1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

3D printed closed headphones with HD800 driver

2 3 4 5 6 7 8 9 10 11
  1. frank2908
    As a long time head-fi follower of modding headphones, like many people here, I've tried 2 of the most popular headphones mod: the grado cup mod (in aluminium) and fostex T50RP (aluminium cup mod with Audeze/Kennerton earpads)

    Now it's time to try something new, to my knowledge, no one has tried customizing the HD800 before, so I've got to try it.
    To start with, before starting the project, I've only auditioned the HD800 for 5 minutes, but based on the reviews the treble can be quite hot. My inspiration for this build is the sony R10, which some guy mentioned that has deeper soundstage than the hd800. So I knew a closed (or semi closed) headphones with great soundstage is possible.
    Here is a sneak peek of the final (until now) prototype:

    I bought a pair of replacement driver online and started to design the cups and baffle. The headband was taken from a superlus hd660. And eagerly I asked fellow head-fi member Rudy- Calico88 to CNC the cups from rosewood. Seeing his work on pictures, I had no doubt it would be good.

    First prototype:
    I know it is silly to make the first prototype of your headphones in very expensive material. The build quality of the cup from Rudy-from the tolerance, to the quality of the wood and the finish is top notch. I designed the baffle with lots of hole to tune the sound by covering them with tape.

    The cup was then dampened with dynamat. Note that the in the middle of the cup there is a removable disc for open/close conversion.

    First listen: It really showed me the potential of the driver. The hd800 is well built, it weight 35g compare to the plannar driver of fostex at 45g.
    The treble detail is there, but with tuning, either I get a good soundstage but no bass or vice versa. Also the soundstage width is very good but not the depth. I spend 2 months playing with this and came to a conclusion: I made a mistake in the design: the cup's volume was too big, and even though the audeze pad is angled, the driver mounted flat on the baffle does not create a deep soundstage. Then came the difficult decision: I have to abandon the beautiful rosewood cup. At the same time, my Printrbot SImple Metal 3D printer arrived, I knew I have to change the design.

    Second prototype:
    Still in love with the wood cup, I decided only to ditch the rosewood baffle for a new design, this time incorporated an angled driver relative to the baffle. The baffle protrude deep into the cup, therefore reduce the volume inside the cup. This new baffle is 3d printed from ABS. The baffle is left undamped ( not with dynamat or blu-tack)

    Sound impression: The soundstage and bass is good, but there is a weird mid. I suspect there is a reflection on the ear side of the driver, due to a large volume created by this new baffle.
    Now I have to decide to remove the driver from this beautiful cup to a fully 3d printed cup. On the new design, I also want to make the cup less deep. I made a mistake designing 2 huge drums stuck on my head.

    Final prototype:
    WIth the inspiration come from the SOny R10 and the speaker B&W Nautilus, I want to design the cup without the concentric shape and inspired by a sea creature. Behold the sea shell cup.


    The 3d printed cup was in grey, then bottle sprayed black.The design of the baffle also changed( it's the grey disc in the next picture). With new design, there is no reflection of sound on the ear side of driver.
    I later spend 3 months listening and tweaking the sound, by opening/closing venting holes on the baffle and cup, damping the cup with blu-tack ( i will switch to dynamat later). After the dynamat, there is a layer of lambskin leather and cotton wool to further damp the driver.
    THe jack is mini XLR from Rean.
    Sound impression: (briefly compared to a hd800 in a showroom)
    Now I'm totally satisfied with the sound. By mass damping the baffle and and cup, the sound is not bright. To my ear, the quality of the bass and mid is mostly the same as the stock hd800. However I am most proud about the soundstage. Although is closed ( or semi closed due to vents), the soundstage is not as wide but deeper than the hd800, as a result of extremely angled driver (angled ear pads + angled driver from baffle).
    Now I know this is not the final prototype, because tuning headphones make you listen to them for a long time. As a result your ear becomes used to the sound. Therefore sometimes I have to go to the shops to compare to other headphones, and it takes time.
    THe final pic is the headphones with different ditched component. Now I'm not throwing the rosewood cups away, still waiting for the next project.

    Edit: new pictures of the build:
    I made the cable from Double Helix Cable Nucleotide, Switchcraft 3.5mm jack (so I can use with my dx90 conveniently) and rean tiny XLR. For mini XLR, I think Rean is a good choice for it's aesthetic and affordability. The terminals are easy to solder. I use a plastic sleeving for speaker cable, similar to Techflex, but more coarse, you can see the braiding through the sleeves.
    A clear shot oy the yokes assembly. The stock yoke from the Superlux HD660 is too small (the same size as Beyerdynamics), so I made a new yoke from stainless steel.I cut the shape from 2mm steel plate, then grinded and polished by dremel.
    The hinge of the yoke (small rectangular plastic thing) is 3d printed from carbon fiber abs, glued on to the yoke. Then I use M3 socket head screw to attach the yoke to the cup.
    There are 3 holes near the mini xlr connector. I designed the cups with vents at the front, but I need to drill extra 3 holes for tunning.
    I used similar method to the Denon Markl mod, adding foam to the thicker side of the pads and cotton to the thinner side. It just sound good to my ears in this config. Also note the 2 holes I punched out of the earpads, helps thinning the mid.
    The naked shot: the baffle is 3d printed from 3 parts, saving me from printing in 1 go and spend days removing the supporting materials. THe white blutack is there to mass damp the baffle and reduce reflection from the driver. There are small vents connecting to the main cup chamber. this reduces the pressure and again helping the soundstage, but too many open vents and the sound becomes thin with peaky treble.
    There is also a lip for mounting the Audeze earpads.
    Final pics showing the beautifull hd800 driver and my crappy craftsmanship with lots of blu-tack

    ---13th Oct 2016 update------
    Version 3 update:
    Finally got a long holiday and had enough time to finish the new prototype.
    Unfortunately I could not make these out of carbon fiber, since I don't have enough time. Also last time I played with carbon fiber, turned out the epoxy I have hardened too fast ( 10 mins), therefore i can't do resin infusion or even normal lay up because of lack of time for air bubble to disappear.
    The earpads are mounted on plate with magnet attached, similarly there are magnets on baffle,
    Every parts are made from Carbon fiber High PLA filament from FIlabot
    weight: 400g without cable ( the last prototype weight 500g). Without dynamat, only weight 330g
    Old design, driver at a small angle, and remain in the center of the ear pad opening.

    New prototype driver mod
    1. resonator made from 3d printed Brass PLA from colorfab, I can get the exact dimension , also the brass plastic is quite heavy, so it will damp the plastic protector. I had to take the driver apart and cut the screen off the center hole
    2. [​IMG]
    3. the driver is at 35 degree from the baffle plate and push as much forward as possible
    4. the baffle is lighter and has lots of opening. They are then covered with Micropore tape (white)
    Another view of the extreme angled driver.

    The backside of the baffle and the cups is covered in Dynamat, then one layer sound absorbing foam similar to Akasa Paxmate. The holes/ openings are covered with micropore tape.
    Mini XLR connectors.





    The Carbon FIber PLA is really stiff, strong and look really nice. It has a matt finish and looks really cool. On the left is carbon fiber PLA, right is normal PLA



    3D printed HD800 in Carbon fiber PLA
    I was too lazy to make a new headband assembly from scratch, so I used an old Fostex T50rp headband.

    The connector will be covered in Black Sugru. I don't want to make a chunky and heavy 3d printed cover or housing for those connector. So it is kept simple.
    Sound improvement:
    Because the driver is much more forward and angled, the image of this can really improved. It is now very pin point. But there is still issues with lacking bass.
    ---14th Aug 2018 update------
    My latest version, V4
    I completed this last year actually but never posted since I traveled round the world for 1 year, so left this baby at home and just picked them up last month. I have a new design but my 3d printer broke and I'm on the road again,anw these sounds great with Chord Mojo
    It's the first time I switch to open ear design, and there is a removable magnetic cap to convert to closed (not shown in photo). There is a dust screen cut off from stax sr007 earpads fabric. The driver cover is printed with honeycomb pattern
    On the baffle and earpads rings there are magnets so I cant swap pads easily. Now I have the diy Stax pads (Sr007 with memory foam fill ) and Shure Alcantara earpads. The Stax produces a better soundstage and a bit more comfy because the memory foam I used is very soft, but if I'm not in a room with AC, then I switch to Alcantara pads. The alcantara has a bit smaller soundstage but better bass (tiny bit)
    The baffle is left open and sealed with 3M Micropore tape, which has the best balance between sealing (more base) and open (soundstage) IMO.
    The original idea of making the headphones with huge closed cup like R10 and HD800 driver failed, not because of the cup but because the volume between the driver and the ear was not big enough, and the hd800 are designed for this huge volume. Now I plan for next build to be convertible between open/closed and over ear/ on ear. When you have an expensive headphones and a Chord MOjo, why do you have to buy another portable headphones? I believe to make a convertible portable hd800 would sound at least better than any $500 portable headphones anyway
    Last edited: Aug 15, 2018
    Richter Di, dogears, haolevu and 21 others like this.
  2. Oregonian
    Congrats!   Amazing work and I'm going to follow this thread for the next iteration(s). 
    Thanks for sharing. 
  3. frank2908
    Thanks, for now the next modification I can think of is to make the mini XLR socket on the headphones smaller, it looks a bit out of proportion in my opinion. I'm considering 2.5mm mono plug- a la HD700. But so far I can't find a good but cheap 2.5mm connector with strain relief, I afraid the thin connector would break on this headphones. Also the Audeze earpads look ridiculously thick, I tried the Kennerton earpads which looks good, but the leather is too hard can does not seal well, therefore blured the mid.
    I have cut out my own angled earpads from very soft memory foam, and have a piece of lambskin laying there, but I'm hopeless at sewing. Hopefully one day I'll learn how to make leather earpads.
  4. kapanak
    Amazing work. Quite inspiring I must say :) And you're right, nobody I know of has attempted a headphones from scratch using HD800 drivers before, much less a closed headphones.
    I actually going to order a pair of replacement HD800 drivers online for something similar to what you're doing. But then again, $500 for a pair of drivers is a VERY high cost of entry for DIY headphones lol ... I happen to have access to fully featured physics laboratories in a university, so I can do some acoustic analysis and imaging as well to optimize a cup :D Just as you say, a great closed headphones is possible, and just not done. Not enough research. DIY it is :D
  5. frank2908
    The Sony R10 is an evidence that it exist, but at that price for most people, it is nothing but a myth :frowning2:(
    Also if you order the driver, be carefull not to solder the cable to the driver for long. I made a mistake melting the solder post into the plastic and broke the extremely thin wire connecting to the voice coil. I was scared as hell but then managed to fix it by soldering the thin wire back together.
  6. kapanak

    Question: Did you order directly from Sennheiser?
  7. rfan8312
    Great work! Inspirational. Congratulations.
  8. chinsettawong
  9. Mahdi8
    Oh wow. A closed hd800 is the dream for me
  10. koiloco
    ^ Dang.  Very nice work.  If you started back in December of last year, I could have paid you a visit in person.  [​IMG]
  11. nw130d
    Great stuff, true hobbyist!
    Personally always wanted to 3D print closed cups for the HD700. But the 3D printer and measurement tools adds up astronomically!
    Side note some companies like Inventables are making small CNC machines that runs a Dremel so one in theory can produce any wooden cups they want.
  12. frank2908
    To put to perspective, my first prototype in rosewood cost more than the 3d printer. The cost for a desktop cnc is comparable to a printer, but exotic wood cost significantly more than 3d printed plastic. Plus since it is additive manufacturing, you need less material than traditional machining.
    Diy wise both requires lots of experience, cutting with wood might result in wood burn, you have to play with the speed and feed rate. For 3d printing, you have to play with many setting for different material and you have to damp afterward. Ideally wood is acoustically better , therefore I'm interested to learn methods for creating liquid wood like the audioquest nighthawk. I can make the mould out of the 3d printer.i also have carbon fiber laying around but given the resonance, I have to add mass damping material anyway. It will not be light, therefore the only purpose for carbon fiber is aesthetic.
    Anyone know how to mold wood/plastic or epoxy composite?
  13. frank2908
    You mean in December 2013? At that time only the driver has arrived. :) you are welcome the next time you are in Vietnam
  14. koiloco
    No, December 2014.  It could have been fun. I will probably visit next summer.  If I do, I will PM you.  [​IMG]
2 3 4 5 6 7 8 9 10 11

Share This Page