UM 3DT Review and Measurements

Discussion in 'IEM Measurements' started by purr1n, Dec 24, 2021.

  1. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    A few months back, I posted some thoughts on why I felt balanced armature IEMs suck: https://www.superbestaudiofriends.org/index.php?threads/why-balanced-armature-drivers-suck.11485/

    I think or at least I would hope that readers understand where I was coming from; because after all, the CFA Ara and OG Solaris remain my go to IEMs - and these two IEMs use balanced armatures at least in some spots (I'm bummed because I lost my Aras, or I think they may have been stolen by the guy who installed our blinds in the house). Here at SBAF, discussions must all be taken into context. We're not afraid to rip apart the stuff we like because we can see something potentially better. Our ego's aren't wrapped into our gear.

    In that thread, it was suggested that instead of the DD/BA hybrid approaches which are always problematic because of discontinuities between the driver types, a DD/DD approach akin to mutli-driver speakers could be the ticket. In theory, perhaps a multi-driver DD would fix the downsides that seem to come with every DD IEM: peaky highs.

    UM has taken this approach with the 3DT. The 3DY uses x2 DD bass drivers and another x1 DD driver for the mids and highs. Does the 3DT bring us the best of DD tech without the downsides of DD?

    Unfortunately, the answer is no.

    UM 3DT IEM
    Frequency Response
    upload_2021-12-24_19-57-14.png
     
    • Like / Agree Like / Agree x 10
    • List
  2. james444

    james444 Mad IEM modding wizard level 99

    Friend IEMW
    Joined:
    Jun 3, 2016
    Likes Received:
    1,852
    Trophy Points:
    93
    Location:
    Vienna, Austria
    I wonder if (at least in some cases) peaky highs may be an intentional tuning decision rather than an unavoidable DD flaw. Remember the JVC FD01's stock tuning, that was pretty peaky too? And so easy to fix with just a bit of damping - which eventually paved the way for the FDX1.

    Well, the same applies to the 3DT: annoyingly peaky stock treble, but could have easily been avoided with minor damping - as illustrated in this post:
    I wonder why IEM manufacturers keep churning out unnecessarily peaky sounding DDs? Is there a large target audience that mistakes treble peakes for "micro detail"? I wish I knew!

    Anyway... guess what I'm trying to say is: the 3DT, with smoothed-out treble peaks, is a pretty agreeable-sounding DD.
     
    • Like / Agree Like / Agree x 8
    • List
  3. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I couldn't find any 3M micropore tape at the drugstore yesterday so I will have to wait under the Amazon order comes. In the meantime, I will try some mask material and report back.

    No doubt the UM 3DT is going to be compared to the FDX1 because both are DD and both are in the $300 (USD) range, so let's start there.

    I preferred the FDX1 with the blue filter which was the most aggressive in pushing that upper midrange bump down. I know it's a matter of semantics maybe, but I wouldn't have categorized FDX1's upper mids as peaky, but rather as elevated or bumped. The 3DT sounds not just elevated, but also peaky, implying several high Q peaks or unevenness. The FDX1 has a smoother response. Crucially, the FDX1 not as hot on the upper fridges of the upper mids and lower mids.

    UM 3DT vs Drop FDX1 (both stock, or non-colored nozzle for FDX1)
    Frequency Response
    upload_2021-12-25_8-28-58.png

    Where the 3DT subjectively beats the FDX1 in the lows. The 3DT seems to be a little less physicality compared to the FDX1, but the FDX1 can lay the bass down a bit thick at times. The 3DT bass also sounds quite nimble or "fast" for a DD. With respect to transients, the UM 3DT reminds me of the DD drivers used in the CFA hybrids (their timbre and texture is different).

    Finally last few points: The 3DT still doesn't quite sound as cohesive throughout the audio band compared to the FDX1. The FDX1 sounds more resolving, higher-fidelity, than the 3DT. There's kind of a "weaksauce" (relative) sound to the 3DT which is hard for me to explain - other than it sounds like tube families that I do not like, e.g. 12AX7. I'll take distortion measurements and report back comparing to FDX1. I'm betting there is something there.
     
    • Like / Agree Like / Agree x 6
    • List
  4. Merrick

    Merrick A lidless ear

    Friend
    Joined:
    Jan 6, 2016
    Likes Received:
    8,828
    Trophy Points:
    113
    Location:
    Portland, OR
    Interesting that you find the FDX1 to be more resolving than the 3DT. Perhaps it’s the pushed treble but I find the 3DT (with micropore) more resolving of fine details and the bass reproduction is no contest. Not only is the FDX1 slower in the lows, but I hear significantly less definition. With the 3DT I can hear the plucking of the strings of a bass guitar or the difference between two kick drums played in quick succession like you often hear in speed metal. With the FDX1 it all sounds like a series of low end thumps without definition.

    I’ll have to run some more comparisons here when I get a moment.
     
    • Like / Agree Like / Agree x 2
    • List
  5. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Ignore the 200Hz 3rd harmonic spikes (green line) in the measurements below.

    UM 3DT
    Distortion (absolute) at ~104db
    upload_2021-12-25_8-59-0.png

    Drop FDX1
    Distortion (absolute) at ~104db
    upload_2021-12-25_8-58-12.png

    Overall H2 is similar around -70db in the lower mids. FDX1 has a little higher H2 starting below 80Hz. FDX1 also has slightly higher H2 in the upper mids.

    However, I think the main difference is with the higher orders. H4 on the FD1X stays low throughout, a slight rise below 80Hz. Similar for H3, gradual rise below 100Hz - but with a narrow bump at 3.4kHz. But otherwise from 100 to 2.6kHz, the higher orders remain non-existent on the FD1X.

    The same cannot be said for the UM 3DT. There's a nice low spot between 300 and 1.6kHz. From up 1.6kHz, there's a steady rise that keeps going up. Taking into account the H3 200Hz artifact, H3 on the 3DT is higher too below 300Hz. There's also a H2 and H3 distortion peak just past 5kHz.
     
    • Like / Agree Like / Agree x 5
    • List
    Last edited: Dec 25, 2021
  6. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    To be clear, more resolving of microdetail in the amplitude domain in the highs. With transients, time domain, the FDX1 is sluggish. The bass is smeared. The 3DT has better definition, probably because of a mid-treble peak (not totally sure it's there per the FR), but the plankton seems fake to me, that is the Photoshop sharpening effect.

    I'm also wondering if there is an ultrasonic peak somewhere. The timbre of the highs on the 3DT is a bit unnatural to me. I wonder if it's related to that H2 spike just past 10kHz, and also that H2/H3 spike just past 5kHz.

    Now I have yet to mod the 3DT.
     
    • Like / Agree Like / Agree x 3
    • List
    Last edited: Dec 25, 2021
  7. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The answer is simple: it's because people keep sidegrading.

    https://www.superbestaudiofriends.o...d-iem-hate-confusion.11787/page-2#post-369349

    85% of the effort for product development goes toward getting it right, final tweaks, just before production. Most people don't realize how many prototypes from Eddie Current, Schiit, Drop, etc. get left in the dustbin. For every one product that makes it to production there are at least three protos that don't, and five ideas, proof-of-concepts, that got killed early on.

    True stories:

    "Hey Marv, check this out. I want to hear your take on this."
    "It sounds like ummm, how do I put this..."
    "Like shit?"
    "Yeah"

    "Hey Marv, check these IEMs out"
    "Why do the left and right channels sound different, which is it supposed to be?"
    "We'll get you another good sample"
    "Why does this good sample sound yet different from the original one?"
    "Well get you another three samples"
    "WTF, why does everything sound different?"
    "Oh shit"
    "Just price them at $39"
    "Good idea"

    I'd be willing to bet that most of these IEMs manufacturers just kind of say to themselves "fuck it, good enough, write dumb marketing document highlighting our goofy tech (meta materials, meta drivers, meta housings, number of drivers, etc.) send a few free ones to shills on HF, people will buy it anyway".

    I mean, really, I have to find micropore tape to fix a fucked sounding IEM? I'm tired of having to do stuff that the manufacturer should have done themselves in a more elegant manner in the first place (Helmholtz resonators, special chambers and sound guides, Holtzman effect, Axlotil tanks, etc.)
     
    • Like / Agree Like / Agree x 5
    • List
    Last edited: Dec 25, 2021
  8. james444

    james444 Mad IEM modding wizard level 99

    Friend IEMW
    Joined:
    Jun 3, 2016
    Likes Received:
    1,852
    Trophy Points:
    93
    Location:
    Vienna, Austria
    My thoughts exactly. I mean I love tinkering with IEMs, so I don't mind. But I can understand anyone who refuses to slap $2 tape on a $300 IEM to make it sound agreeable.
     
    • Like / Agree Like / Agree x 3
    • List
  9. Biodegraded

    Biodegraded Friend

    Friend
    Joined:
    May 28, 2017
    Likes Received:
    5,330
    Trophy Points:
    113
    Location:
    Vancouver BC
    I haven't heard the 3DTs, but I've found the character of the FDX1s to be very sensitive to what they're driven by, and I don't think it's related just to the output power: ostensibly very similar dongles can make them sound pretty different.

    This. So many of these companies seem focused on churning out new models so they can generate new hype. If they're similar to others that have been hyped, have the bass & upper mid boosts that have succeeded especially in the Asian market before and which allows them to say "Harman", and if they have treble that masquerades as detail as @james444 alluded to, a new train on the timetable is almost guaranteed.

    Could you do an impedance measurement? I'm not expecting the variations to be significant for amp choices, but it'd be interesting to see how the 3 drivers tie together. Perhaps phase shifts could partly explain the perceived lack of cohesion?
     
    • Like / Agree Like / Agree x 3
    • List
  10. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I don't mind minor tweaks, e.g. tape mod on Grado bowl pads. However, finding the right tape, cutting it, and dicking with tiny IEM shit, that takes time and is annoying. At least Grado listened and provided F-pads for the SR325X and SR225X. And then there's the GAudio Nair, where I did absolutely nothing, nada, zero effort, and they sounded wonderful and right.
     
    • Like / Agree Like / Agree x 1
    • List
  11. Merrick

    Merrick A lidless ear

    Friend
    Joined:
    Jan 6, 2016
    Likes Received:
    8,828
    Trophy Points:
    113
    Location:
    Portland, OR
    I mean, Grado’s been producing headphones since the late 1980s and they’re just now giving us F pads.
     
    • Like / Agree Like / Agree x 2
    • List
  12. james444

    james444 Mad IEM modding wizard level 99

    Friend IEMW
    Joined:
    Jun 3, 2016
    Likes Received:
    1,852
    Trophy Points:
    93
    Location:
    Vienna, Austria
    I don't usually post distortion measurements, 'cause I know too little about them. However, in case of the 3DT it's pretty striking, how that slight amount of treble smoothing via micropore tape has a quite noticeable beneficial effect on higher order distortion across the FR.

    FR comparison (red = stock, green = micropore modded):
    [​IMG]

    Distortion stock:
    [​IMG]

    Distortion post mod:
    [​IMG]
     
    • Like / Agree Like / Agree x 1
    • List
  13. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Yeah, that's why everyone on Changstar/SBAF shit on or ignored them for past 10 years and preferred the Joe Grado HP1000 from the early 80s.

    ---

    UM 3DT
    Frequency Response
    RED = Stock
    CYN = x2 layers outer mask material
    ORA = x1 layer out and x1 layer middle mask material
    upload_2021-12-25_11-6-54.png

    UM 3DT
    x1 layer outer and x1 layer middle mask material
    Distortion at ~104db
    upload_2021-12-25_11-31-20.png

    Much better now. The 5kHz peak looks to have been eliminated. Not only that, the 5kHz distortion is gone. The distortion characteristics actually look very nice with H2 being dominate, and H3 and H4 being well below H2. There is a slightly rise in H3 in the upper mids, but nothing's perfect.
     
    • Like / Agree Like / Agree x 2
    • List
    Last edited: Dec 25, 2021
  14. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Now we are making some progress. I took two layers of the three layer mask material sandwidth. I'll try the micropore when it comes in since this x2 layer sandwich could be a bit on the aggressive side with making the tonal response more amenable. Then again, I don't think the micropore will be enough for me - it doesn't kill that 5kHz peak - so I could be stuck in a bad place.

    The main downside with shoving stuff in front of the nozzle is attenuation of the last octave and some killing of liveliness (perhaps we can call this the Aeon effect). This is any different here, but the effect isn't severe. It's like most DD IEMs don't have plenty of air anyway. Even without mods, the lack of air and shimmer with metal percussion was notable, especially compared to a proper headphone HD6**. The sound is a bit stuffier though compared to IEMs that just sound good without having to do tape or material mods in such a heavy handed way.

    With respect to subjective tonal response, the HD600 and CFA OG Solaris, even with their imperfections, sound so much smoother and grown up. The 3DT sounds less refined and rather sizzly. The timbre is much better with the modded 3DT, however, there's still that spotlit peak that won't go away. Compared to the OG Solaris, it's like pick your poison: some BA timbre vs. spot-lit sizzle. I'd call it even. The DD goodness still comes with its drawbacks: peakage. Rawer BA timbre still comes with its advantages: clarity and transients. Although it's probably not fair to compare with the CFAs since their BA implementations are among the better ones (and also in another price class).

    The 3DT does have some "plankton" that only DDs seem to be able to pick up. It's not on IE800 or Dunu Zen (I have not heard Pro yet) level. It's most certainly not on HD6** level. But it pips any BA, including any high-end ones. I'd say about on par with Grado SR225's ability to resolve low-level information which isn't bad - quite good really (This is really for you @crenca, IEMs are not ready for you).

    Component chain: Yggdrasil LIM -> Magni 3+
    Music used: Daft Punk, Alice in Chains Unplugged, lots of old big band stuff from Eartha Kitt and Bing Crosby.
     
    • Like / Agree Like / Agree x 1
    • List
    Last edited: Dec 25, 2021
  15. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    In a nutshell:
    • The price is right, especially with the recent price reduction.
    • Conditionally highly recommended, provided you are willing to put in some elbow grease. Difficult highs. timbre, and distortion are largely addressed with tape mod.
    • Doubtful everything can be fixed. Mid-treble peakage cannot be totally fixed. There are downsides with taping stuff over the nozzle, stuffier sound, and attenuation of last octave. Some experimentation for ideal material to suit personal tastes and find the right trade-offs may be necessary.
    • Nice DD timbre, but DD downsides (treble peakage) not fully addressed
    • Good resolution considering price.
     
    • Like / Agree Like / Agree x 3
    • List
    Last edited: Dec 25, 2021
  16. purr1n

    purr1n Super Friend

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    75,548
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The FDX1 seems to prefer desktop amps that can cut through, otherwise the lows are a mushy mess. Also, desktop gear is more resolving, which the FDX1 can take advantage.

    x2 7mm DD for the lows.
    x1 10mm carbon nanotube DD for the highs.

    The Carbon Nanotube stuff is the latest crazy, now that folks realized EST sounded like shit.

    UM 3DT
    Impedance and Electrical Phase
    upload_2021-12-25_12-44-48.png
     
    • Like / Agree Like / Agree x 2
    • List
    Last edited: Dec 25, 2021
  17. Merrick

    Merrick A lidless ear

    Friend
    Joined:
    Jan 6, 2016
    Likes Received:
    8,828
    Trophy Points:
    113
    Location:
    Portland, OR
    For me, ~$300 is all I’m willing to spend on IEMs because there are too many compromises for me to spend $1k on a pair, even if the $1k is objectively better than the $300 pair. So for me, good enough at this price point is reasonable.
     
    • Like / Agree Like / Agree x 4
    • List
  18. Merrick

    Merrick A lidless ear

    Friend
    Joined:
    Jan 6, 2016
    Likes Received:
    8,828
    Trophy Points:
    113
    Location:
    Portland, OR
    Well, I can confirm this. On my RU6 and ZX300, the FDX1 has mushy, in distinct lows in comparison to the 3DT. Plugging it into the Vali 2+ and the bass becomes more defined, clear, and discernible. It’s still slower than the 3DT but the performance is much closer out of the Vali. The 3DT lows sound the same out of all three devices. It’s frustrating that an IEM requires so much power to sound its best, goes against the whole idea of these as portable.
     
  19. Biodegraded

    Biodegraded Friend

    Friend
    Joined:
    May 28, 2017
    Likes Received:
    5,330
    Trophy Points:
    113
    Location:
    Vancouver BC
    Gonna stick with my contention that the FDX1s are not all about power. Their bass response is quite different (worse) from the Helm Bolt (which otherwise I think is the better dongle) than from the Shanling UA 1, and even the much lower powered Audioquest DragonFly Black I find better for their bass than the Bolt (even though the Black is hella warmpoo). And from the balanced output of the EarMen Sparrow (higher output, sure) their bass is great.

    Power, sure; but synergy, too.
     
    • Like / Agree Like / Agree x 2
    • List

Share This Page