The Soul of a Music Box- The Comb

The Soul of a Music Box - The Comb (Vibration Plate)
UPDATE: 2024/09

The soul of the music box lies in the comb, as its material, thickness, shape, and processing methods all influence the final sound quality produced during playback.

This article will take you deeper into the connection between the melodic features of music boxes and the design of the damper, as well as our experimental process dedicated to optimizing the durability and sound quality of the comb. From the development of the N20 music box with twenty notes in 2018 to the N40 music box with forty notes developed in 2023, the design of the comb has evolved to meet the expectations of professional musicians and top music box collectors.

The Feature and Limitation of a Music Box

Have you wondered why music boxes always play slow melodies? Is it because the slow melodies sound more romantic or it is simply due to its mechanical design’s limitation?

I think the real reason is that the comb of a music box has its limitations in mechanical design, so it is not suitable for playing fast melodies that repeat plucking the same notes, which will generate a sharp clicking sound. Therefore, when a melody has to pluck the identical note many times, its tempo needs to be adjusted to slow it down in order to make it produce a pleasant sound.

For example, in the following video, we conducted a test of the music box striking the same note continuously. Although the melody speed in the video is not fast, you can already hear noticeable noise in the test without the damper, which we refer to as continuous noise.

The Purpose of the Damper

The continuous noise in the music box is caused by the metal pin of the next note contacting the vibrating comb, producing a metallic clashing sound.

The function of the damper on the comb is to work as a cushion to reduce the noise from the pin’s collision on the vibrating tooth of the comb.

In other musical instruments, for example, a piano has its own unique damper pedal for the similar function. It is responsible for the end of a note, to make it a smooth sound.

However, because the size of a music box is smaller and simpler, there is no perfect solution for its damper design yet. If you are interested in learning more about how a damper works in a piano to maintain its sustaining sound effect, please watch this video:

The Limitation of the Damper

If the damper solves the clicking sound problem, why do music boxes still play slow melodies? It is because the damper only alleviates the problem. For example, with a damper, it may reduce the shortest repeating interval of notes from 2 seconds to 1 second, but it is still not enough for a music composer’s needs.

The damper is subject to wearing. The most fragile part of the music box is the damper. A Sankyo branded paper strip music box has an official life span of 400 hours due to the damper start wearing. The music quality dropped after the wearing happened.

The following picture is a close-up shot of the damper from the music box brand “NICOLE FRÈRES”. The metal string at the tip of the comb is the damper. From the photo, we can see that many of the dampers are damaged. The durability is a great challenge in music box’s damper design.

Can We Fix the Problem Without Using Dampers?

Ok, if the damper is an issue, are there any other solutions that can be used instead? I believe the answer is Yes. For example, if production cost is not an issue, the bigger and more expensive music boxes have dual-disk or dual-comb design, and I think it is aimed for the problem we are discussing. To see a music box with multiple combs in performance, please watch the following video: https://youtu.be/AUmz4GnxuB4

It is an intuition that double the comb can double the volume, however the true benefits are far beyond that. For my understanding, the major goals for the multiple combs are for two purposes: consecutive notes and sublime harmony.

It is related to consecutive notes because the following note can be plucked from a different pin/comb, and it avoids the problem of the following pins from collision on the previous vibrating metal tooth.

The Muro Box is structurally capable of utilizing this design, but ultimately it was not adopted. The simple reason is due to cost issues. We have indeed received many customer requests for this development, and if there is an opportunity in the future, we will execute it. Stay tuned!

Sublime Harmony

Sublime harmony, on the other hand, is intended to enhance special sound effects. The luxurious music box had the feeling of the “vibration” of the tones that oscillated in the space. This is not just because it has a better quality wood case, it is mostly due to the use of two teeth that are not exactly tuned at the same frequency but are slightly out of tune.

More accurately speaking, two teeth are different in exactly 14 cents (100 cents = 1 semitone). If you are interested in this topic, please take a look at the following video. Our N40 Sublime music box uses this technique in our comb design. (Thanks to Siegfried Pattyn for providing this great suggestion, if you are interested in why he can give us such professional advice on mechanical design, you are welcome to read Siegfried’s story.)

Muro Box Uses a Custom Version of the Comb

The comb used in the Muro Box is actually based on the comb of the Sankyo brand 20-note paper roll music box produced by Kyooh Precision Industry Co., Ltd.. We customized their combs and upgraded them for Muro Box music boxes. The main improvements focus on two aspects: the damper designed to solve the continuous striking noise issue, and the brass tuning weights related to the N40 tuning.

We have invested several years and significant effort into the development of the comb. This article records the various developments surrounding this small comb.

(Note: This article was updated in September 2024, including changes made to the comb in the product development journey of the N40 music boxes.)

The Development Process of the Comb (Vibration Plate) of Muro Box Music Box

Regarding the speed limitations of continuous strikes in music boxes, we consulted with music arrangers during the initial design phase and established a standard of four consecutive strikes per second for a single note. While this speed is not sufficient for playing heavy metal rock, it is more than adequate for most pop songs.

Indeed, we have received numerous requests from customers to further allow the continuous strikes of the same notes to achieve more than 4 beats per second. Achieving four strikes per second without noise is already a significant challenge. Initially, we did not fully realize the difficulty of meeting this standard; otherwise, we would have set a slower speed limit for the product development to go on more smoothly.

There is no turning back, we have to move on. Because the consecutive notes noise has a great impact on the experience of music listening, we have people leave comments in our YouTube video questioning if the music box is broken to have this noise.

The following video clip shows our earlier N20 music box prototype using Sankyo’s comb without adding our newly designed damper to reduce the consecutive notes noise.

In the beginning, we misunderstood that the cause of the noise came from some unknown defects of the electromagnet, and we made an assumption that it will disappear naturally after we optimize the structure of the cylinder. Unfortunately, the problem finally broke out in our last stage of the mass production and greatly delayed the product delivery.

That was devastating to our team. We have spent years on making a programmable cylinder for Muro Box, now we have to make the comb on our own, too? If the problem cannot be solved, then we have to reduce the max speed of the melody, then this music box makes no difference to the model made 100 years ago. What should we do?

How Muro Box Solves the Problem

Since our vibration plate was initially designed based on Sankyo’s paper strip music box, we were aware from the beginning that the original damper design from Sankyo was insufficient to meet our needs. Initially, we attempted to slightly alter the shape and thickness of the plastic damper in hope of enhancing its shock-absorbing effect. However, we quickly realized that this approach was not feasible. If simply modifying the shape could solve the issue, the original designer would have already implemented it.

We then have a second thought on the material of the damper. The original damper is made of PET. How about changing the material to something with better vibration absorption? When the key material was changed, it also entailed that we had to spend time finding a new supplier to test more samples.

With the help of a die-cutting factory, we designed a new mold and started to test various kinds of plastic materials. PVC, Mylar, PC, PEEK…, we have evaluated and tested everything in their factory and nothing worked. Fracture, wearing, deformation, and weak bonding, all kinds of issues occurred and more than you can imagine. We have exhausted every option and the result is worse than the original PET. We have to believe the original designer’s wisdom that they already found the best solution. We still have to use PET.

The First Major Failure

Finally, after endless going back and forth between all sorts of problems, Muro Box finally embraced its first official shipping date. (This was our first ZecZec crowdfunding in Taiwan in 2018 before the 2020 Indiegogo campaign). Just a few days later, the first report from our customer brought the bad news. The damper fell off from the tooth of the comb. Our UV resin (the glue to attach the PET) cannot subtend the extended music playing scenario in Muro Box.

We immediately stopped the product shipping, and put the Muro Box back to the stress test table to fix the quality issue. Our new stress test standard is to continue playing the Muro Box for 16 days, 24 hours per day. This length of stress test time was suggested by Kyooh.

We started experimenting with all kinds of industrial adhesive gum. From Super Glue, Epoxy to UV resin, we tested every possible kind of glue we can get. However, the working condition of the damper is extremely hard. All kinds of extreme conditions, like tiny contact surfaces for the glue, excessive duration of vibration of the comb and rubbing by the cylinder’s pins. The most difficult part is the joint of dissimilar materials of plastic and metal. All methods have failed.

Our entire team is feeling demoralized. It’s really hard to endure the endless challenges and setbacks in product development over the years, and we never expected that even at the final moment before shipping, this major obstacle wouldn’t let us go.

Design the Metal Damper from Scratch

We already spent a few months on the damper. We started from the original design of Sankyo’s damper, replaced it with various materials, and still unfortunately did not reach our acceptable quality standard. We finally understood that we have to think out of the box to get through this nightmare. We decided to abandon the existing design and started a new design from scratch.

Since we have completely abandoned the previous damper design, there are no reference materials to follow; all we can do is test, test, and test again. I have conducted crazy experiments with various materials: stainless steel, carbon steel, brass, copper, bronze, beryllium copper, titanium alloy, and even various combinations of electroplating. I often think that perhaps my research training during my PhD was meant for this moment.

I know others have begun to look at me strangely; they do not understand the connection between a small damper and a music box. They only see someone who has been sitting here for six months, surrounded by piles of metal scraps, with no clear idea of what I am trying to accomplish.

Imperfect, but Enough to Keep Me Moving Forward

Six months have passed, and although the defect rate is still very high, we have finally identified a combination of damper materials and assembly methods that allows us to eliminate continuous striking noise while maintaining sufficient durability. Although the comb still occasionally produces abnormal sounds for unknown reasons, we have managed to filter out defective products, reaching a compromise between quality and yield.

Starting Over from Scratch Again

Time returns to early 2024. As we made adjustments to the comb during the development of the N40, we unexpectedly encountered a resurgence of issues with the damper. Previously, we could resolve the high defect rate by selective sorting, but now the new design of the brass tuning weights in the N40 model changed the weight distribution, and with two combs in a N40 Standard music box, and four combs in a N40 Sublime music box, the overall assembly defect rate increased significantly, making rework no longer a solution.

The real root of the problem was our lack of understanding of comb design. It was only through years of customer feedback that we began to grasp what good sound should be like, and this understanding came to us after the N40 music boxes were shipped to a small batch of backers. Because the selling price of the N40 did not allow for any compromises on sound quality, we found ourselves in a predicament of substandard vibration plate quality, endless rework, and an inability to ship products.

Revisiting the Design of the Damper

In desperation, I returned to the front line of production. For three months, I endlessly assembled and disassembled the dampers, hoping to discover the pattern behind the noise.

We couldn’t understand the problem; why did this particular comb produce noise when it appeared identical to others? All we could do was repeat the assembly process, hoping to find hidden clues about the cause of the noise.
The inability to ship out products soon had become so severe that it was demoralizing. Employees had developed an unspoken agreement to avoid discussing shipping issues. They knew I had been sleeping in the office without going home for several nights, and they didn’t want to add any extra pressure on me.

Despite not being able to solve the problem, I considered lowering our standards for shipping them sooner because, in blind tests, most people couldn’t detect the noise issue. As long as we didn’t mention it, there was a high chance there wouldn’t be any problems. To my surprise, our quality control personnel stepped up and refused to allow us to ship it, and their insistence left me both amused and exasperated. All I could do was to continue searching for new solutions.

Explanation of the Damper Issues

We are facing a dilemma with the damper issue. While we want to increase the shock absorption strength, the damper itself can also create noise. Conversely, to reduce the noise generated by the damper, the shock absorption strength becomes insufficient.

Our design uses a metal spring as a buffer for the contact between the comb and the striking flexible pins, where the size of the “shock absorption capability” depends on the size of the front opening. A larger opening allows the spring to deform more, absorbing more impact force. However, if the opening is too large, it becomes challenging to control the relative position of the spring and the front end of the comb, which can lead to noise.

In simple terms, if the damper protrudes too much, it will extend the comb too far, resulting in louder sound. This problem is relatively mild, as it only sounds louder and does not reach the level of noise. On the other hand, if the position is too shallow, the striking pin will first hit the damper, and then after disengaging, will strike the comb a second time, producing a noticeable abnormal sound that is unacceptable.

All of this occurs within a range of 0.1 mm at the front end of the comb, and we should not underestimate even a small amount of abnormal sound. The human ear is much more sensitive than the eye, and this 0.1 mm distance can definitely be detected.

Past Attempted Solutions

We have been struggling with how to precisely control the position of the damper right at that sweet spot. Each N40 Sublime has 44 dampers installed across its 44 notes, and even if the failure rate for one note is only 1 in 10, the probability of successfully assembling all dampers correctly in one go is only (9/10)^44 = 0.00969, which means there is less than a 1% chance of success. This is incredibly low, and it has led us into a cycle of endless rework, which has been our previous predicament.

Note: The N40 Sublime uses 4 sets of 20-note, totaling 80 tuning forks. Each tuning fork only has space for dampers in the lower octave, accommodating 11 notes for installation. Therefore, no dampers are installed in the higher octave, resulting in a total of 44 dampers for each N40 Sublime.

The Final Solution for the Damper

The key is to let go of the previous fixation on positioning. Since we have reached the limit of controlling the position, we decided to abandon that approach. Instead, we focused on another cause of the noise: the two impact sounds.

If the issue is the double impact, we simply need to find a way to make it a single impact. By changing the shape of the damper, we can ensure that it ‘cannot’ hit the music comb again.

In simple terms, we raised the front end of the damper to ‘wrap’ around the front end of the vibration plate. This way, the vibration plate has no chance of coming into contact with the striking pin; it will always contact through the damper.

The Process of Discovering the Solution

It may seem logical that we arrived at this solution step by step from the problems we faced. However, that wasn’t the case; this solution was discovered by chance.

The reality was that we went through repeated failures and endless rework, causing the production team to be on the brink of collapse. At this point, Mr. Wang from the metal stamping company Airwise suggested that we must address the issue fundamentally. Thus, we began modifying the damper’s mold for new testing. The first design we tested, as shown in the picture below, was modeled after the curved steel wire design of classical music boxes. The failure of this design was due to the damper vibrating and producing sound after impact, and simply relying on the metal piece’s damping effect was insufficient.

After I reported the test failure, I realized that our original design was actually closer to success. I proposed a solution to bend the front end at a small angle based on the original design, allowing the striking pin to “slide smoothly” along the damper to the vibration plate, thus reducing the sound of the second impact.

However, to my surprise, Mr. Wang (the owner and CTO of Airewise Ltd., a metal pressing factory) misunderstood my description and bent it in the opposite direction, creating an almost vertical angle that caused the striking pin to hit the vibration plate directly. As expected, the results were worse than before; the impact sound nearly drowned out the music, making it completely inaudible.

This test required a significant investment in mold creation, and the outcome left me in a daze. One failure couldn’t simply end the process—I needed to provide Mr. Wang with an explanation. With a determination to find a solution, I took the new sample and pondered whether there were any other ways to break through the dilemma. Eventually, I recalled a principle that had been repeatedly validated during our journey: “Solutions are often hidden in impossible methods.” Since we cannot change the strange shape of the damper, why not extend it just a bit more to fully encompass the vibration plate?

To my surprise, this incorrect bending angle opened up a completely different way of thinking. Ultimately, the solution of having the damper envelop the vibration plate was discovered accidentally, emerging from the failed design.

Design Changes to the Vibration Plate in the New N40 Model

The design changes in the N40’s comb mainly involve two aspects: the selection of musical scales and the addition of brass weights.

For the selection of musical scales, we chose non-continuous scales to allow traditional 30-note paper music boxes to be directly used on the N40. Additionally, to enhance the bass sound, we added weights to the comb teeth.

The biggest difference between the N40 Sublime and Standard is that both have 40 musical scales, but the Sublime has 80, producing a sublime effect during performance. What is the principle behind this?

Here, we will disclose the development process of the special design of the N40 vibration plate.

Let's listen to the unique “Sublime Harmony” effect in mechanical music boxes!

The following video plays the same melody “My Grandfather’s Clock” on three different Muro Box music boxes.

0:00 N40 Sublime
1:07 N40 Standard
2:15 N20 Standard

To learn more details about the differences among the three models, please refer to “which music box suits me”.

Should a New Comb (Vibration Plate) Be Developed for the N40?

First, we would like to thank Mr. Siegfried Pattyn for introducing us to the concept of “Sublime Harmony”. To put it simply, the principle of Sublime Harmony is to pluck two comb teeth with slightly offset frequencies. Due to the physical properties of sound waveforms adding and canceling each other out, the harmonics will fluctuate in volume depending on the frequency difference, creating the illusion of a sound that seems to vibrate.

When we first heard about it, we were actually quite confused and did not understand how it worked, particularly why a -14 cents offset would produce a sound that vibrates. Therefore, we initially kept this concept in mind without implementing it.

The original plan for the N40 involved commissioning Kyooh Precision Industry Co., Ltd. to develop a new 12-centimeter long comb that could support 40 different pitches. Of course, the cylinder would also need to be redeveloped.

However, after some careful consideration, it suddenly dawned on us that the current design of the N20 cylinder could directly utilize the existing sound comb and cylinder mechanism by reversing the installation of two vibration plates, thereby achieving the configuration for Sublime Harmony. This realization indicated that redeveloping a 12-centimeter sound comb was unnecessary.

Will Redeveloping the Comb (Vibration Plate) Result in Better Sound Quality?

What are the differences in sound based on the length of the vibration plate? Would a 12-cm long vibration plate be better than two 6-cm long vibration plates? My conclusion is that the total width of the comb itself is not the key factor affecting sound; rather, the tooth shape of each note is crucial to the sound characteristics.

The comb used in the Muro Box originates from the 20-note paper strip music boxes of Sankyo, which gives the Muro Box a relatively thick sound characteristic:

1. The width of the tooth is wider (3 mm) to match the hole width of the perforated paper.
2. The original design aimed not to add extra weights; thus, the teeth are intentionally designed to be thicker to increase their own weight, with thinner root parts.

However, high-class music boxes with more notes included from the Sankyo brand do not have these restrictions, so their tooth widths are usually narrower, and their bases are thicker, resulting in a more piercing sound characteristic.

Once I finally understood the principles of sound characteristics in the comb, I no longer insisted on redeveloping a 12 cm 40-note vibration plate. Due to our structural limitations, the spacing for each note is fixed at 3 mm and cannot be changed, and currently, we cannot use weights on the comb. Therefore, the thickness of the teeth will not change significantly. As long as these two points remain unchanged, the sound effects of two shorter combs will be similar to that of a single longer comb.

Final Decision: The Multi-Comb Solution

Due to the rarity of music boxes with multiple combs, our first consideration was why existing music box models do not adopt this design.
Multi-comb configurations are actually quite common, but their higher unit price results in lower circulation, making them less visible in the market. For example, the image below shows a music box housed in the Modern Music Box Museum in Taichung, which consists of two bases with 50 notes each, forming a 100-note music box. This design is clearly created for Sublime Harmony. The price for this music box is NT$110,000 (about $3419 USD).
After assessing the development challenges and risks, we decided to use the existing N20 music box vibration plate as our starting point, designing the new N40 model with Sublime Harmony as the main focus.

N40 Standard’s Music Scale Design: Discontinuous F3~C7

We proposed different scale options during the design phase, as shown in the image below, which was presented to customers for voting.

Initially, we considered using Muro Box-N40(0) for the scale selection. However, after considering the need to support existing 30-note music box pieces (with Music Box Maniacs being the largest library for music box melodies), we ultimately changed our plan to Muro Box-N40(1).

The scale features a continuous range from D4 to C7, encompassing every whole and half step, while the bass section from F3 to C4 is discontinuous, omitting certain half steps.

With a wide range of half steps, the 40-note music box can more accurately reproduce full songs and offers more options for original compositions. The application of melodies and harmonies expands the usable notes. However, the 40 notes still have their limitations, so we have chosen C major and A minor as the two scales most suitable for music box arrangements. These two keys are the most commonly used in simple notation and are also the most beginner-friendly.

Of course, you can still create songs in other keys with the 40-note music box. Notes that are not available in the bass range can be found in other octaves.

The 40 notes on the combs of a N40 music box are arranged in an alternating pattern rather than a sequential one.

Did you know? The two combs of the N40 Standard or the four sound combs of the Sublime are arranged in an interleaved manner. The left side starts with the low note Fa, and the next note—low Sol—is placed on the sound comb on the right, and so on. The 40 notes of the N40 are interleaved rather than arranged sequentially like in standard keyboard instruments.

This arrangement is not a strict requirement, but it is the most suitable method for tuning, making it the optimal arrangement for production.

Please take a look at the following animation demonstrating the “interleaved scale arrangement”:

Note	2	4	6	8	10	12	14	16	18	20	22	24	26	28	30	32	34	36	38	40
Note Name	G2	B2	D3	E3	F#3	G#3	A#3	C4	D4	E4	F#4	G#4	A#4	C5	D5	E5	F#5	G#5	A#5	C6
Original Frequency	196.0	247.0	293.7	329.6	370.0	415.3	466.2	523.3	587.3	659.3	740.0	830.6	932.3	1046.5	1174.7	1318.5	1480.0	1661.2	1864.7	2093.0
-14 cent Frequency	194.4	245.0	291.3	326.9	367.0	412.0	462.4	519.1	582.6	654.0	734.0	823.9	924.8	1038.1	1165.2	1307.9	1468.1	1647.8	1849.7	2076.1

Note	1	3	5	7	9	11	13	15	17	19	21	23	25	27	29	31	33	35	37	39
Note Name	F2	A2	C3	D#3	F3	G3	A3	B3	C#4	D#4	F4	G4	A4	B4	C#5	D#5	F5	G5	A5	B5
Original Frequency	174.6	220.0	261.6	311.1	349.2	392.0	440.0	493.9	554.4	622.3	698.5	784.0	880.0	987.8	1108.7	1244.5	1396.9	1568.0	1760.0	1975.5
-14 cent Frequency	173.2	218.2	259.5	308.6	346.4	388.8	436.5	489.9	549.9	617.3	692.9	777.7	872.9	979.8	1099.8	1234.5	1385.6	1555.4	1745.8	1959.6

Development Details of the Brass Weights

Because the N40’s lower range notes have insufficient weight on the teeth, tuning them directly without additional weight would make the teeth roots too thin, resulting in inadequate strength and making it prone to deformation. Additionally, the bass sound would be weaker. Therefore, we aim to add brass tuning weights to the two lowest notes on the comb.

This approach raises two questions:

(1)Why use brass?
(2)How will the weights be secured to the teeth?

The Main Reason for Using Brass is That Lead is Prohibited.

The reason for using brass is that we cannot use processes involving lead; hence, lead cannot be employed. Brass, being a denser material that is readily available and has stable processing properties, is commonly used. However, not using lead affects the fixation method. The traditional way of combining weights with the teeth is either through epoxy adhesive or soldering.

We cannot accept the adhesive method, as it is a temporary fix rather than a stable mass-production solution.

We also cannot use traditional soldering methods. The reason is that the teeth undergo heat treatment to enhance hardness, and excessive heating can cause annealing, softening the material and affecting the sound. Traditional use of lead is feasible because its melting point (327°C) is lower than the annealing temperature of the teeth (400°C). As long as the temperature is controlled, soldering half-melted lead can be done without annealing the teeth. However, brass has a melting point exceeding 900°C, which makes it impossible to melt the copper pieces below the annealing temperature. Moreover, using only tin solder to secure the connection lacks sufficient strength, resulting in joints that cannot withstand repeated vibrations and may loosen.

To achieve this, we proposed a new method for securing the brass weight by using a mortise joint. This involves drilling a hole at the front end of the sound teeth, allowing the brass to be inserted and directly locked onto the teeth.

Epilogue

We are fortunate to receive encouragement and support from many people, and endless support from our key suppliers. We redefined the manufacturing process and material of the damper to make it beyond the capability of the existing one. We have patents of the new damper design in R.O.C (M594783) and China (202020038579.1).

Although only a part of the damper’s development process is documented here, we believe that you should be able to realize how much effort is required to design a good damper to resolve the noise from consecutive notes.

This is why the Muro Box is such a valuable possession. It is a gift from so many people’s endless effort, and only comes to life after countless miracles happen.