TH900 mod extravaganza (stream of consciousness)

Very useful and interesting thread. What combination of material from your point of view increased virtual volume at max level?

 

I'm not sure I understand your question @Garuspik . I didn't notice a big difference in terms of efficiency with the mods if that's what you mean. Maybe find myself wanting to turn them up more since the treble's less hyped, but that's about it.

 

Pardon my english. I mean what combo of dampening materials made Fostex to sound more open then stock?

 

I'd start with removing the paper damping on the back of the driver and finding a synergistic pair of pads. I actually ended up removing all of the paper on the X00s yesterday (on the driver, not the baffle of course!) as opposed to what the picture shows on the 900s - helped narrow the dynamic gap a little bit.

Removing the paper is a bit scary since it's irreversible and you risk puncturing the driver while trying to tear it off, but you can always use micropore tape to restore it to stock (more or less) if you decide you don't want it fully open. Less paper covering = improved macro dynamics, reduced treble peakiness, at the expense of deepening the 3k dip (which I don't mind).

The cup stuffing and lining that I've done definitely yielded improvements compared to stock fluff (or empty) and most simple fillings, but if you just fill the cups with some moderately packed cotton or acoustic fiber (more stuffing on the sides, less stuffing in the center) that'll probably take you 90% of the way there once you find good pads and remove some or all of the rear driver paper.

These headphones like firm and shallow pads, normally with a fairly small and circular ear hole. Bigger ear hole means less net slope but also less linear treble if you're going without front damping (which is recommended if possible). Pretty much all of the deep pads I tried gave jagged treble, which was a bummer because I have lots of ZMF pads I thought would give good synergy but didn't to my ears. The AP2000Ti pads were kind of the exception in terms of not having a circular earhole.

 

@E_Schaaf Thanks. Now i'm repeating your way of damping with simillar materials. Very good results.

 

Something I forgot to mention -

The density of the front foam surrounding the driver basically will adjust everything below 500hz in a very shelf-like manner. The stock foam density on TH900 is 35:1 (mine has 45:1), but it appears the stock foam density of the X00 is 25:1. So when I don't swap the foams and use the X00 stock foam, perform all the mods as pictured (except with zero rear driver paper), and with AP2kti pads, you get a really cool graph -



Now that's what I call flat!... well at least up to the 3k area. Pretty cool regardless if you ask me, and an interesting signature nonetheless. Yes, you could do a dead-flat TH900 this way if you wanted to as well, save for the dip. I couldn't cut the denser than 45:1 neoprene to shape to play with (might do some laser cutting or something else), so now I'm either stuck with 25:1 or 45:1.

 

One last tweak (seriously I'm done... at least until those 8 other pad variants come in) -

Removed hairscreen of AP2kTi pads, placed diagonal RAD-sourced acoustic paper covering about 50% of pad opening.

Helped bring 3.5k bounceback in-line with 2.4k dropoff point ( as opposed to a rise ) Lower treble balance feels better, less sharp, less etch and hash (though it was barely there before - my ears are just sensitive), better integrated with the mids. Bumpy bits around ~8k likely just positioning differences. More volume and less fatigue without any sense of choked dynamics (even 1 layer TP can make them feel a bit veiled). Nice! 10Hz? No prob.



bUt I tHoUgHt OnLy PlAnArS cOuLd Do BaSs!1!1!

without these tweaks for refernece:

 

So @E_Schaaf just came by my house (with masks and social distancing ofc, we’re not barbarians) to let me listen to his modded headphones for a brief listen and here are my impressions on them! Thanks again for letting me try them, it was a lot of fun!

Listening Rig: Pi2AES> Sonic Frontiers SFD2 > BHA1 > JAR650 | Evan modded TH900, HD600 | Elegia

Evan’s modded TH900

• I've always wanted to hear the crazy slam of TH900s and today I finally got the chance to try them. Let's just say the legend lives up to its hype. Biocellulose drivers FTW! This was one of the coolest, unique headphones I've heard. It's definitely not reference tuning, but it was a very immersive and fatigue free listening experience. Because my ears were so used to the Sennheiser 6x0 sound, it took me a bit for my brain to adjust to its downslope FR.
• Subbass is very extended and feels pronounced in comparison to his modded 600 and my jar650. This thing punches and kicks with SFD2 and BHA1 lol I can turn this up quite a bit without nasty peaks murdering my ears and just let the cans blast me off to another dimension.
• Adding on from this, yeah this thing has insane macrodynamics, possibly one of the best I’ve heard from a dynamic can. Bass is certainly pronounced, but not tasfeully done without adding a significant mid bass bump. It's clean and punchy. Borealis comes close, but Evan’s fostex is a good level or two more slammy. I don't even want to mention the Elegia with those two because it's just embarrassing.
• I was not sure if I would like the downsloping FR but it’s actually very pleasant to listen to. Treble is not as smooth has his modded 600 but still really well behaved, like how it shows in his measurements.
• In some tracks, vocals sound slightly withdrawn/congested in comparison to his modded 600, probably because of the upper mid dip, brain burn in, closedback vs. openback, etc. I thought it would sound dark, but there’s actually plenty of treble sparkle.
• Staging isn’t the biggest, but understandable since closedback and it’s not Evan’s utmost priority in his mods. However, it feels very immersive and has great image separation.
• Another very notable aspect is how much i can turn this up without getting any listening fatigue. Blade runner 2049 soundtrack was very immersive to listen to without the bass overpowering everything else.
• no issue with timbre and resolve. Can’t speak much since I only had an hour to listen to them.
• I wish I had more time with this, but nevertheless, very impressive how he turned such cursed tuning of the stock TH900 into something so enjoyable.

Evan’s modded HD600

• I've heard various different mods on 6x0 variants, so I didn't expect to like this one so much. My notes are short compared to the TH900 due to limited time to audition and since I'm familiar with the general Senn tuning.
• Compared to his TH900, this one is more reference tuning that most people will prefer. Sounds pretty linear other than the subbass roll off, but who tf cares, no one expects it to have that linear bass extension from these drivers since they're not one of its strengths. It's slightly more livelier and forward, aggressive sounding than my JAR650
• Excellent clarity and black background reminiscent of JAR600.
• Overall very smooth throughout the FR, but energetic listen. I don’t think I hear any bothersome treble peaks or dips. mids reminded me of HD650 because of how liquid it sounded while maintaining HD600's treble presence.
  
• Notable aspect from this mod was that the upper mids are pretty much tamed, even better than JAR600. This is really surprising to hear. Even when I played my peaky upper mid heavy korean ballad tracks, I heard no high pitch ringing or fatigue that I usually get with stock HD600.
• Another interesting observation was how he was able to keep the lively sound with a set of pretty worn out stock pads. Did Evan basically solve the pad wearing out issue? -gasp-
• I was so impressed by this mod, I plan on replicating it in the near future. It's not enough to replace my JAR650, but I think his mods for the HD600 deserve more attention and praise.

 

Just wanted to shout-out @tommytakis for taking the plunge after taking a listen and commissioning this mod for a pair of TH600s I'll load with 900 drivers and a small tweak to exclude the external mods. Thanks to his contribution, I'll be able to order some extra drivers and cups, maybe even an XLR cable, so the loaner pair has a green light! Bit of lead time on parts for the loaner, but expect it to ship out from me in 3-4 weeks. I'll ask @ChaChaRealSmooth to start a loaner signup thread once everything arrives here.

 

This mod took quite a bit of doing and undoing after starting with a metal-cup TH600 (I had a short term metal cup romance until I put on some vocals HONK HONK), but here's a one-of-a-kind LTD Monkeypod @tommytakis edition SuperBest900.

These have deeper cups than the stock TH900 or TH600 (helps with lower mid response and general spaciousness), transplanted TH900 drivers, and sheepskin 'flat pads' for a less yok'd bottom end and smoother highs with no front damping. There's some new secret sauce on the inside of the cups unlike my own pair - these have a simpler damping scheme that's both more breathable directly behind the magnet assembly and also more controlled overall. Was able to open up 10 out of 12 driver ports as opposed to only 7 on mine with the improved damping scheme for even more kick in the face.





I'll be dropping these off with him this weekend! Already have all the parts needed for the upcoming loaner pair ready to go aside from the drivers (which are supposedly backordered for 6-8 weeks). That being said, if anyone wants some mods done and can find TH900 drivers some other way, I'll be happy to work some mojo - just hit me up!

Thanks again Tommy - if you decide to flip these at a later time, I'll be first in line to buy them back from you. Cheers!

 

Loaner pair teaser. Finally came up with an experimental mod that 'fixes' the upper mid dip quite a bit.





Still working on some custom front baffle foam rings and a potential pad swap that'll boost the sub presence below 50hz. But the response here is about as even as I can get it. Pretty jazzed!

 

Looks yummy, good job.

 

These front baffle foam rings make a huge difference in the lower third of the response. The stock rings (left) have 4 small vents to allow some airflow for a more 'semi-open' design. Depending on where those vents are facing relative to the stock baffle port, you can dampen or open everything below about 500hz. Firmer foam also means more presence below the 500hz mark. I used these on Tommy's and my red pair noted in the first post of this thread.

The ring in the middle was a stock ring that I taped up with micropore to reduce airflow and seal the vents. The effect of the greater seal is that the bass emphasis shifts from the sub to the midbass, the upper bass and lower mids come more forward, and you get more energy in the 1.5k-2k area. That being said, too much seal and can induce driver flex, which chokes the sub presence a tad and can make the mids glare a bit if you crank it up. To avoid trapping too much energy with the sealed rings, more baffle ports needed to be opened. With the right baffle venting and the sealed rings, you get a warmer tone and a 'rounder' dynamic response. This is what was installed in the above post.

I wanted to cut the difference between the robust sub presence and dynamic sharpness of the stock rings and the more balanced tone of the sealed rings, so I went with a foam that's less porous than the stock foam (for a greater seal) but also cut 2 vents into the sides of the rings and positioned them adjacent to two ports I cut into the outer area of the baffle for more airflow. The center diameter is slightly more open as well, which slightly tones down the middle mids and helps with congestion and nasality. I made these with custom-shaped leather cutting dies. Those are the rings on the right.

Anyway, here's the result -


I never set out to turn the TH900 into a headphone I could use as my daily driver, but with this response I've hit the perfect target of dynamic slam, extension, and tonal balance for my own personal tastes.

Next week I have a few more pairs of drivers showing up to make a clone of this for my own usage (the other I'll be holding for the mostly-open back mod I'm working on). I was considering shipping 2 pairs out for the loaner - one more 'fun and basshead' and the other more 'playful reference', but I think I'm going to make this recipe my standard. Once I have the clone, we'll get the ball rolling on the tour. I know it's been a long time coming but I feel it's been worth the wait - now I just want to produce and share as many as I can and get the open version started!

 

The cups on the loaner pair look awesome.

 

I love the slightly downward yet linear slope! Is it possible to get my pair tweaked like that as well Evan?

 

I'll craft you some rings and install the new 'sushi mod' i've concocted that helps with the 3k area. Not sure if the deeper cups will shift things around or not (the palo santo cups on this loaner pair I've been working on are the same dimensions as the stock cups), but I'll do my best. We have some other things to discuss too.

 

I've had a few profile posts about alternative cup shapes and designs for the foster variant drivers but just wanted to show an example of a different tuning from some of the others I've done. More 'neutral' so to speak. I'm also going to break down an example of what something like this would cost in terms of materials.



These are vented cups - 4 5mm vents about halfway through the depth of the cup. They're African blackwood with a camphor burl stripe in the middle. Build cost on these cups was $300. AP2kTi Pads cost about $50. Pilot pad (optional) $30. This was installed onto a used THX00, which I got for $275. You could find an EMU teak or Denon D2000 for around the same price. We're at $655. TH900 drivers (optional, not installed in these headphones currently) are becoming unicorns but they could be had for $275 shipping inclusive from Full Compass. That comes up to $930. Then everything on my end:

Materials list for stuffing / damping (can differ depending on your target transient and tonal response) - cotton balls, acoustic wool, thin stiff adhesive felt, thick stiff adhesive felt, porous open cell foam, high-density neoprene, dynamat, micropore tape. 50mm cutting die. 75mm cutting die. Hole punch, leather hole punch, scissors, double stick tape. 16 fresh screws. About 4-6 hours of labor (disassembly, cutting, placing, reassembling, measuring, etc) assuming it goes perfectly the first time. Each revision takes about 30 minutes (15 min per cup).

Cup stuffing is very important given one of the mods I use involves opening many of the rear driver vents - there's a lot of backwave energy. I didn't think my so-called 'sushi mod' magnet vent extension / filter would work on the X00 driver because the magnet doesn't protrude as much from the rear of the driver, but I got it to fit and it does help make the midrange-treble transition more cohesive as you'll see in the graphs. I rescind my previous statement about the 600/X00 series driver not being worthy of a good mod:

Empty cups, 6 driver vents open, no sushi mod, sealed baffle, and my custom front neoprene baffle rings with pancake pads. Ignore my speculative scribbles -



And now with stuffed cups + sushi mod with all else the same -



You can use the measurements earlier in this thread or in my sig to reference to other headphones. This is perfect for someone treble-sensitive like me, and transplanting 900 series drivers in the near future should narrow the 3khz dip even more with more presence in the sub region. Right on target. It's possible to get more treble with alternative driver vent and cup damping mods, but I like this.

edit - What you can't tell from the graph is how much the stage opens up with the deeper cups and AP pads, and how the bottom half of the response has incredible solidity and this gorgeous woody tone without any glare or weird audible distortion peaks you'd get from undamped or underdamped cups. It's like if you took an OG LCD-2's bottom half, but with HD650 treble and Auteur-level resonance play with that smooth bio timbre and HE6-level slam. This headphone is pure pleasure, and totally crankable with hotly-mixed pop tracks. Oh, and it gets blaring loud and plenty dynamic from every DAP/portable I've tested it with (Shanling M0, ZX2, Mojo, an iPhone w/ dongle, even Dragonfly)

If you have materials already and want to DIY, I can offer advice via DM. Fair warning - these headphones are very fickle about minor alternations. The same methods don't always work depending on the cup shape, wood, minute changes in materials, etc. If you can't measure yourself, I don't suggest it. It is very tedious. I like the challenge, but if you're easily frustrated by super anal meticulous things, stay far away.

 

The loaner pair blew a driver (extremely old pair, grills fell of the drivers when they arrived, mounting driver glue was flaking off, SN under 300) so I replaced with a fresh pair and did a quick retune. This is now finalized. Inspired by the vented cups, I installed a 'pseudovent' cup mod, which involves risers on the rear of the baffle and longer cup screws so that the driver/baffle unit floats a bit in front of cup while still holding secure fit. Sorry about poor photo quality.



Normally there is a gap here that allows venting from the front of the baffle out to the sides. The baffle is now sealed with the solid front foam rings. The section of the gap that's accentuated is between the back of the baffle and the cup as opposed to the front of the baffle and the pad mount / headband assembly. This helps reduce glare, driver flex, and secondary cup reflections. It sounds more 'open' and less strained overall. The smoother dynamic response (less audible ringing than before) let me swap to the deeper AP2kti pads, which are a more traditional pad shape with a more spacious cavity than the pancake pads I was using before. More comfortable and a more secure seal overall.

Graph:



It's shipping out to @DEATHxMACHINE today or Monday. @ChaChaRealSmooth please start a loaner signup thread. DEATHxMACHINE can decide how much time per member and how many members. I'd like at least 5 strongly interested Friends + Marv and ChaCha, USA48 only for now. I'd prefer those who may be interested in commissioning my services and/or who have a long track record here. Please, no disassembly of any kind (I can send pictures on request via DM to anyone considering a mod). He can pick the rest of the terms.

 

Posting on this thread instead of the loaner thread since the proposed new mod I'm about to mention isn't quite finished yet.

For loaner round 2, I'm going to aim for 100% flat on EARS w/ SBAF early comp curve. Got pretty close here before sushi mod:



and here, with the sushi mod and several rear driver vents open:



To bring up that treble I might close all the vents aside from the one filtering through the sushi mod to bring the treble up. Or stuff the heck out of the cups. Hopefully I don't get any peaks as a result.

If I can get a flat line response with no front damping, I'd be curious to hear what that's like given I have some nearly flat-measuring HD580s which are murderously bright to my ears. Maybe those >13khz peaks on the senns. Maybe the bio timbre is the solution. Or maybe traditionally shaped open cups. Maybe a flat line isn't actually what I'd consider neutral on this comp curve. Lots of questions. The finalized rev 1 on the tour (actually the sixth pair I've made) is just the beginning.

 

Interesting experimental new mod I've been working on since shipping out the loaner pair. Finally putting the D5000 enclosure to good use with some glossy ebony cups and 548843 (TH900) drivers transplanted. This one's modded in the same way as the loaner pair with a few adjustments. I was going to make a clone, but why not try something new?



This iteration was inspired by a few threads where Marv was debunking the myth of 'fast bass' - in an open-air environment, we expect for bass to be 'slower' than the midrange, which is 'slower' than the treble, etc. One aspect of the 'uncanny fakeness' of the headphone sound might be related to the lack of room response and non linear decays across frequencies (hence why CSDs are important, why ZMFs sound like ZMFs, etc). This mod is an exploration of these concepts, despite my inability to measure CSDs well.

To dry out the treble and reduce bloom in the mids, these have a few rear driver vents taped over (7/11 open vs 11/11 on the loaner or 2/11 stock). Rug liner on the cup back wall, with some self-adhesive elastic medical tape (this stuff is a PITA to use) instead of the extra thick felt I used on the loaner. With less backwave energy overall these simply have a dryer, less 'wet' sound, with more definition and transient energy in the treble even though general treble quantity is similar. Matter of preference here vs the loaner.

To bolster the sustain and reinforcement of the low frequencies, I extended the rear flange of the sushi mod. This directly couples much of the primary driver vent's backwave energy to the back cup wall. This turns tactility up to a 10 in the bottom end and this type of coupling wouldn't be possible with XL-type cups. Bass conduction in the cups is quite high - you can feel low frequencies through your fingers on the outside of the cups, though midrange and treble attacks don't seem to have as much conduction (unlike with a 100% sealed cup and baffle) and have no perceptible trails. This flange extension was problematic at first with the sealed front baffle rings as I was getting some driver flex when applying pressure on the pads. So, I vented the ring despite sealing the baffle and that fixed the flex issue without effecting FR.

(light line is a modded Drop HD6xx, darker line is ESX900 #7) Please don't prank call whoever was calling UA at the time lol



There actually is a remarkable resemblance in the general curvature here - if you tilted the 6XX graph clockwise a couple degrees centered around 800hz we'd only be a few dB off in the LF and HF extremes, with a mild 1khz bump on the ESX which I don't mind in a downsloping well controlled response.

edit - tried the above just to see what would happen (this is a bit gorey so apologies, I don't know a better way to do this type of thing)



Anyway, despite the yok'd bottom end on the ESX, with the pseudovent mod + heavier cup stuffing handling the backwave energy from the secondary driver vents, we get a very rapid decay in the top half of the response with just a touch of bloom in the mid band (dialed back maybe 15% vs the loaner). Bass here is indeed more sustained and resounding than mids and treble, but I feel I may have struck a nice trade-off between traditional closed-back style bass with a more open-air style top half.

Thankfully the drivers are so responsive and dynamically 'anti-veiled' as DEATHxMACHINE put it, you can cheat a bit on the graph with a much more robust bass presence than you could get away with by using EQ with another headphone without things feeling bloated. Seriously, nobody says they ever want more midbass than a 650, but somehow these feel less sleepy than the 650 tonally and dynamically despite the reduced treble and enhanced bass. No sense of bloat. Loving it. Happy to ship these out for a demo too. Maybe I'll line it up so Marv has these and the loaner at the same time for a comparative impression.

I also have updates with the XL vented cups and am currently working on a pair of XL passive radiator cups. I'll post those soon.