LYRE: The Story of a DIY headphone
Hi everyone!
After a long time working on the headphones, I got to the point, where I think the headphones are perfected enough to be considered finished.
They went through a lot and I'd like to share it here now. And I'll be honest, I did forget about this thread, but I hope the story will still interest you all.
Please keep in mind, there are a lot of graphs measured with the MiniDSP EARS with HEQ calibration - this as a measurement rig is far from accurate.
1 Drivers
1.1 Single-Sided Magnet Array Prototypes
1.1.1 First Driver Prototype
The first driver was a very interesting and complicated process. Thankfully, I am lucky that my school allowed me to use their equipment to make the drivers with, otherwise I wouldn't be able to prototype the drivers this easily. When I made the first driver, I honestly wanted to cancel the whole project - it was muddy, undetailed and had no upper treble. Thankfully, I told myself that it's the first driver and it's never perfect on the first try.
The first driver was a 106mm failure. It measured an impedance of 52 Ohms and sensitivity of 87.4dB/mW. The full frequency range was surprisingly 9Hz - 32.1kHz. If I wanted a truly good headphone, this was not something I could use, nor let anyone listen to.
The driver was unlistenable to and was recycled, and its parts were used in several other driver prototypes I made.
1.1.2 A possible DSP to help with tuning
While working on several prototypes, I was still far from the sound I truly wanted to reach. At that point, I thought to myself... what if I would sell the headphones with a small DSP? While working on several prototypes, I experimented with making a DSP - successfully, but it didn't feel right to make a headphone good only thanks to a DSP - which still did not improve anything like imaging, soundstage, dynamics or detail - just the tuning. I decided to not go this route and keep the drivers as natural as they can be.
On the graph, you can see how I managed to use a DSP to make the tuning of a prototype driver better.
1.1.3 Last Single-Sided Driver Prototype
After many prototypes that were slowly improving after the first one, I got to the point, where I told myself, this is the last single-sided magnet array driver I'll make. Despite me liking its sound, I had a lot of people dislike it due to a part of the low-mid treble area being raised and 1.5kHz having a small dip. Compared to the first prototype, this one is miles ahead and it's visible I got some experience to make a usable driver. Unfortuneately, despite the fact I liked the sound, it still did not feel like a true high-end headphone because of the lack of soundstage and detail.
It was a 105mm driver with the impedance of 30 Ohms and sensitivity of 90.8dB/mW. The frequency range was 6Hz - 60kHz. It was good sounding to my ears, but I knew it was still far from the final driver that I would like to present myself with around the world.
After many prototypes, with a single-sided magnet array, it was time to move over to a double-sided magnet array for an improved performance - possibly the best decision I could've made while working on the headphones.
1.2 Double-Sided Magnet Array Prototypes
1.2.1 First Double-Sided Magnet Experiment
So why did I switch to a double-sided magnet array? I've been invested a lot into the making of the driver and I've been reading a lot of discussions about how the double-sided magnets improve bass performance, sensitivity and detail. All these were confirmed to me when right when I made the first driver with a double-sided magnet array. The tuning was not amazing, but it was considerably better. This was the first time when I thought to myself that this is a high-end driver. The bass performance was much better than with the single-sided array drivers and I was surprised how detailed the sound can truly be.
The 104mm driver measured at 50 Ohms impedance, sensitivity of 107dB/mW, frequency range of 8Hz to 51kHz. Mostly because of the tuning and soundstage performance, I was still sure that this is not the final driver and that I still have a fairly long way to go - just switching to a double-sided magnet array is not the end of the story.
1.2.2 What I thought might be the final driver
With this prototype driver, I almost kept it as the final one. I spent a lot of time with it to perfect it as much as possible, but something was simply off - I wasn't satisfied enough to call it the final driver - every time I listened to it, I was amazed, but still felt like this is just a prototype. The sound was not bad - both tuning wise and technicalities wise - the soundstage was wide, the imaging was precise, they were detailed, the dynamics did lack a bit, the bass was punchy and the treble was nice and sparkly, giving them a very open feeling. I doubt it was the dynamics that kept me thinking of it as a prototype.
This 114mm driver measured with a 48 Ohm impedance, 106dB/mW sensitivity, frequency range of 8Hz - 52kHz. The treble never felt like a problem, it was a brighter driver, but in a good way.
1.3 Final Driver
And we get to the final driver. The driver I am the most proud of, the driver that performs the best, the driver that I perfected. Offering a wide soundstage, precise imaging, good dynamics, a lot of detail and a tuning that offers flat bass all the way with basically no rolloff and an overall sound that follows the sound of real instruments instead of a target like Harman.
The driver is 114mm, measures the impedance of 42 Ohms, sensitivity of 108dB/mW, frequency range of 8Hz - 49.6kHz. It's something I am proud of and very happy that after so many prototypes, I managed to get all the way to this point.
Sheepskin Earpads Velour Earpads
After such a long time, having a driver that is as good as this feels truly amazing and I can finally say, I have a high-end planar magnetic headphone.....that I made myself.
2 Build
2.1 Cups
2.1.1 How it started
With drivers, you also need cups that they sit in. And this part was honestly just as complicated. At first I did not pay any attention and just went with a design that was simple to print, easy to model and didn't consume much material - mistake. The sound felt hollow, resonated and it made the well performing driver seem bad.
The first prototype cup +wood & connector +wooden grill
2.1.2 What could've been the final cup design
Later, that cup was upgraded with one, that looked the same from the outside, but had different densities, fills and was modified with acoustic foam around the walls inside. The upgraded cup was also painted silver - this was supposed to simulate a metal look, but in my opinion, that unfortuneately did not happen. These cups kept the sound feeling as if it was coming from speakers in a very small, but empty room. It was not optimal, but much better than the original prototype cups.
The upgraded cup The inside of the upgraded cup Both cups with grills and connectors
2.1.3 Finalised cups
The third and final cup design was made with some edits to the overall design. The inside space was now properly calculated for the drivers to perform well, the connector was changed from the triangle at the bottom to a "connector tube" and the driver now sat fully inside instead of on top of the edges.
2.2 Headband
2.2.1 A try at a fully custom headband
At first, I wanted to create a fully custom headband and I was trying for a long time using parts from Beyerdynamic, Audeze and using some of my own parts. I got very close, but unfortuneately I did not manage to finish a custom headband and eventually went on the journey to find the best, most comfortable DIY headband on the market. My custom headband went through a few changes - it started as just the DT770 bow and DT770 buttoned cover screwed into wooden blocks freesliding on Audeze's yokes.
After some time, this got updated due to the design being disliked. The buttoned cover got replaced by the Custom One velcro cover and the wooden blocks were replaced by 3D printed ones which were seperated into 3 different parts and were screwed in together.
Later, after moving onto newer cups, I went back to the wooden blocks, but kept the Custom One velcro cover on the DT770 bow. Arguably, this is the best looking design out of the bunch.
After this, I went back to the beginning and used the wooden blocks combined with the buttonable cover due to comfort issues with the Custom One cover. The design was worse, but the headband was much more comfortable.
2.2.2 Finding the best DIY headband on the market
After being unsuccessful with making a custom headband, I decided to move on and find the best DIY headband on the market. This included me getting and testing several headbands available to be purchased and testing their comfort after hours of use. After all of this, I ended on the headband, that is very Audeze-like looking, because its thin leather strap distributes the weight the best and the headphones feel like they're floating on the head without too much clamp force. This did also finalize the build.
2.3 Earpads
2.3.1 AliExpress general use earpads
Choosing the right earpads was a challange. There are many options and many possibilities of attaching them. The best AliExpress earpads I used with these were CP-YANG's general use sheepskin earpads. I modified them to attach onto the headphones with velcro - this made any kind of replacements and pad rolling easy and fast. These earpads kept the drivers fairly close to the ear resulting in a bit less treble, but very nice bass and mids. The earhole was big, but it was more in the front which did not quite allign with the driver's openings and resulted in the sound always slightly feeling like as if it was coming from the back.
2.3.2 Dekoni Audeze earpads
After using the CP-YANG earpads for a long time, I decided it's time to upgrade them. Since the cup size was identical to an Audeze, I decided I'll try the Dekoni Audeze pads. This also meant ditching the velcro system and taping the earpads onto the headphones, which makes padrolling a pain. I still think it was a good decision though. The bass became much punchier, they allowed the bass to extend all the way without rolling off, they increased the treble a bit, the soundstage became a bit wider. Dekoni's foam allowed to already comfortable headphone to become even comfier. Now it went from floating on the head to floating with pillows around your ears.
3 Final words
Overall, this project had a lot of ups and downs. I wanted to end it several times because I couldn't get good performance or stuff was not going well, but in the end, I'm glad I continued and finished them. They are currently the only headphones I know with circular 114mm drivers. I hope you all like the headphone despite only being able to look at fairly inaccurate frequency response measurements from the MiniDSP EARS and images of the headphones.
DMS listening to the headphone (prototype headband)
I'd appreciate any feedback! If you'd like to know more, feel free to comment - I'll try to respond to everyone or if you'd like to possibly even buy them for yourself, you can privately message me.
