Yet Another DIY Isodynamic (Planar Magnetic) Headphones
Sep 16, 2022 at 5:21 AM Thread Starter Post #1 of 9

Ulfar4

Head-Fier
Joined
Mar 29, 2018
Posts
86
Likes
217
Location
Czech Republic
Hi All,
I have been working on this headphones for some time and I want to share some results.

I want to clear from start that I won't give away everything that I learned, mainly not revealing the process of the membrane creating. Although it can be guessed from photos. It just cost me too much money and time to get usable results. So for now I want to keep it but I don't rule out that I won't reveal it later.

So why I choose to try build this type of headphones? I already built an electrostatic ones and I was always curious about planar magnetic drivers after I saw the thread about Verum One headphones and many others.

It was good to set some goals which I wanted to achieve.
- Verify that I can build membrane with usable range of impedance. Without this step, it would have no sense to continue with the project.
- Get at least some sound from the driver as I did not expect any high quality sound.
- Then maybe tune design to get some usable sound that will match planars in lowest price range.
- In the end final goal was to have fun during building those headphones and get some new knowledges about headphones.

1) First membrane
1.jpg2.jpg3.jpg
This was actually the biggest challenge to create first usable membrane. It took me almost three months to do this since it was time consuming and one have also other duties than only to diy something. So when I achieved first good result, I bought then some magnets, printed some housing and used some old headband. I was actually surprised that it worked but the sound was not usable at all. Those headphones was bass heavy. I did some measurement on my DIY headphone rig, which was in no way accurate but it helps a lot.

4.jpg5.jpg

2) Improving the design to get more balanced sound
6.jpg7.jpg
Now it was all about to do better design since the first was only for verification. I did membrane with new shape and new housing. Then it started to look so much better. And indeed the sound was much better after sime tinkering with membrane tension. It was definitely the signal to continue with this project more.

After some othe iterations and experimenting with damping I was really satisfied with the results. Sound was good with lot of bass extension. It had dark signature, so it was totally opposite of my electrostatic headphones.

8.jpg9.jpg10.jpg11.jpg12.jpg13.jpg
Since I was satisfied with headphones, I decided to manufacture better mechanical package. Membrane was glued to FR4 spacer and housing was created in way that it was comfortable to wear it. Impedance was 11Ω, weight without cable 470g and magnetic structure single sided. It was comfortable even for long session despite higher weight.

To this moment I was using my DIY measurement rig for tunning the sound but since it was impossible to rely on results I ordered MiniDSP Ears. I know it is not as precise as some other rigs, but it is affordable and consistent at least in some ways.

3) Measuring with MiniDSP Ears rig
I think it will be good if I shortly explain the compensation curve I am using for Ears so there will not be any misunderstanding.
Ears offer three types of compensations, RAW which is essenitaly only level calibration of mics. Then HEQ (Headphone EQ Compensation), this will point you to equalizing the headphones for neutral sound. Last is HPN (Headphone Measurement Compensation), this will give you results that is most simmilar from other measurements rigs. I did not like either of them, so I did my own compensation, which combine HEQ and HPN, it is distantly simmilar with SBAF compenstaions.

compensation.png
I named it MPN and to around 1kHz it is using HPN curve. After this frequency it is using HEQ curve. This is what suits me best when tunning and measuring headphones. So any further measurements will have this compensation curve. It is not accurate or best way how to measure the headphones, but I just can't afford measurement rig that costs ten times more than EARS since this is mainly for fun project. Measured headphones above are some Beyerdynamics for clarification.

mgn silver l-r.png14.jpg
Finally there is measurement from the silver(gray) headphones. As you can see, it is still relatively dark signature headphones.

4) Tweaking current design
I ordered multiple earpads and damping materials and I try to improve the sound even more. I also did minor redesign of the housing (pictures are with sheepskin earpads)
15.jpg16.jpg17.jpg
I actualy ended up with using velur earpads and two layer damping under protection grill. The sound was more even while still maintaining dark signature. I always imagined that LCD2 from Audeze have simmilar sound to this. So naturally I had to borrow ones. I have luck that we have one audio shop that is offering this service so I got LCD2 for 5 days to listening.

18.jpg
The sound was actually similar which pleased me. My headphones were slightly harder to drive and much more comfortable than Audeze, but I must say this new revision with thinner earpads is not actually bad. Audeze has better detail in sound, for some tracks I prefered Audeze and for other my headphones. I am actually not good in describing the sound so I can not comment it more. And also there is one other thing. Maybe the feel of that this my creation is clouding my judgement of how they sound.

19.jpg
Anyway, there is comparison of LCD 2 and my headphones from EARS.

20.jpg
There is one more comparsion with LCD2 and my headphones with velur and sheepskin earpads.

5) What next
From comparison with LCD2 I have some new ideas. I will try design with double sided magnetic structure to see if there will be any improvements. The headphones should be then easier to drive and have reduced distortion due to symmetrical design. I also hope that level of detail will be close to LCD2.

MGN-2 Assy crbn.JPGMGN-2 Assy 2.JPGMGN-2 Assy 3.JPG
I did some first drafts of new version. It will be double sided but the cups will be smaller than before. First design with single sided structure had 110mm diameter cup. New one has 103mm diameter.

New parts are on the way so when I have updates I'll add them to next posts to keep you updated.

I hope that you enjoyed reading this post.
 
Last edited:
Oct 3, 2022 at 11:13 AM Post #3 of 9
Thanks. Yeah, isodynamic is only naming convention, it is classic planar magnetic driver. Isodynamic is this type with bar magnets and orthodynamic is the type wit circle magnets like Final Audio D8000.

5) Progress on the second version

PXL_20221008_074820462.jpg
I received all necessary parts for new design so I started designing new membrane because of different magnets. The result is a membrane with an impedance of around 9Ω (left). But I tried to push it more so I did thinner traces and I was able to get impedance of 34Ω (right).

PXL_20221008_145728944.jpg
Now I am slowly building the whole assembly whenever I have time for it.

Assembly is almost complete, I am really curious how it will sound.

308799938_832046084462160_4254405148844442063_n.jpg

During the last two weeks I completed the headphones and I was quite satisfied with the result. So I reprinted the headphones and did some surface finish. On first listen I immediately thought that it sounded a lot like the Audeze LCD2 with the new thinner earpads. So I did new measurement and compare it and it is very close to LCD2.

PXL_20221030_112039864.jpgPXL_20221030_112107158.jpgPXL_20221030_112144010.jpgPXL_20221030_112214473.jpgPXL_20221030_112245235.jpg

There is comparsion with LCD2.
300226562_423619399845914_4459287470949705684_n.png

Headphones are still missing some damping so I will play with this more, also I will try other earpads. Also I maybe try to connect some RC filter like Audeze LCD2 Closed have.

Finish is really near the corner. I am playing with damping and it seems that I found the combination that I like most. It is the green line in frequency response.

Red - Perforated sheepskin earpads without any damping
Green - Perforated sheepskin earpads without damping filter
Red - Solid sheepskin earpads with damping filter

Porovnani filtr.png

Headphones with installed damping filter.
315227313_850326955999887_8743152936745899681_n.jpg

Now I need to let the headphones play for some time to see if they are reliable.
 
Nov 12, 2022 at 10:44 AM Post #5 of 9
It is open back design. There is comparison with DT770.
dt770 comparsion.png

After all, it was a bit too much with that filter. So I was still working and trying to lower the treble because it was very tiring for longer listening.

MGN-2 mod.png

Small update after some time. Headphones was sent to few people to get some feedback. You can read some of it in local forum there (first page, posts nr. 4,5,10). It is written in Czech, but you can use a translator, hopefully it will be at least somewhat usable.
https://forum-avmania-zive-cz.trans...1328551&_x_tr_sl=auto&_x_tr_tl=cs&_x_tr_hl=cs

After the headphone tour I noticed that the paint was worn out in some spots on the headphone so there will be need of better way to paint them.

av14.jpg

The biggest problem of this headphones is probably the headband, it is too large for almost everyone and the quality is also not very good. It is already starting to fall apart in one spot.

av13.jpg

So it was time to decide that I have to create my own headband, which will suit the headphones much more. I spent some time inspiring from various designs and then create one. It took few prototypes before it was usable and I was satisfied.

av12.jpg

Now it is in state of raw 3D print and have not yet decided what to do next but I can say that the new headband is much more comfortable. Weight of the headphone is now 390 grams.

So probably the last update for MGN 2,5. The biggest change since the last update is the surface treatment. It is acetone smoothed.
1aj.jpg
1bj.jpg
1cj.jpg
 
Apr 11, 2023 at 6:22 AM Post #6 of 9
"After the headphone tour I noticed that the paint was worn out in some spots on the headphone so there will be need of better way to paint them."

I have had good results with lechler megamack to protect paintwork. So some kind of automotive clearcoat may give themsome extra protection
 
Feb 20, 2024 at 12:42 PM Post #8 of 9
It is open back design. There is comparison with DT770.
dt770 comparsion.png

After all, it was a bit too much with that filter. So I was still working and trying to lower the treble because it was very tiring for longer listening.

MGN-2 mod.png

Small update after some time. Headphones was sent to few people to get some feedback. You can read some of it in local forum there (first page, posts nr. 4,5,10). It is written in Czech, but you can use a translator, hopefully it will be at least somewhat usable.
https://forum-avmania-zive-cz.trans...1328551&_x_tr_sl=auto&_x_tr_tl=cs&_x_tr_hl=cs

After the headphone tour I noticed that the paint was worn out in some spots on the headphone so there will be need of better way to paint them.

av14.jpg

The biggest problem of this headphones is probably the headband, it is too large for almost everyone and the quality is also not very good. It is already starting to fall apart in one spot.

av13.jpg

So it was time to decide that I have to create my own headband, which will suit the headphones much more. I spent some time inspiring from various designs and then create one. It took few prototypes before it was usable and I was satisfied.

av12.jpg

Now it is in state of raw 3D print and have not yet decided what to do next but I can say that the new headband is much more comfortable. Weight of the headphone is now 390 grams.

So probably the last update for MGN 2,5. The biggest change since the last update is the surface treatment. It is acetone smoothed.
1aj.jpg
1bj.jpg
1cj.jpg
Wow! This is really incredible, thank you for sharing. Love the final design 😍
 
Mar 3, 2024 at 12:07 AM Post #9 of 9
Hi All,
I have been working on this headphones for some time and I want to share some results.

I want to clear from start that I won't give away everything that I learned, mainly not revealing the process of the membrane creating. Although it can be guessed from photos. It just cost me too much money and time to get usable results. So for now I want to keep it but I don't rule out that I won't reveal it later.

So why I choose to try build this type of headphones? I already built an electrostatic ones and I was always curious about planar magnetic drivers after I saw the thread about Verum One headphones and many others.

It was good to set some goals which I wanted to achieve.
- Verify that I can build membrane with usable range of impedance. Without this step, it would have no sense to continue with the project.
- Get at least some sound from the driver as I did not expect any high quality sound.
- Then maybe tune design to get some usable sound that will match planars in lowest price range.
- In the end final goal was to have fun during building those headphones and get some new knowledges about headphones.

1) First membrane

This was actually the biggest challenge to create first usable membrane. It took me almost three months to do this since it was time consuming and one have also other duties than only to diy something. So when I achieved first good result, I bought then some magnets, printed some housing and used some old headband. I was actually surprised that it worked but the sound was not usable at all. Those headphones was bass heavy. I did some measurement on my DIY headphone rig, which was in no way accurate but it helps a lot.



2) Improving the design to get more balanced sound

Now it was all about to do better design since the first was only for verification. I did membrane with new shape and new housing. Then it started to look so much better. And indeed the sound was much better after sime tinkering with membrane tension. It was definitely the signal to continue with this project more.

After some othe iterations and experimenting with damping I was really satisfied with the results. Sound was good with lot of bass extension. It had dark signature, so it was totally opposite of my electrostatic headphones.


Since I was satisfied with headphones, I decided to manufacture better mechanical package. Membrane was glued to FR4 spacer and housing was created in way that it was comfortable to wear it. Impedance was 11Ω, weight without cable 470g and magnetic structure single sided. It was comfortable even for long session despite higher weight.

To this moment I was using my DIY measurement rig for tunning the sound but since it was impossible to rely on results I ordered MiniDSP Ears. I know it is not as precise as some other rigs, but it is affordable and consistent at least in some ways.

3) Measuring with MiniDSP Ears rig
Mình nghĩ sẽ tốt hơn nếu mình giải thích ngắn gọn về đường cong bù mình đang sử dụng cho Ears để không có sự hiểu lầm nào.
Tai cung cấp ba loại bù, RAW, về cơ bản chỉ là mức hiệu chuẩn của micrô. Sau đó là HEQ (Bù EQ tai nghe), điều này sẽ hướng dẫn bạn cách cân bằng tai nghe để có âm thanh trung tính. Cuối cùng là HPN (Bù đo tai nghe), điều này sẽ cho bạn kết quả gần giống nhất với các thiết bị đo khác. Tôi không thích một trong hai điều đó, vì vậy tôi đã tự thực hiện khoản bồi thường của riêng mình, kết hợp HEQ và HPN, nó rất giống với các khoản bồi thường SBAF.


Tôi đặt tên nó là MPN và ở khoảng 1kHz, nó đang sử dụng đường cong HPN. Sau tần số này nó đang sử dụng đường cong HEQ. Đây là điều phù hợp với tôi nhất khi điều chỉnh và đo tai nghe. Vì vậy, bất kỳ phép đo nào nữa sẽ có đường cong bù này. Đây không phải là cách chính xác hoặc tốt nhất để đo tai nghe, nhưng tôi không đủ tiền mua thiết bị đo đắt gấp 10 lần EARS vì đây chủ yếu là dự án giải trí. Tai nghe được đo ở trên là một số Beyerdynamics để làm rõ.


Cuối cùng là phép đo từ tai nghe màu bạc (xám). Như bạn có thể thấy, đây vẫn là chiếc tai nghe có đặc điểm tương đối tối.

4) Tinh chỉnh thiết kế hiện tại
Tôi đã đặt hàng nhiều miếng đệm tai và vật liệu giảm chấn và cố gắng cải thiện âm thanh hơn nữa. Tôi cũng đã thiết kế lại một chút phần vỏ (hình ảnh có miếng đệm tai nghe bằng da cừu)

Trên thực tế, tôi đã quyết định sử dụng miếng đệm tai bằng nhung và lớp giảm chấn hai lớp dưới tấm lưới bảo vệ. Âm thanh đều hơn trong khi vẫn duy trì dấu hiệu tối. Tôi luôn tưởng tượng rằng LCD2 của Audeze có âm thanh tương tự như thế này. Vì vậy, đương nhiên tôi phải mượn những cái đó. Tôi thật may mắn khi chúng tôi có một cửa hàng âm thanh cung cấp dịch vụ này nên tôi đã có LCD2 để nghe trong 5 ngày.


Âm thanh thực sự giống nhau làm tôi hài lòng. Tai nghe của tôi khó lái hơn một chút và thoải mái hơn nhiều so với Audeze, nhưng tôi phải nói rằng phiên bản mới này với miếng đệm tai mỏng hơn thực sự không tệ. Audeze có âm thanh chi tiết hơn, đối với một số bản nhạc, tôi thích Audeze hơn và đối với các tai nghe khác của tôi. Thực ra mình không giỏi miêu tả âm thanh nên không thể bình luận gì thêm. Và ngoài ra còn có một điều khác. Có lẽ cảm giác rằng sự sáng tạo của tôi đang che mờ khả năng phán đoán của tôi về cách chúng phát ra âm thanh.


Dù sao thì cũng có sự so sánh giữa LCD 2 và tai nghe EARS của tôi.


Còn một sự so sánh nữa với LCD2 và tai nghe của tôi có miếng đệm tai bằng nhung và da cừu.

5) Điều gì tiếp theo
So sánh với LCD2 tôi có một số ý tưởng mới. Tôi sẽ thử thiết kế với cấu trúc từ tính hai mặt để xem có cải tiến gì không. Khi đó, tai nghe sẽ dễ lái hơn và giảm độ méo tiếng do thiết kế đối xứng. Tôi cũng hy vọng rằng mức độ chi tiết sẽ gần bằng LCD2.


Tôi đã thực hiện một số bản phác thảo đầu tiên của phiên bản mới. Nó sẽ có hai mặt nhưng cốc sẽ nhỏ hơn trước. Thiết kế đầu tiên với cấu trúc một mặt có đường kính cốc 110mm. Cái mới có đường kính 103mm.

Các phần mới đang được triển khai nên khi có bản cập nhật, tôi sẽ thêm chúng vào các bài đăng tiếp theo để cập nhật cho bạn.

Tôi hy vọng bạn thích đọc bài viết này.
Hello. What kind of glue did you use to bond the membrane and the conductive layer?
 

Users who are viewing this thread

Back
Top