Designing and prototyping my own high-end headphones

[djentbat]

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.



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.



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.

Thank you so much for the detailed write up, I thought that I would need to solder pins 2 and 4 onto the driver for it to work. I really appreciate the detailed explanation and the resource

 

[martelanc24]

I did see some examples of people shorting pins 1+4 and 2+3 on the 4-pin mini XLR on the cable side for Audeze headphones. I don't know if there is something to that, maybe do a bit of extra research there. As I wrote, I prefer a 3,5 mm connector, so I didn't research much about mini XLR. But yeah, the driver only needs two wires coming from your chosen connector.

 

[martelanc24]

7. Headband​

I wanted to challenge myself to design and make my own custom headband. I knew it would be hard, and it actually was the hardest part of the headphones to design and make, even if the final product looks simple. My goal was designing a headband with as few parts as possible, that is attractive, durable and could be easily made within my skill set and current knowledge. Made “in house” or ordered from a supplier. The option of just using an off-the-shelf headband was always there, but I challenged myself to go through with designing my own headband. I am working on a version of my headphones that do use an off-the-shelf headband, and I will share that here in the near future. For the headphones after the second prototype, I will continue refining this custom headband.

The first idea was making a headband with a spring steel arch at the top, on which two CNC milled pieces are mounted. These two pieces would have smooth rods going through them for sliding the cups up and down. A system similar to Grado headphones, early ZMF headphones, Meze Empyrean headphones, and many others. After buying some rods, bushings and other parts, in a month of designing and testing I concluded this system would be complicated and expensive to manufacture by me. One big issue would also be getting the tolerances of the rods and the hole or bushing to be just right. On one hand, with tight tolerances, the rods would barely be able to be moved, on the other hand with loose tolerances the rods might just slide down, which would be very annoying.

With the original goal of making a headband with as few parts as possible, I went in the direction of cutting it from spring steel on a water jet cutter or laser cutter. The first updated idea was a three-part headband, where a centre top piece would hold two smaller side pieces that would somehow slide up and down the centre piece. After some days of thinking and drawing how to solve this, I realized I need to simplify it further. In the end the simplest and realistically manufacturable by me was a headband from a single piece, bent in shape, heat treated and painted, or powder coated in the end.

After determining the final concept, I had to think about how to mount the headband to the cups. As I initially did not want to bend the headband in a complicated shape, I had to find an off-the-shelf part to mate the two together. In the end I found a part, an M3 bearing with M3 thread, I just had to shorten the thread as it was too long. It is not the prettiest mounting hardware, but it enables a smooth and completely silent movement of the cup. So, on the head there is zero squeaking, grinding…, while using the headphones. In the future, I will experiment with bending so I can just mount the headband straight to the cups without these bearings.

After choosing the way I would mount the cups to the headband, I started with the detailed design of the headband. The headband has a relatively simple shape, the centre part is thin with a long hole in the middle and at the sides it breaks out in a fork. I made a few models, and then the first step was choosing the “fork” shape. I 3D printed two versions to better visualize the shape and made a choice.

When I chose the shape, I drew the full headband. The model is made in a way where I can just shorten or lengthen the centre part and get a different length that way.

To decide on the initial length of the headband, I measured a few popular headphone headbands like Sennheiser HD 580/ HD 600 and Beyerdynamic DT 770 Pro for example. In addition to that, I went through some research papers and studies, where I found head sizes and other statistics concerning head measurements. Even after all that “benchmarking” and reading, I made the headband too big. The very first headband I had cut, I thought would be too small, with my interpretation of the research in the papers, but it turned out way too big. After it was cut, even If I knew it would be too long, I bent it into shape as I wanted to actually see how it would fit and look with the cups.

After this I still tested this too big headband version on over 50 people, observed the fit and asked questions about fit to all of them. (In addition to sound impressions.) I took notes which helped me decide and amend the final length, shape, holes positioning for the sliding comfort strap, spacing between the cups that affects the clamping force and comfort strap length.

Another thing I did was inverting the mounting position of the headband on the M3 bearing from the outside towards the inside. Now I have the final measurements, geometry and corrected CAD model for my next prototypes. I can now experiment with further bending, heat treatment and choosing a painting method where the paint does not chip with repeated bending.

For my future prototypes, I will continue work on this headband and refine it further. But for my second prototype, as I said, I went in a different direction, by going with an off-the-shelf headband.

 

[martelanc24]

8. Comfort strap​

From the very beginning, I wanted to make a headband with a comfort strap, even before I started designing my headband. For high weight headphones, like my first prototype, I feel it is the best option as a lot of surface area is actually touching your head and spreading the weight as much as possible. One more advantage of an as big as possible surface area touching your head is you minimize the problem of a hot spot on your head. Also, as my headphones have a light clamping force, most of the weight is carried by the strap.

The shape of the comfort strap for the first prototype is simple; wide centre, tapering down to where the screws holding the strap to the headband are. The material is soft, yet strong leather that does not stretch with time. With over one year of almost daily use on my current strap, it did not stretch around the M4 screws or anywhere else.

The strap is mounted on the headband with two M4 screws. At first, I mounted it straight to the headband like in the picture below, but I ran into a problem. When I loosened both thumb nuts to adjust the strap height, it would just rotate to the front or the back and flop down.

To solve this issue, for a first test, I designed, and 3D printed a strap stopper, that would fit over the nut and hold the strap upright when I loosened it to adjust the height. After I confirmed the concept, I designed a similar, but simplified piece that can be milled out of aluminium for the next prototypes.

To prevent the strap sliding up while the headphones are on your head, there are some M4 PVC washers on the screws that hold the strap in place after the thumb nut is hand tightened. It is not the most advanced system, but it holds the strap in place as it never slipped out of position and offers “infinite” adjustability. You do need to take the headphones off your head to set the height.

This was the last post about my first prototype, next will come the second prototype. It took me a while to decide in which direction to go with it, and I will soon share what I have been working on.

 

[martelanc24]

DIY R-2R DAC​

I should have maybe named this thread a more general DIY thread for my projects, but my headphones are my main focus and everything else comes as an occasional build, so I will leave it. All this is still connected to my headphones, as I want to get the most out of them and have previously built an Aegis headphone amplifier partly for that reason. The next major thing was upgrading my DAC.

I wanted a single ended DAC for under 1.000 €/$ to use with my Aegis headphone amplifier, as I also have a preamplifier output on my Aegis, with which I then feed two other single ended amplifiers. As there are not many commercial single ended only DACs under 1.000 €/$, after much research and reading about DACs, I ultimately decided to build an R-2R DAC, using a Soekris dam1021 DAC module (0.01%/0.02% resistor version).

It is relatively simple to build this DAC, you need a dual rail power supply, +/- 9,0 V is recommended in this case, some wires, additional components and put everything in an enclosure. I wanted a linear power supply with the transformer built in, to make the build simple and quick, so someone recommended this power supply to me, as they already measured it in the past and found it to be very good.

In the back, I have a Neutrik powerCON connector for power input, an ON/OFF switch and a green LED to indicate the DAC being turned ON. For most internal power connections and RCA output connections, I used NEOTECH SOCP-20 UP-OCC copper PVC 20AWG wire. The switch is on the back because the DAC is turned ON most of the time, being an R-2R DAC and having very low power consumption. For digital inputs, I have a BNC connector, as I am connecting to the DAC from a DDC with a BNC cable. Outputs are two pairs of RCA connectors, coming right off the board, the bottom pair are 1,4 V; Z-out 640 R and the top pair are 2,0 V; Z-out 10 R.

I also made a shielded power cable with a powerCON connector, using OLFLEX 110CY 3x2,5mm2 cable, like I did for my Aegis build.

To decorate the enclosure, I got a piece of leather and cut it to size for the front panel. It adds something visually interesting, at certain angles, to an otherwise plain black metal box.