos you can see both the ripples in a pad prior to steaming and one that has been seated successfully.
Photo 2. Two types of key show two methods of retaining pads. Pad at lower right shows ripples prior to steaming.
Photo 3. A steamed or “ironed” pad, together with some unused ones. The small pad has to be glued in place.
Nearly all problems that occur with padding are down to poor geometry rather than the pad itself. The pad normally protrudes from the key surface a small amount and this must be allowed for in the relationship between the key surface and its hinge point. Poor design or workmanship can result in either the back or the front of the pad making contact first, leading inevitably to a crescent shaped gap at the opposite side. Such gaps can be eliminated either by a different pad thickness or a paper shim behind it but I shall address this later on.
New pads sometimes move a bit during their first few days. They have been compressed under wet heat and are presumably longing to retreat somewhat back to their original state. The best remedy for such insubordinate behaviour is to play the flute! Gradually they get used to their new life and will provide 100% airtightness for a long time. What they don’t like is being subjected to a bone dry atmosphere and total neglect. So don’t leave your flute in the desert, come back in 3 months and expect it to work.
Once padding is complete there is one day left. All the little bumper felts are glued on to key tails, clutches are fitted with either thin cork or nitrile rubber interfaces, then the whole flute regulated such that all the keys open the correct amount, work nice and easily with no hesitation or undue noise. Finally the tube is polished using rouge. Keywork is taken off for this but springs and posts remain, so it’s a tricky job working around all the obstacles. Having gone over it again with 4000 cloth I then use miniature buffing wheels made of chamois leather. These last a lot longer than calico but are fairly aggressive so care is needed. I also make my own wheels up to about 2” using discs of cotton. These don’t last but are very soft and gentle. There are many kinds of metal polish and even T-cut has its place but there is no doubt that silver responds best to rouge.
The very last job is to make a wooden case. I shall not linger on this except to say that mine are always made of walnut, are lined inside with velvet and given concealed brass catches. My wife who is an accomplished leather worker makes a gig bag to protect the case. So every part of the whole commission is made by us with the exception of pads.
I mentioned testing for pad leaks. For this I use a simple water manometer which I designed many years ago specially for flute testing. The idea is to isolate a short section of the tube underneath a tone hole by means of rubber seals, then introduce a very slight air pressure into the space with the key closed. This will register on the manometer as an inch or two of water. If there is the slightest leak the water levels return to their original equal heights either quickly or slowly depending how bad the leak is.
Photo 4. The “Flute Ferret”. At centre is the manometer. To the right is the probe that isolates a portion of the tube under a tone hole. The syringe inflates the two red rubber seals that lock the probe in place.
Photo 5. The Flute Ferret in use. The lowest C key is under test. The gauge shows a pressure of less than 1” of water is holding.
The pressure is applied by blowing very gently down a rubber tube into the device. This is a beautifully simple and sensitive test that can be applied to every pad very quickly. It will not tell you exactly which part of the pad is leaking so for that you need a tiny feeler gauge which is slipped between pad and tone hole and then gently tugged while the key is closed. I use a short section of cassette tape having a thickness around 0.001” or less, which gives an idea of the precision required.
Photo 6. Using a feeler gauge to find the loose spots. The highest sensitivity is required.
Somewhere around the circumference there will be a loose spot where the feeler comes out too easily. In such cases the pad is removed and a segment of thin paper shim inserted underneath the light spot to bring the pad up to an even height. This work is extremely delicate and of course time-consuming but has to be done. It is more commonly needed on instruments that have come in for servicing or ones in which the pads are becoming old and worn. With luck, a brand new flute should seal straight off.
Photo 7. Shimming a pad. Note the segment of shim placed in the key to raise that portion of the pad by about 0.001”. An extended needle is used to manipulate the pad as they should not be touched directly.
It is now time to look at what can go wrong with a flute and what the repairman (or woman) can do to put it right. After 35 years in the trade I have become convinced that the skills needed to service and repair woodwind instruments are very rarely found among musicians themselves. There are a few who can do it but the vast majority prefer to pass even quite minor adjustments to a specialist. They are probably right to do so because it’s all too easy to make things a lot worse through either a poor understanding of mechanical things or worse, having neither the time nor patience to do it properly. There you are, a gifted player let’s say, practising in the green room just before an important audition or recital. Suddenly a key appears to have come loose, or a note won’t sound properly. What on earth to do? You look around hopelessly at the other musicians none of whom have much clue either. Luckily you brought a spare flute but if only you had kept up with the regular servicing…….!
There is no defence against pure bad luck but being alert to potential trouble and the very first signs of its development is something that can be learned. I tell my customers that flute maintenance is not unlike dentistry – have your regular check-ups with someone you trust and hopefully the worst will never happen.
Probably the most likely event to cause a player’s consternation is the sticking key. It closes easily but then won’t snap open, either staying down or rising sluggishly. Several things can account for this: old oil that has become too viscous; dirt, grit or fluff that has found its way into a pivot, a loose spring that has lost its anchorage or, most commonly, a mechanism that has been built or set up with too little clearance. Flute makers tend to pride themselves on precision, sometimes forgetting that in the real world temperature changes and natural flexures of the instrument do require a certain looseness for things to work smoothly. I have learnt this lesson myself but you see the problem everywhere from clocks to steam locomotives.
Then there are the pad leaks: these usually appear very surreptitiously, beguiling the player into thinking her playing has gone a bit “off” recently. She will blame herself up to a point but compensate by blowing differently or pressing slightly harder on certain keys. In the end there’s nothing for it but to take it down to the repairman who will inevitably find the leak, probably several. He might shim the pads or replace them with new. His thankful customer will delight in the new-found resonance and power that this simple work will release.
Flutes get dropped. Foot joints work loose and fall off. One of my customers earns her living playing on cruise ships. The confined environment, press of people, changes of climate and constant salt breezes do their business. Every now and then an enormous wave overcomes the ship’s stabilisers and everything goes flying. The poor old flute comes through it all though, constantly collecting a patina of tiny dents and scratches. Last time I saw it the whole tube needed straightening.
Dent removal on a flute is straightforward as it can be slipped on to a steel mandrel that fits the cylindrical bore. The dent is then burnished out with a variety of polished homemade tools. A deep sharp dent will probably leave traces however as once the metal has been stretched it can’t really be pushed back into itself; instead the dent spreads out into a host of micro ripples. Once flattened and polished these become virtually invisible. Scratches too can be much reduced in a similar way.
8. Severely dented portion of tube
9. Dents being burnished out on mandrel. Downward pressure on tool causes inside surface to rise. Fine papering shows progress.
10. Tube is now fairly true again. Reflections of the strip lights above magnify the imperfections.
More serious repairs and sometimes alterations to the keywork are sometimes required. If silver soldering is involved then pads and plastic inlays must in general be removed first. Occasionally the latter isn’t feasible or may take too long but thanks to the poor heat conductivity of stainless steel, nearby plastic parts can usually be immersed in a bowl of water while the part to be heated remains just above the surface. This is a useful technique that avoids too much deconstruction of an assembly if only one small area needs reheating – however high the flame is set, parts below water level will remain at no more than 100°C, probably less.
I am often asked about the best way to clean a flute and whether the owner should oil it occasionally. It goes without saying that any mechanical device is less likely to give trouble if it is kept clean. Pad leaks are often caused by nothing more than a small piece of fluff or grit that has settled under the sealing surface. Combined with finger grease, general grime and excess key oil, the likelihood of serious leaks is great. Once a year the instrument should be stripped down and thoroughly cleaned using IPA (isopropanol) followed by polishing compound. Some repairers use an ultrasonic bath for the first stage. Oiling should be very sparing indeed; on a well-made flute it does little other than attract dirt. On older flutes made before the introduction of stainless steels for rods and springs oil was needed to resist corrosion. In either case, this work is better carried out by a professional repairer.
If you have read this far you may well be wondering what an article about flute making is doing in a magazine largely devoted to the hobby of model engineering. I happen to indulge in both activities finding that the one constantly informs the other. The tools, machinery, techniques, materials are all very similar. The approach to a problem may require a lateral thought process of a kind that model engineers are so good at; for example the making of a jig to hold parts together during soldering, or a peculiar set-up in the lathe. One difference is the economic one, for to become a professional maker it is vital to work fast and accurately as mistakes cost money. More importantly the average buyer of a flute has little technical knowledge and expects a flawless mechanical performance for many years whereas the products of a model engineer’s workshop (including mine) tend to stay with their owner who can tend to their leaky valves and other foibles as needed. Many years’ experience has taught me that good reliability of any mechanism is usually down to surface finish, especially at edges and corners where the slightest burr formed during manufacture, possibly invisible to the naked eye, will eventually either break off or wear down the neighbouring moving part. When this happens, a lightly loaded mechanism will probably seize, while the exact cause remains elusive. Another factor is corrosion. The products of corrosion are always hard, granular and either sticky or abrasive; wherever there is condensation and lack of use for a protracted period Nature does her best to reduce even our finest work to gum and grit. A thin film of oil is no guarantor against this process.
Finally I should like to say how immensely privileged I feel to be able to make things by hand. This is a gift that presumably all readers of this magazine possess but which may be less common than it once was. We should use it, value it and encourage others to do the same. Whenever we pause to wonder why we may be spending so many hours doing frustratingly difficult things in our workshops when we could be nattering away in the pub, the only answer is that we have been put there to do it. Others can’t.