Designing and prototyping my own high-end headphones

Nov 17, 2024 at 9:57 AM Post #16 of 35
Same as you. I think anode is the best way. But I don't know how to get the smooth anode layer like the big brands. I tried anode but it shows traces of CNC machining on the surface of the part. I'm looking for a way to fix it.
 
Nov 17, 2024 at 10:50 AM Post #17 of 35
On my cups, the rear CNC pattern is kind of random, as I did not give any instructions to the CNC machinist. I wanted to see what he does. But the finishing can be as I want; smooth, straight lines in any direction, circular and as fine as I want. But the finer the lines, the longer the machining process takes for each piece, adding cost. There is always a trade-off.

Anodization only adds a layer of 10 – 15 microns (this is the thickness suppliers in my area, and my particular supplier offer) so any pattern in the aluminium will still show. For me, it is kind of a positive or a selling point, as you want to show the part is CNC machined and not cast, for example. If you look at the Meze Elite or Empyrean 2, for example, the machining lines are quite visible and part of the marketing.

In the end it comes down to personal taste, for me, it is finding a pleasing pattern that the machinist can finish the part off with and then this pattern is still visible after anodization.
 
Nov 30, 2024 at 3:05 PM Post #18 of 35

5. Grille with dust protection


Continuing on a similar theme from last time, now I had to somehow mount the dust protection mesh to the rear opening of the cup. Like many headphones, I did this with a grille. For final testing, I used Plexiglass, but I can use aluminium, steel, carbon fibre or any thin material that can be laser cut, or water jet cut to my design. I made the rear opening of the aluminium cup as big as I could with my design, and designed the grille as an insert that would fit on the inside of the cup. Designing and fitting the grille in a way that no screws were visible (if I decided to use screws) but is still securely mounted. The reason is longevity, but also because many people hold headphones by the grille or apply pressure on it while handling the headphones, so this part must be mounted very securely. I have observed this in video reviews of many headphones and while demonstrating my own headphones to people.

For mounting the grille to the inside of the cups I did not want to use screws, as that would add complexity and weight, so I opted for an adhesive. The adhesive I chose is an odourless silicone adhesive-sealant available in a few colours, that I have a lot of experience with. I know, from testing, the adhesive can withstand temperatures from -40°C to 120°C, temperature shocks, humidity changes and high humidity, without hardening, losing adhesive strength or discolour in any way. It is also a great choice for repairability, as it can be easily cut and cleaned of any material by scraping, and by using the correct solvent, fully cleaned off any surface. More work is required if a repair is needed, but there are always some compromises that must be taken. In this case there were fewer compromises with this, as by adhering the part I also ensure there is no rattling of the grille, as there are no screws to get loose over time.

As usually, I started the design on paper and then moved to CAD. A few iterations of 3D printed parts were needed, to get to the final part on the first prototype. The design of the rear grille follows the location of the magnets on the driver, so the horizontal grille lines are exactly in line with the magnets, while the openings on the grille are exactly in line with the openings of the driver. The grille lines are a bit narrower than the magnets themselves. I did experiment with a few openings, where the lower limit of the lines width, before they are not strong enough, depends on the material and the material thickness. For the second prototype I am choosing a material, that can make the grille as open as possible, without sacrificing durability. For the first prototype, I used 3D printed parts and then Plexiglass parts. Plexiglass is not the best choice, as it scratches easily and if it is black gloss finish for example, there are fingerprint marks visible all over the part. Carbon fibre or a thin aluminium part are a better choice.

With the first version I discovered I made the lines too thin for the material used and found that I had excess material where it was not needed.
1732995251765.png


In the second revision, I made the lines a bit wider and removed the unneeded excess material. The two cut-outs at the bottom are there to make room for the 3,5 mm connector housing. The holes are on both sides, so one part can be made and used for the left or right cup. This is great for lowering costs and simplifying logistics of only having to order one universal part for both sides of the headphone.
1732995259408.png


Below are laser cut parts from Plexiglass with different openings sizes. Protective film is still applied, that is why the parts look scratched.
1732995265441.png


To mount the dust protection mesh, I adhered two grilles together, with the dust protection mesh in between them. The same adhesive is used for this as adhering the grille inside the cups. I made a jig to precisely position the two halves and a cover that goes on top of the “sandwich”, on top of which weight can be added for ~24h so the adhesive cures. Excess adhesive is trimmed in the end.
1732995278422.png

1732995285339.png

1732995290994.png


The end result is a solid part, but a heavy part, as I have two grilles and a stainless-steel mesh in between them. Each finished grille with the mesh weighs 25 g to be exact, 50 g for both sides of the headphone, which is too heavy. Also, the adhering and assembly process is too slow and complex. That is why I am looking at a different material for the mesh and will mount it in a different way, by using only one grille part. The way it is made in this first prototype is very strong, but too heavy.
 
Jan 5, 2025 at 4:55 PM Post #19 of 35

6. Cups


I knew from the start I wanted my cups to be CNC milled from a single piece of aluminium. I researched to find the best aluminium alloy type and grade for my application, checked for general availability for my choice and ordered two blocks, cut to size, just slightly larger than the finished cups.

1736111108643.png


The shape of the cups follows the shape of the driver itself, and the thickness of the cups is mostly determined by the connector I chose, which was the thinnest, high quality, of the shelf connector I could find. Inside the cups, there are posts with threads that hold the driver and the pad holder in place. The first version of the cups had screws visible on the outside of the headphones, which I did not want. So, a few versions later, by going this route (posts with threads) I also eliminated a few screws, washers and nuts at the same time. Now in the final design, 6 screws hold everything together, which also makes assembling the cups quick, easy and consistently repeatable. To mount the headband to the cups, there are two holes on each cup for a screw to go through, a proven and reliable way to mount cups to the headband.

Like for other parts, I started the design on paper and then moved to CAD. By this time, I already figured out the things I described above, so when the first version of the cup was drawn, I quickly 3D printed it. When I got the 3D printed cup I tried assembling it, saw I made a few small mistakes, corrected them and 3D printed again. A few more corrections and prints later, I had the final cup for the first prototype and could start looking for a CNC machinist. Throughout this process of correcting the 3D cups, I also performed listening tests and made corrections regarding that, not only tolerances. For the second prototype, I will make even more corrections, now that I tested the headphones with the actual CNC machined aluminium cups. A big thing is making the cup shape slightly rounder at the back.

One thing I spent quite some time on is the position of the 3,5 mm connector, which sets the angle the cables come out from the cups. The easiest would have been to position it at the bottom, that way the cups would be mirrored, but the connectors and cables would hit your shoulders any time you would move your head, which I did not want. Positioning the connector too far up the cup would put strain on the cable and also pull the headphones forward if a thick and heavy cable would be used. Trough experimenting with 3D printing multiple cup versions, I found an ideal position, not too low to hit your shoulders and not too high to put much if any strain on the cable or pulling the headphones forward.

1736111128240.png


At this point I also had to make a technical drawing for the CNC machinist, review and finalize all my tolerances on the drawing. After finding the machinist I sent the files, he confirmed he can make the parts, and then I just delivered the aluminium blocks. A few days later, the cups were made. Like I already wrote in this thread, on my cups, the rear milling pattern is kind of random, as I did not give any instructions to the CNC machinist. I wanted to see what he does. But the finishing can be as I want; smooth, straight lines in any direction, circular and as fine as I want. But the finer the lines, the longer the machining process takes for each piece, adding cost. There is always a trade-off.

For my personal taste, I like there to be some fine lines visible, so you know you have a CNC machined piece of aluminium in your hands. For me, it is finding a pleasing pattern that the machinist can finish the part off with, and then this pattern is still visible after anodization.

1736111142787.png


For the cups, the raw material mass for both cups was 2,47 kg (5,45 lbs), but realistically slightly smaller blocks at 2,00 kg (4,41 lbs) or even less for both would be enough. After CNC milling, each cup came out at slightly over 90,0 g (3,18 oz). This is why CNC milled headphones, or any other CNC milled products are expensive, it is not the material itself, it is the machining time and skill of the machinist you are paying for. For my cups, I made a few walls way too thick for this application and already thinned them out for my second prototype to lower the weight. The cups are still extremely strong, and I have to save any gram I can, because my drivers are heavy and that is the only part I can’t lower the weight of.

For protecting the raw aluminium, like I also already wrote in this thread, I looked at a few options (anodizing, painting, powder coating and Cerakote), but chose anodizing after looking at the strengths and weakness for all four options. A combination and balance of supplier availability, pricing and durability played a big role in my choice. I found a local service provider, that is quick to perform the service of anodization and accepts orders from one piece to large batches. They are quick as they only offer natural or black anodization of 10 – 15 micron thickness.

1736111161128.png


For the first prototype, I chose natural anodization (this mat silver/ grey colour of my current cups), but like I mentioned in the very first post, I will go with black anodization for my second prototype. Many people seem to prefer black or a darker colour for this type of headphone.
 
Jan 30, 2025 at 5:49 AM Post #20 of 35

1. Pads


Before I started this project, I spent some time researching drivers, finding driver suppliers and ordered samples. At the time and now, my preferences lean towards planar magnetic drivers, so I focused on that. After choosing the driver, I started work on the mechanical parts of the headphones.

The first thing was choosing the size of the pads to be used on my headphones. Because of the driver size, the opening diameter of the pad had to be big enough not to cover any part of the diaphragm or at least not too much of it. I knew the pads would be above 100 mm outer diameter. So, I started looking at my options for pad sizes at 105 mm, 110 mm and 120 mm.

At first, I wanted to go with the biggest size, as I thought it would be the most comfortable and spacious for most ears. What I quickly found out was there are not many 120 mm pads on the market and that did not fit my plan of using off the shelf parts where possible. I also did not want to go this route of limited options, because if I found out I needed a certain combination of materials in the pads to tune my headphones, I did not want to go with custom pads. I also suspected 105 mm pads would be a bit small for some people, but still did some searching for them. In the end, I landed on 110 mm pads.

I kind of knew from the beginning I wanted to go with 110 mm as there are many pads available in this size, from generic pads to pads offered by manufacturers like Brainwavz Audio, Dekoni Audio or ZMF Headphones. They all offer pads in the ~110 mm size that have a lip on the back used to hold the pads on the headphones. That lip is 110 mm, the actual pad is usually a bit larger in diameter, but I will be referring to these pads as 110 mm. (Spoiler for my future post about my pad mounting system, it is a simple ring system with a notch, so pad changes are quick, painless and done by twisting the pad onto the pad holder.)

I went ahead and ordered a few 110 mm pads made from different combinations of materials and with/ without perforations. I still ordered some 105 mm pads just to see how they would fit on my pad holder. In the end, I sized the pad holder in a way that 105 mm pads get stretched just a bit, but as long as the lip is slightly larger, they easily fit. 105 mm pads with smaller lips are a bit finicky to fit, but once fitted are securely held on the pad holder. An interesting observation is the similarity of the opening diameter between 105 mm and 110 mm pads I ordered.

All pads I ordered are non-angled with a round outer shape and round inner opening, so I could firstly focus on comparing perforated/ not perforated and PU leather/ velour. Another reason I ordered round, non-angled pads is my personal preference. All my past and current headphones had and have non angled pads. I did try pad rolling on all my headphones, but always came back to non-angled pads for all of them. Just to explain my reasoning for this choice. When I will have the second prototype and a tuning I like, the idea is also trying some angled pads from the above-mentioned manufacturers to see how the headphones react to that. Audibly to me and measured on my rig.

List of pads for testing:
Outer diameter
Opening diameter
Thickness
Outer material
Inner material
Frontal material
110 mm​
65 mm​
28 mm​
PU leather/ NOT perforated​
PU leather/ NOT perforated​
PU leather/ NOT perforated​
110 mm​
60 mm​
35 mm​
PU leather/ NOT perforated​
PU leather/ NOT perforated​
PU leather/ NOT perforated​
110 mm​
60 mm​
28 mm​
PU leather/ perforated​
PU leather/ perforated​
PU leather/ perforated​
110 mm​
60 mm​
32 mm​
PU leather/ NOT perforated​
PU leather/ NOT perforated​
PU leather/ NOT perforated​
110 mm​
63 mm​
30 mm​
PU leather/ NOT perforated​
PU leather/ perforated​
PU leather/ NOT perforated​
110 mm​
65 mm​
32 mm​
Velour/ NOT perforated​
PU leather/ perforated​
Velour/ NOT perforated​
105 mm​
60 mm​
30 mm​
PU leather/ NOT perforated​
PU leather/ NOT perforated​
PU leather/ NOT perforated​
105 mm​
60 mm​
30 mm​
PU leather/ NOT perforated​
PU leather/ perforated​
PU leather/ NOT perforated​
105 mm​
63 mm​
32 mm​
PU leather/ perforated​
PU leather/ perforated​
PU leather/ perforated​

Some pads for testing:
1729346053993.jpeg

I like the idea of having a high compatibility with many ~110 mm pads, as you can change the sound if you want. Generally, on the first prototype, I found that:
  • Fully perforated PU leather pads are the most balanced with a little less bass but give you the most lifelike timbre and a feeling of a larger sound stage (I like Blues music with these).
  • PU leather pads that are perforated on the inside but not perforated outside or on the front give more bass, timbre that is a bit off and a closed in feeling (Nice for EDM).
  • A compromise between the previous two are velour pads that are perforated on the inside but not perforated outside or on the front. Good bass, timbre and a feeling of a large sound stage (I like them for Rock and metal from the '60s to the '90s).
With further tuning, I want to find an ideal combination of dampening in the cup and pair that with the ideal pads, at least for my preferences. But why not have almost limitless options, similarly to what people like about some brands of headphones. My current personal preference is fully perforated pads. The pads from the picture above are the first set of fully perforated pads I used for over one year for internal testing, testing the fit of the headphones on over 80 peoples heads and for general demonstrations of the headphones. I’m guessing a lot of disinfecting and rubbing down PU leather pads does accelerate their ageing.

As already mentioned, next time I will be describing the pad holder and pad mounting system.
Where did you buy the 110mm spacer? I tried searching on GG but couldn't find anything like yours.
 
Feb 7, 2025 at 2:20 PM Post #25 of 35
Yesterday I just finished my Aegis DIY Tube Headphone Amplifier build, I will be posting in the Aegis DIY thread soon. Now I can focus more on fully finishing the second prototype of my headphones. Next project might be a DIY tube DAC, but I am still deciding if I want to start that or just buy a DAC.
 
Feb 8, 2025 at 9:06 PM Post #26 of 35
Man do I injoy reading this… Kudos to you for even starting such project! I just lack skills to jump into something like this, my dream headphones would be large wooden cups, massive 100mm beryllium drivers, I kinda like JM editions mods, at least he definitely rocks in ‘bang for the buck’ category
 
Mar 1, 2025 at 11:29 AM Post #28 of 35
Mar 10, 2025 at 2:00 PM Post #29 of 35

3. Connector


Because I am into this hobby and generally follow high-end headphones, I knew the most popular connectors for high-end headphones are 3,5 mm and mini XLR, but I wanted to know the details. The reason is choosing a popular connector and wire it internally in the headphones in such a way it will be highly compatible with existing or aftermarket cables for other headphones.

I looked at headphone models from popular manufacturers, specifically at 15 manufacturers and all their wired headphone models from 1000 €/$ to 8000 €/$. Outliers are Abyss Headphones with 2-pole 2,5 mm and 3-pole mini XLR connectors, DCA with 4-pole Hirose connectors, Sennheiser with ODU connectors and Fostex with their 2-pin connectors. What I learnt was, that the most widely used connectors are 3-pole 3,5 mm and right behind them 4-pole mini XLR.

The next thing was researching about these connectors and choosing a specific model for each. I ordered a few pieces of each connector, did some testing, including destructive testing to see from my own tests which are the most durable connectors. In the end, I chose Rean RT4MP and Multicomp Pro MJ-073H. Similar connectors are already known and proven to be high quality and if these two manufacturers chose to discontinue these specific connectors there are the same connectors available from other manufacturers. The choice for me in the end came down to the external diameter of the connector, mini XLR was 11,2 mm, 3,5 mm was 7,1 mm. As the connectors in my headphones are between the driver and the back of the cup, the connector thickness itself would impact the cup thickness. So, I chose the 3,5 mm connectors.

1730651949449.png

Another advantage of going with these specific 3,5 mm connectors is designing the cups and connector mounting system in a way, where I can just screw them in the cup. The cup has a thread in which the 3,5 mm connector screws in securely. Another advantage is easy replaceability if the connector ever gets damaged, the wires get desoldered, the connector unscrews, new one screws in, wires get soldered, and it is done.

1730651966218.png
1730651971835.png

As the connector is round, it does not matter where the contacts land after being screwed in, at least for this specific connector. When the headphone cable is plugged in, the connection is secure without any movement in any direction, but at the same time easy enough to unplug. On my first prototype after hundreds of plugs and unplugs there is really no difference in the feel of the connector and there is still no movement or unwanted wiggling of the male 3,5 mm connector.

The pinout on the headphone side of my 3-pole 3,5 mm connector (TRS connector); the tips are L and R, the sleeves are GND. Before the first prototype was finished, I already made two cables, a simple one for general testing and a nicer one for headphone demonstration purposes. I posted the builds here, for anyone interested in the details:

1730651985389.png
https://www.head-fi.org/threads/diy-cable-gallery.71148/post-17550277

1730651991643.png
https://www.head-fi.org/threads/diy-cable-gallery.71148/post-17582535
Hey do you have any tips when it comes to solder the connectors to the driver? Left and right is fairly simple, but I'm wondering what to do with 4 pin mini xlr. Where do grounds go on the drivers?
 
Mar 10, 2025 at 4:30 PM Post #30 of 35
I am using a 3-pole 3,5 mm connector (TRS connector) and wired it, so the tips are positive, the sleeves are negative or ground. The middle ring is left unconnected. I can then use Hifiman cables for example and many others. I could also use a 2-pole connector, but I chose a 3-pin for wider cable compatibility.

1741635863931.jpeg


For 4 pin mini XLR, pin 1 is positive (+) and pin 3 is negative (-) or ground. Pins 2 and 4 are left unconnected. The numbers are on the connector itself. I know ZMF uses this and probably Audeze, Meze any other manufacturers are using this same pinout. But do some extra research for what you are trying to do. This is how I would do it for a wide cable compatibility.

1741635874968.jpeg


The driver only needs two wires, positive and negative (or ground). No mater if it is a single ended or balanced cable.

Here is a great resource for many pinouts: http://www.diyaudioblog.com/2016/02/headphone-connectors-pins-pinouts-for.html

For the driver. Usually, there are markings for positive and negative on the driver itself, at least the ones I tried. Otherwise, there should be at least a basic datasheet for any driver and in it the polarity is usually mentioned (Peerless drivers for headphones, for example). Even if you buy a driver from AliExpress/ eBay, there is usually a picture in the listing with the polarity.

The last resort, a trick from car speakers, when I had to find out the polarity without taking the speaker out, I tested with a battery. If you touch the positive wire of the speaker to the positive of the battery, the speaker will move forward, away from the frame. But for tiny, delicate headphone drivers, I don't know if this is a good idea.

Hope that helps.
 

Users who are viewing this thread

Back
Top