String Making – An Introduction to Alternative Methods and Materials for Qin Strings


INTRODUCTION

In my previous post String Making – The Beginnings of a Journey, I discussed some of the background of my journey into string making for the qin, mainly using alternative synthetic materials. This post will focus more on some of the actual details and discussion relating to several topics of the process that I use to make the strings, including material selection, the process of making the strings, and the results so far. This post is fairly lengthy in detail, but is still not an absolute and complete guide to the process that I have used for making strings so far – I will be creating several in-depth pages dedicated to each specific stage of the process, with full pictures and detailed explanations. However, for the information provided below, I have included the presentation PDF I introduced in my previous post as well for reference – I would strongly recommend to open up the PDF and follow along throughout the read of this post to get a better grasp of some of the concepts I am talking about. The PDF includes comparison charts of material properties, stress-strain curves of various textile fibers, pictures of my actual string making process for these types of strings, example calculations for string diameter estimation, microscopic pictures of various strings, and preliminary data on the strings presented in the PDF. The PDF can be accessed here: engineering-sound-a-look-into-alternative-string-materials-string-making-techniques-and-harmonic-analysis-of-the-qin, or at the very bottom of the post.


THE MATERIALS

Perhaps the first and most critical of decisions to be made is the selection of the materials to use for the strings. This includes significant research into the fundamentals of strings, as well as the physical properties of the materials. The first step was to look up the mechanical properties of silk, and compare them with other known textiles. Some of the most important graphs to look at first includes the stress-strain curves of a material, which shows how a material responds to developing stress. There are several regions of the curve, in which the material undergoes temporary deformation, or elastic behavior, permanent deformation, or plastic behavior, and ultimately, failure. These graphs for textiles generally give the average results for a particular textile fiber, but the basic information is perfectly valid and important for these initial stages. Eventually, I would like to build my own universal test machine to develop my own stress-strain curves of the actual strings I test and make, so that I will not only have the full harmonic profile of the strings, but the material and mechanical characteristics as well. Each material has different responses. To simplify my initial attempts, I wanted to choose something as close as possible to the response of silk, since I would be exploring this route first. Metal has a radically different curve, as do high-tenacity materials, so synthetics that are closest to silk may give me the closest response. In regards to average stress-strain curves, nylon and polyester end up being the closest in response. The responses are different, but are right around the same range of stress-strain profiles and failure. You can refer to the stress-strain diagrams in the presentation PDF below, specifically pages 9-13.

The next important factors to look into is the material properties themselves. Density is particularly important, as this is one of the major terms used in the equation for calculating the fundamental tuning of a string. The equation for this is as follows: f=0.5L*SQRT(T/u), where f = frequency, L = vibrating length of the string, T = tension, and u = linear mass-density. Note that this is a simplified equation and mainly applies to monofilament strings, so if yo are calculating for a multi-core string or a string with wrappings these have to be taken into account in the equation as well. Rope twisted cores will also be tricky as well, since the tension may be non-linear, and the linear mass-density may change more depending on construction, due to increased flexibility and stretching when under tension. In regards to density for a simple string however, as density increases, the frequency will decrease, all things being equal. So with a material of the same density, a thicker string will result in a lower tuning for the same tension, and for different materials, a higher density string will yield a thinner string for the same tuning. This is why metal-core qin strings are much thinner than silk. So it will follow that a synthetic with lower density than silk will need to be thicker on average for roughly the same string and tuning, and a higher density material will be on average slightly thinner. Note that you can adjust the strings to be thicker or thinner depending on the tone you want, and can result in a variety of gauges for the same material and type of string. Silk has an average density of about 1.3g/cm^3. Nylon has a density of about 1.15g/cm^3, and polyester has a density of about 1.37g/cm^3. Note again that there are many different silks, nylons, and polyesters, so these will vary a bit, but on average these are the densities that are generally common and referred to. This means that a full nylon string would be a bit thicker than its silk counterpart, while polyester will be a bit thinner. And this is exactly what I have seen when making these strings. You can refer to pages 5-8 in the PDF below for material properties of various textiles compared with silk.

The next thing to answer is what material of each do I want to use, and where do I source this material from? Essentially, musical strings, particularly those made of organic materials such as silk or gut, are really just specialized ropes optimized for musical note production. They are essentially just a bunch of tiny fibers twisted together in larger bundles, stretched, glued, and dried under various conditions and manufacturing methods. I knew that monofilament strings already existed and would be the easiest to source and work with for materials, however research has shown that multi-core strings will give rise to a much richer harmonic content than equivalent monofilament cores. This applies regardless of material, and is why instruments with very long, thick strings such as the cello and bass have moved towards these types of helical twisted cores. In addition, while the decay is faster and resonance is less, the strings exhibit much higher flexibility. This is key for metal core strings, which can help overcome inharmonicities present in overly-thick strings, which often results in a dissonant sound. Individual silk fibers from the bombyx mori silk worm are about 13um, or microns, in diameter. To put this in perspective, the average human hair is about 181um in diameter – these are some really tiny fibers! However, it takes many of these fibers to create the substrands of a string, which are further twisted together to make a full string. To figure out how to correlate this to threads of nylon and polyester, I had to find out how many strands of silk are used for a substrand for a string, and how to correlate them together so that I could figure out what diameter of textile thread to get and how many threads were needed to create a string of equivalent size to silk.

To do this I used very simple estimations based on information I could find about silk filament numbers in strings as well as circle-in-circle calculations. On John Thompson’s website, there is a wealth of information on silk strings, the best resource on the internet that I have found so far on silk strings for the qin. One page, found here, http://www.silkqin.com/02qnpu/05tydq/ty1b.htm#strands, has a very useful table on collected information for the diameters of many types of silk strings. These diameters would provide the first basis of calculation estimates. In order to simulate the silk string as closely as possible with synthetic materials, I would want to choose a thread that is as small diameter as possible that I can find – the thinner the diameter, the more threads I can use for a single string. Silk filaments are extremely thin, so they are bundled together in larger numbers to make a silk thread, to be further twisted with similar threads. These numbers can vary, and from John Thompson’s site it is cited that this could range on average from 9 to 12 filaments per thread. Let us pick the 7th string, standard size, as an example for calculation estimates. You can find the table either in the link above or on page 17 of my PDF presentation in the link at the bottom of the page below. This diameter correlates to 0.85mm. Let us also pick a construction technique that just uses a single twisted bunch of threads which are each made from a particular number of filaments. Assuming an average diameter of 13um for a single silk filament, and uniform cross-section (silk has kind of a rounded-triangular cross-section, but for simplicity and rough estimate we go with round), this comes out to 0.013mm. Now we need to choose how many filaments to estimate per thread. If we choose 9 filaments, hen using the circle-in-circle calculator found here, http://www.engineeringtoolbox.com/smaller-circles-in-larger-circle-d_1849.html, we plug 0.013 into the outside diameters of inside smaller circles, and plug an arbitrary number into the box above, labeled inside diameter of outside larger circle box, say 0.1, to get the number of filaments per thread – in this case, it comes out to 43. Our goal is 9, so we decrease the diameter of the larger circle until we get to 9. In this case, we get about 0.047mm diameter for 9 filaments, and for the other estimate of 12 filaments, we get 0.0541mm diameter. Now we go to the calculator again, except this time we plug in 0.85 into the inside diameter of outside larger circle box, and 0.047 or 0.0541 into the outside diameters of inside smaller circles box to get the total number of threads. For the 9 filament, 0.047mm diameter thread, we get a total of 251 threads, and for the 12 filament, 0.0541mm diameter thread, we get 189 total threads.

For the filament count using this method, we get a total of 2259 filaments for a 9 filament per thread, and 2268 total filaments for 12 filaments per thread. These numbers should be equal, but are very slightly different due to the discrepancy in any residual excess spacing between substrands in the calculator. Note though how very different this total number is and the implications for it using a different construction method described below.

Note however that these numbers are for one method of construction, if all of the threads are twisted together as a single bundle. This results in a smoother string, but is less common. I believe Suxin strings use this method, and this is actually the method used for making traditional gut strings for instruments such as the violin. The other way to do this is to create larger substrands, and twist these together first from the small threads, and twist these substrands together to make the final string. This is more along the lines of a “true” rope structure. Generally, you will find that 3 or 4 is a common number. I have found that silk qin and shamisen strings, made from this method, are most often made with 4.  Note that this method will result in a decrease in the total number of threads, which can have large implications on the response and properties of the string, as described shortly below. In order to get the total number of silk threads, we must back-calculate from total diameter, to substrand diameter, to thread number. Let us take 4 substrands for this example of a 0.85mm diameter silk 7th qin string. In the calculator, we first put 0.85 into the top box, and some arbitrary number into the bottom box, say 0.4, which gives us a total of 2 substrands. We need 4, so we increase the number in the second box of the calculator until we reach 4. This ends up being about 0.352, which is our diameter per substring, in mm. We then plug this number into the top box, and our thread size into the bottom box. For a 9 filament thread of 0.047mm diameter, we get 41 threads, and for a 12 filament thread of 0.0541mm diameter, we get 31 threads.

So for our silk estimate for an average standard silk 7th qin string, we have a range of filament numbers to choose from, based on silk thread size. For a string that uses thread substrands made from 9 filaments each, with 41 filaments per substrand, and 4 substrands total, we get a total of 1476 filaments, and for a string that uses thread substrands made from 12 filaments each, with 31 filaments per substrand, and 4 substrands total, we get a total of 1488 filaments. Note the similarly very small discrepancy between these two numbers like in the estimate above due to spacing discrepancy in the calculator. Also note however that for an equivalent diameter string, this method results in a bit more than half the total number of silk filaments used over the first method. This difference between two construction techniques for the same string, assuming the same glues are used for each results in one string NEARLY TWICE THE DENSITY OF THE OTHER, despite being the same diameter, material, and thread size! However, other factors, such as internal damping and flexibility will be drastically altered as well, which all end up affecting the overall timbre of the string. Also note that since more silk material is being used for the first method than the second, the overall cost of the string will increase as well.

We can now apply the same principles to a synthetic thread. I ended up choosing nylon and polyester threads with a diameter of about 0.0048″, or about 0.122mm. This equates to a textile thread size of tex 16, which is the smallest size I could find for these two materials. Note that in comparison with our silk thread estimates, using the circle-in-circle calculator, we find that a 0.122mm nylon or polyester thread will be about the size of 4 bundled 0.047mm silk threads with 9 filaments each, or about 3 bundled 0.0541mm silk threads with 12 filaments each. Again these are not exact figures, but rough estimates for use as a guideline in decision making for strings. With a total string diameter of 0.85mm for a standard size 7th silk thread, plugging this number and 0.122 into the calculator, we get about 35 total threads per string. However, this is NOT our correct answer. Since I am not making these strings with glue, but rather as twisted ropes without glue, I need to first make larger substrands to twist into the final string. For this number, you can generally choose either 3 or 4 substrands. I found that for the material I am using, 3 works the best and results in a much smoother strings, and is all around easier to work with. The number of threads you use per substrand is then found out through a bit of trial and error and optimization, by just plugging in numbers and refining the result until you reach your desired outcome. It is easier in this case if we backtrack from final string, to substrands, to threads. So in our calculator, we plug in 0.85 in the top box, and some random number into the bottom. Let’s just say 0.1, which results in a total of 54 substrands – way more than 3. We keep lowering this number until we reach 3 – for this example, it turns out to be 0.394mm. We then plug this number into the top box, and our thread diameter of .122 into the bottom box. This gives us a total thread count of about 7 threads per substrand, for a total of 21 threads per string, divided into 3 substrands. Yet we are still not done here, and this again is NOT our final answer. Based on the rope making method I employ, there will always be an even number of threads per substrand, since you wind the string between the two hooks for one complete pass, and to tie them together at the same end, this results in 2 threads per pass. Therefore, we need to choose to round this number, in this case 7, either up or down. Remember earlier that I talked about density – since nylon is less dense than silk, it would require an equivalently slightly thicker string. Therefor, rounding the number of threads per substrand up to 8, I now get a total diameter of 0.405mm per substrand, and with 3 substrands per string, I get a total string diameter of about 0.875mm – low and behold, this falls very slightly above our original silk 7th string diameter of 0.85mm!

So, our first estimate for our average standard size 7th string for the guqin made of nylon with a thread diameter of 0.122mm, with 8 threads per substrand, and 3 substrands total, we get a string diameter of 0.875mm. Our back calculations and estimates falls right exactly in the range we want when comparing to its similar silk string counterpart of 0.85mm! This example turns out to be the exact numbers I used for my optimized nylon 7th string, which is noted as Trial #13.

Something else to be mindful of is the total strength of the string. A rope will be strongest with parallel laid lines – so the theoretical max strength of our string BEFORE twisting should be the strength per thread multiplied by the number of total threads. 3 substrands of 8 threads each gives us 24. The nylon I am using has a strength of about 2lbs per thread, so 2lbs x 24 threads = 48lbs total. However, twisting the strands together will weaken this structure, so use this figure very conservatively in your calculations, and plan for enough overhead in strength that your string can handle the tension of stringing. For both nylon and polyester for the strings I have made, I have not had a single issue in breakage yet, even tuning the thinnest 7th string up to modern standard tuning of D3.


THE PROCESS

While I won’t go over the full details in length here, I will provide a detailed overview of the major parts of the process. I am currently working on new posts and full pages dedicated to an in depth detailed look of the process for those who are interested in trying the process for themselves and exploring alternative strings as well. For reference to the process described in this section, you can look at pages 18-23 of the presentation PDF below.

I knew pretty much from the start of my decision to make the strings that I would need to employ some sort of twisting mechanism to achieve the results I was looking for. There is some material on the process of silk string making, for the qin and for other instruments, and knowing that I would need to pursue the second method as described above in the previous section using 3-4 larger substrands twisted together, I started looking into rope-making technology. Essentially, these types of strings are indeed miniature ropes, and the process of twisting a simple rope is the same regardless if it is the size of a thread, a musical sting, or a large utility rope. In this regard, there are many numerous resources and tutorials for constructing your own rope-making machine. I knew that my setup needed to be modified to incorporate a way to: 1.) run multiple threads back and forth over some distance to create a substrand; 2.) be able to twist multiple substrands together in the same direction simultaneously, while keeping them separate until I twisted them together; 3.) have a second mechanism that adds tension to the strings via weight and pulley system; 4.) design a way that I could do the primary and secondary twisting without a need for extra hands, and create a device that keeps the strings spaced and separated until they are twisted together, at a specific tension and angle. I ended up designing and building a very simple 3 part setup that allowed me to achieve my above goals and reproduce my string results to high accuracy and consistency.

The first part of my rope-making machine is a motorized hook assembly that includes 4 rotating hooks which spin at the same time. This arrangement would allow me to make ropes anywhere from 1 substrand all the way up to 4. If you are pursuing string making, and are looking to experiment at this method, I would highly recommend automating the primary twist mechanism – you will need to achieve a couple thousand of turns, and unless you have a heavily geared hand-crank that will result in many turns per rotation, you will be at it for a very long time. Using a motor-driven system also has a key advantage of being able to precisely count the number of turns you place on the substrands, as well as having the option for adding additional automation and control. Instead of sitting there and counting upwards of several thousand turns, there is a very simple way to set and calculate how many turns you put on the substrands for the primary twist. Simply by knowing how many rotations per minute, or RPM, the motor output turns on the hooks, and setting a timer for the number of minutes for the motor to run, you can get the total number of turns. The total number of turns of the primary twist of the substrands is one of the few very critical parameters for accurately replicating your results and fine tuning them. You also need to make sure that your assembly and motor can handle the load of twisting the substrands under weight – the thicker the string you are making, the more weight is needed. For my setup, I chose to use a powerful 12vdc windshield wiper motor with a RPM of 150 revolutions per minute. You can refer to page 18 of the PDF for a front picture of this module.

The second piece of equipment for this setup is a simple pulley tower that can be clamped to a table and run a rope with weights on one end and a swivel hook on the other. You can refer to page 19 of the PDF for a picture of the setup for this piece of equipment. Why do you need a pulley? As the substrands of the rope get twisted, they become shorter – if both ends are fixed, then they will most likely snap before you reach your desired number of twists. Adding a weighted hook on a pulley allows you to also set the exact tension the strands are under when being twisted, which also allows for much greater consistency and control for making strings. For weights I literally just bought the cheapest pairs of hand-weights from Walmart that I could find, the small rubber-coated dumbbells, in various small increments, from 1lb to 5lbs. I fixed these weights to the end of the rope by just tying the rope around them. The weights should be suspended off the floor so that the string starts at the desired set tension. I also made the pulley tower about 2′ tall, to give me additional elevation above the floor when clamped to the table so I have enough rope to account for the shortening of the string. For the hook on the pulley end, it must be able to swivel or rotate freely between the point where the rope and weights are tied and where the strands are threaded through. This is very important so that just the strands are twisted and not the whole rope with weights itself, which would not be able to twist and cause damage to the motor as a result. It also needs to swivel so that the secondary twist can be made on the substrands by hand to lock them together after.

Now, how much does a rope shorten when we twist it? There turns out to be a universal mathematical derivation of this which dictates this number, and purely relies on the number of substrands for the rope. The length of reduction vs. the number of rotations is calculated and plotted out in a parametric curve giving proportions for ideal conditions. Fortunately, we do not have to figure this out for ourselves since this has already been done. On page 15-16 of the presentation PDF, you will find this information. This information was found in a very interesting paper titled “The Ancient Art of Laying Rope” by J. Bohr, and K. Olsen. The link for the full paper can be found here: http://iopscience.iop.org/article/10.1209/0295-5075/93/60004/pdf. This is a very interesting paper and one that I would highly recommend anyone interested or involved in the art of string making to take a look at. Based on the paper, for the ideal zero-twist point of a rope, it’s total length is as follows: for a 2 strand rope with a zero-twist point of 39.4 degrees, the length of the final rope is 63% of the original strand length; for a 3 strand rope with a zero-twist point of 42.8 degrees, the length of the final rope is 68% of the original strand length; and for a 4 strand rope with a zero-twist point of 43.8 degrees, the final length of the rope is 69% of the original strand length. As the number of strands trends towards infinity, the zero-twist point angle trends towards 45 degrees. The zero twist point for a rope is the maximum number of turns for an ideal rope, in which the rope can no longer be twisted without folding upon itself. However, this is again a very highly ideal and perfect scenario – real ropes, made from real materials, are not perfect, and will exhibit stretch and flexibility, which will change these numbers. In my experiments, with the nylon and polyester threads that I use, I found that this number can vary between 77.50%-85.90% for string #7, 77.50%-87.50% for string #6, and 89.17%-92.08% for string #5, given the conditions I used for optimizing my strings. I have also found that the nylon I am using, for a single thread, has a breaking point of about 12% beyond an initial set length, and about 4% for polyester.

Another very key and important thing to note is that all of the number I site for my string parameter optimization is based off an initial starting length of 10′, or 120″. This is the length that I start with for all of my strings, and including the size decrease from primary and secondary twisting, offset by stretch, gives me a final string with plenty of extra length. This extra length is also needed as I have found that I can keep the overall twists of the sting harder and stiffer by tying off a larger amount from the motorized end, rather than right at the looks themselves, since the residual stresses will release the twist built up partially on the string, and the further the tie-off point is from the hooks, the better quality of the resulting string. If you decide to use any length other than 10′ with my initial parameters, found on each string trial page, your results will not be the same! Making the string too long, such as 12′, presents issues with higher probability in strands breaking, and shorter may not give as much excess needed for a reusable qin string. 

The third piece of equipment is something of my own design, which is a modified version of a typical strand separator when making ropes. Generally, with traditional rope making, the strand separator is held in place while the substrands undergo a primary twist, and when a certain point is reached, the separator is pulled towards the twisting side while it is continued to be twisted, resulting in the strands locking themselves together due to the nature of placing multiple twisted strands next to each other – the twisting essentially forces them to twist upon each other and lock into place. Afterwards, the open end is knotted off, securing the shape of the rope, unless adhesives are used along the length of the rope. My separator has three critical functions: 1.) to separate the strands; 2.) to set the twist angle of the strands; and 3.) to provide back-tension to the strands while they are being twisted together. You can view my tensioner in action on pages 19-23 in the PDF. At the moment, in order to achieve the secondary twist which locks the strands together, I must twist this part by hand from the swivel hook at the pulls end. There is much less twist needed in this stage than the first stage, and I do not count the exact number of twists, but go by trial and error to find the optimum starting point, which usually involves twisting it until the separator is all the way to the motorized hook end. I have dubbed the separator for my setup as a “trigonometric tensioner”, for reasons that will be explained as follows. While I am twisting the rope by hand at one end, I cannot move the separator at the same time. Therefore, I needed a way to have it move on its own. This is done by exploiting the force exerted from the strands coming together at a particular angle, which acts on the separator, overcoming the friction holding it in place, and pushes it back. The separator consists of 2 squares of wood with notches on each side for the substrands to pass through, and a bolt in the center that protrudes out to a nut. By knowing the height of the substrands from the central axis, and by adjusting the protruding nut and bolt distance, I can determine and set the exact twist angle I need, which is the angle from the central axis that the strands all come down and meet at, which rests at the very tip of the bolt and nut. Normally however, there is not enough friction to keep the separator from sliding back with even just a little bit of twisting force on the substrands. In order to increase this, and create “frictional back-pressure” to counteract this force, I use a rope tied several times around the substrands between the two pieces of wood and behind the separator, towards the motorized hook side, which acts as a tensioner and increases the friction of the separator on the strings. The amount of tightness that I tie the rope is determined through trial and error, to find the right amount of tension that causes the separator to move on its own when there is just the right amount of tension applied through twisting. Too loose a rope, and it moves too easily, thereby giving me not as tight secondary twists on the string. Too tight a knot results in too much friction and could cause the string to bunch up on itself when twisted and the tensioner does not move. The excess rope has an additional small weight, such as a few large steel washers at the end, hanging suspended above the ground at the center of the length of the tensioner to balance the tensioner. You will want some small weight on the rope to lower the center of gravity enough so that when twisting the strands together, the whole tensioner does not rotate.

The basic process for making a complete string is as follows, for a 3 substrand string. It can be a challenging and tricky process, and I will dedicate a page with pictures for each of the following steps in the future:

  1. Set up the motorized module and pulley module at a distance of 10′, each tightly clamped to a table, and aligned. They can be either at the same height, or the motorized module slightly lower than the tower.
  2. Taking the thread, and facing the hooks directly, tie one end of the thread to the leftmost hook with numerous knots, leaving several inches of excess thread.
  3. Proceed to run this thread all the way the the swivel hook, loop it around, and come back to the hook. Loop around the opposite side of the hook from where you started. So if you tied the thread on the hook and began running it from the right side of the hook, facing the hook, then the return will be on the left side of the hook. This is a single run, making 2 threads per pass.
  4. Repeat this process for the desired total number of threads per substrand. You must make sure that each pass has the same amount of tension as all of the other passes, and should not be overly-tight or too loose, and should not have any threads hanging excessively loose from the other ones. The threads should also be lined up in order one by one on the swivel hook side, with no runs overlapping.
  5. When the final run is made, take the excess thread from the first knot at the motorized hook end and rotate it around to the opposite side of where the thread is on the hook for the last run for that hook, and proceed to tying numerous knots together at this end. Cut the excess thread, leaving about 0.5″ or so excess.
  6. Repeat this process going clockwise on the next hook, which is the center top hook for this configuration.
  7. Repeat this process going clockwise on the next hook, which is the right most hook for this configuration.
  8. Place a preliminary separator between each of the substrands to keep them separated. This will be replaced with the trigonometric tensioner after the substrands have started.
  9. Start the motor. For my setup, with a 150 RPM motor, I let this first run for about 2 minutes to start twisting the substrands together.
  10. After this initial twist phase, I switch out the small separator for the large main one, bringing it as close to the swivel hook end as possible, and setting the tensioner ropes as required. Do not tension too tightly – just lightly enough that there is friction to keep it from sliding too far back
  11. Continue with the rest of the primary twisting phase with the motor. Carefully observe that no threads have broken, and that all the substrands are twisting evenly and are the same tension, and that none of them have failed. Also keep an eye on the timer – the last 1 minute or so is the most critical and is the time that the string is most likely to fail. Failure points will be either at the hooks, or in the tensioner if it is set too tight during twisting. When the time desired is reached for the proper number of tests, switch off the motor.
  12. If the string survived the primary twist phase and each substrand looks good, proceed to fully tensioning the ropes on the separator. Also if the secondary phase requires additional weight on the pulley side, add the weights now.
  13. Hold the one side of the swivel hook with the weighted rope tied to it, and proceed to twisting the other side which the substrands are looped around. Continue to twist in the correct direction until the separator reaches the other end of the string. Carefully watch the string and interaction with the tensioner, and make adjustments to the tension as needed to optimize the tightness of the secondary twists. When the separator reaches the other end, continue to adding additional twists until you feel tightness in twisting the swivel hook – the more extra twists you add at the end, the stiffer the resulting string will be during the tie-off phase, which results in a better sounding string.
  14. Wear a glove with grip and a shirt with sleeves for this next phases to prevent rope burn. Tightly grip the string right at the apex where the substrands come together at the string separator, which should now be at the motorized hook end of the rig. To get better grip, especially with strings with high tension, wrap the string several times around the sleeved arm and glove to increase grip and friction. Cut the individual substrands at the hooks – be very prepared, as you will feel a sudden yank as tension is released from one end and pulled down by the weights at the other end! Absolutely do not let go at this point before the knots are made!
  15. While still tightly gripping the string at the now freed end, with the other hand, choose a point up to 1.5-2 feet away from the cut end, grip and wrap that tightly. Do not let go with the first hand yet –  as you loosen tension on the string with untied ends, the substrands will begin to come apart! Pick a point an inch or so away from the second hand point, and proceed to carefully tying a knot, moving it as close to the second hand as possible. It is absolutely critical that you do not slack any tension in either hands or the string as a whole! It helps if you have someone help tie this knot, and this is possibly the most difficult step of the entire process! Slipping here will potentially ruin the string!
  16. When the knot is tightly formed, slowly unwrap the sting from around your hand and arm. While still holding the string under full tension and not letting the weights go or hit the floor, starts to slowly slide your hands down, facing the pulley end and ending the free end. This stage releases the excess twisted tension in the string and will cause it to spin. This step is necessary to keep the string from twisting on itself. Continue to tightly drag your hands back, from the center of the string to the open end, working your way slowly until you reach the swivel hook end. If you did the past couple of steps right, the resulting string should be a bit stiff and will not curl back onto itself. The end towards the swivel hook will be the stiffer end – this is where you tie the fly-head’s knot that is fixed through the rongkou of the qin. The other end, which is the end you first cut from the motorized hook side, and the current free end, will be a bit softer and flexible – this is the end you use for wrapping around the wild geese feet on the qin. This is very important, and will result in much better tone from the string than if used in reverse.
  17. Side the string off of the swivel hook, while still keeping tension on this part of the string. Make sure to grip the string tightly as close as possible to the swivel hook to keep tension. Tie off the end tightly with a single knot. Proceed to trimming off excess string if needed from the other side underneath the knot: do not cut the string between the two knots! If the string breaks at some point in the future, then just tie of the string a few inches down, and reuse the string like you would silk.

THE RESULTS

If the process is done correctly, you should end up with a very strong, durable, slightly stiff but still very flexible qin string. The process has many steps and parameters, and basically comes down to lots of practice, lots of experience, and lots of trial and error. Once the process is down and solid, iterating and making changes becomes much easier and quicker, and opens up the door for fine tuning the response through the making process. The strings currently do not use over-wrapping or glue, and require knots at each end. However, both over-wrapping and glue could be used as a potential tonal modifiers in customizing the tone of the string, though proper glue and treatment will need to be selected since nylon, and even more so polyester, are significantly less absorbent of water than silk, and are much harder to apply glues to. However, in their current state, while still a bit experimental, work perfectly fine as is, and have been successfully used on both the qin and the shamisen. Due to the way the strings are made, there is a proper orientation, and will result in a slightly stiffer part and less stiff part. The area towards the swivel hook will be stiffer, and is the end you tie the fly-head knot at. The other end, at the motorized hook side, will be less stiff, and this is where you wrap the strings around the wild geese feet.

The strings do have a “break-in” period, which the tone will open up as they are stretched at full tension for a certain period. I find that they work best after about a week of stretching and playing under full tension. Polyester will have much less stretch than nylon, and will keep its tuning better and longer. However, once stretched and played for a while, they should keep tuning reasonably well.

When stringing these strings, they are considerably easier than metal nylon and Longren Binxian strings, and close to the same as silk. Due to their structure, they have an interesting non-linear feel in terms of stretching. When first pulling the strings into tune, they will stretch a significant amount. However, they will become increasingly and more rapidly stiff as they approach their proper tuning. This varies based on material as well as string size and final tuning.

Also note that these particular strings end up being a bit rough, since they are rope-twisted structures. Thicker strings, from 1-4 would not present this issue due to over-wrapping, but 5-7 might if they are not over-wrapped. I have not yet made strings 1-4, and plan on experimenting sometime in the near future on over-wrapping these strings, both to improve physical smoothness, as well as to modify and customize the tone. You can refer to pages 24-30 in the PDF to see microscopic views of various strings, including silk qin and shamisen strings, metal-nylon qin strings, and a nylon and polyester experimental qin string using the above described process. 

In regards to tone, they ended up following my initial predictions quite well: this style of synthetic string falls exactly in between silk and metal-nylon in regards to tone, volume, and decay. Both nylon and polyester twisted strings are noticeably brighter than their monofilament core Longren Binxian nylon composite counterparts, with faster decay, less fundamental, and more mid to upper-range complexity, which is also to be expected. Polyester is brighter than nylon for these types of strings, as expected due to increase density and material stiffness. A look into the harmonic content reveals a very unique response profile when compared with silk, Longren Binxian, and metal-nylon strings, almost as a hybrid between silk and metal-nylon. I find this response to be quite fascinating, and this opens up the door to future prospects of refining and customizing the strings to achieve specific response and timbre. This also opens up possibilities for utilizing other materials to achieve particular results, such as Kevlar, or even carbon nanofibers. As the response of strings will be different for different qin, having more strings available, with the potential to tweak and customize string response, opens the door to even more options for the qin, and providing a wider range of options for players to select from. In addition, after the initial investment in making equipment, the cost for strings is significantly smaller, and further proves that customizable, use-able, and quality strings can be made using cheap, common materials, for a large fraction of the price to what they are offered for currently on the market. You can refer to pages 32-37 of the PDF for a comparison of the harmonic power spectrum and spectrograms between a Yuesheng MN string #7, a silk string #7 from Lawrence P. Kaster, an experimental nylon string #7, and an experimental polyester string #7, all at various tunings.


PDF PRESENTATION FOR REFERENCE

engineering-sound-a-look-into-alternative-string-materials-string-making-techniques-and-harmonic-analysis-of-the-qin


 

Leave a Reply

Your email address will not be published. Required fields are marked *