Replacing Op-Amps in a Topping D10 DAC

Discussion in '[H]ard|OCP Front Page News' started by cageymaru, Sep 24, 2018.

  1. cageymaru

    cageymaru [H]ard|News

    Messages:
    19,490
    Joined:
    Apr 10, 2003
    Audiophiles have been rolling (replacing) op-amps in their DAC units for years. In the computer audio world, high end sound card units from Creative and Asus have had replaceable op-amps in years past and their owners swear by the increased audio fidelity from changing them out. Audio Science Review has taken the challenge to roll the op-amps in a Topping D10 DAC as the op-amps in the unit are socketed. Then afterwards the unit was tested using a $28,000 APx555 Audio Analyzer. The op-amps that were tried in the unit included the OPA-2134 (shipped in the Topping D10), Muse 8820, Muse 8920, LME49720, OPA-1612a, and a 994Enh-Ticha Dual Matched Discrete Operational Amplifier. These ranged in price from $4.79 for the ones that shipped with the Topping unit to $94 each. Here are the results presented in chart form.

    This chart is hugely amplified (vertical axis) relative to what I normally show. Despite that, the differences are tiny. Assuming you want to go by that difference anyway, the best performance was achieved using the op-amp shipped with the unit, the TI OPA2134!!! In other words, the designer knows best. :) I also looked at SINAD differences (not shown) and there was no difference to speak of there. The unit itself varied in performance by 1 dB there and that was what I observed as I changed op-amps.
     
    Stryker7314, AlphaAtlas and Brian_B like this.
  2. tungt88

    tungt88 [H]ard|Gawd

    Messages:
    2,019
    Joined:
    Jan 14, 2008
    Well, there's being an audiophile, and then there's being an audio-fool, like those who pay hundreds for snake-oil cables.
     
    Talyrius, BloodyIron, PenGunn and 9 others like this.
  3. Eickst

    Eickst [H]ard|Gawd

    Messages:
    1,772
    Joined:
    Aug 24, 2005
    No shortage of brand snobs in any hobby/industry
     
    GoodBoy likes this.
  4. PhaseNoise

    PhaseNoise Gawd

    Messages:
    1,002
    Joined:
    May 11, 2005
    And every EE already knew this, including the design team.

    But there is absolutely no deterring people who want "warmth", "feel", "depth", and "presence" and can only find it by spending more money.
     
  5. Riccochet

    Riccochet Off Topic Award

    Messages:
    25,033
    Joined:
    Apr 11, 2007
    talk about getting bent over
     
  6. nyt

    nyt Limp Gawd

    Messages:
    373
    Joined:
    Jan 28, 2008
    Not that THD+N is going to show you everything about how an opamp performs, but this is silly anyway. Better off just using a better dac that isn't so noisy in the hf ranges in the first place. The D10 is not very good.
     
    clockdogg likes this.
  7. iceperson

    iceperson n00bie

    Messages:
    8
    Joined:
    Feb 22, 2007
    You know what they say, "An audiophile and his money..."
     
    viscountalpha and clockdogg like this.
  8. clockdogg

    clockdogg Gawd

    Messages:
    695
    Joined:
    Dec 12, 2007
    "high end sound card units from Creative and Asus"

    Creative? Asus? High End? LOL! How the high standards have fallen deep into the mid-fi marketing mud. Topping is the definitive low-end Chinese brand with mid-fi pretensions.

    Swapping op-amps is a low-end pursuit. Not a high end one. Far too affordable. $94 op-amp. Puuuleeze!

    Swapping rare NOS tubes like an ECC32 or Tungsol Roundplate 6SN7s in your 5k headamp is a more lofty pursuit worthy of the description High End. Although that example is pretty entry level. Proper High End DACs begin above $2k and don't stop until all of your money, sanity and will-to-live has been drained.

    Calling Creative and Topping high end audio is like pretending an Intel HD4600 is high end gaming.
     
    GoodBoy and Incontentia Buttocks like this.
  9. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,753
    Joined:
    Jun 25, 2004
    The 2134 is great! I use them fairly frequently in some of my filter designs. I will say there are some other properties of op amps not being looked at here. THD is a useful measurement until you hit certain levels. I use the LME49720s in my output amplifier modules, and it performs quite nicely. I don't graph my results though. For me, I just lay out the circuits, try a few of my usual favorite parts, and see how it sounds. The 49720 will drive higher loads though, which is handy. I'm working on an output module off and on in my spare time that will use these, plus the 49600 high current driver, and a DC servo, so that I can eliminate the AC coupling at the output. This is all for fun, and this stuff does make it into some of my synthesizers.

    The OPA1642 is great too.

    Thing is, in a lot of cases you can use bog standard op amps, and get really good results with a bit of extra circuitry. I just go for the sweet spot though. I go for something that's proven to be good for the application, that's not WAY too expensive, and call it good. The LMEs are cheap for what they are, which is nice. I use the older LF series National ones quite a bit for general purpose. For the few projects that I'm working on though that are intended to be mass produced (I use that loosely based on the niche space I work in for this (typically low three digits)) we are designing the circuits to behave nicely with TL07x op amps. We need so many of them in one project, and in such a small space, that even things like the 2134 will boost the cost through the roof. (One design is using close to 50-some op amps spread across dual packages)

    Anyway, it's always fun to see things like this no matter the results. I'd recommend taking a look at Douglas Self's Small Signal Audio Design book. It's pretty interesting if you're into this sort of thing. It could probably use an update with some of the newer devices around, but it's got some interesting ideas and designs in it.
     
    NoNRG and clockdogg like this.
  10. IOTW: Below audible threshold.

    Save your money. Take this from someone who used to design & build his own amplifiers in college.

    Where things get interesting is how output amplifiers hold up on high current complex loads. Scopes don't show that well. You can drive 32 Ohm headphones, and 8 Ohm Nom speakers. But truth being told the resistance modulates with the frequency. So if you are playing a wideband fq, then the power output delivered will vary. My nautilus's go to 3Ohms, but they are 8 Ohm Nom. If you push them on all but the best amplifiers to volume, you could overload your amp and get over 1% THD.

    Relative damping factor, total VA, and all channels driven WPC are the numbers you should look at.
     
    Last edited by a moderator: Sep 24, 2018
  11. You know tubes just round off the high end due to electron flow over increasing the THD right? That's what gives it a more mellow sound. Sort of like filling up a glass with water and washing it slush over the side unevenly. Bob Carver created a similar effect on receivers he sold little over a decode ago. (Amp driven versus voltage driven) But it introduced THD.

    They are great if you have hard to drive loads that don't require a lot of power (Maybe headphone). But it's like the hispter movement that still insist on buying vinyl. It's bull shit. Nothing personal.
     
    Last edited by a moderator: Sep 24, 2018
    haste., Betaboy1983 and alxlwson like this.
  12. Darunion

    Darunion Chin Poon Specialist

    Messages:
    3,331
    Joined:
    Oct 6, 2010
    It is almost like when they designed the amplifier stage they picked what opamps they wanted and then choose the supporting components based on what the opamps required. So changing out the opamps would essentially 'detune' the circuit.

    IMO if you aren't looking at the datasheets for these before putting them in, you probably shouldn't be doing this.

    If you can get the schematic and BOM for this DAC, that would be a great place to work from and see where improvements could be made. Though I would honestly just move to a DAC that does what you want it to do. I know this isn't the [H]ard route but there are better things you could start with before messing with the opamp stage of an existing product.

    The OPA2134 is a decent chip on its own.
     
    J3RK likes this.
  13. nyt

    nyt Limp Gawd

    Messages:
    373
    Joined:
    Jan 28, 2008
    That's just not really true. There are ess and akm dacs that are incredibly transparent and inexpensive.
     
    J3RK likes this.
  14. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,753
    Joined:
    Jun 25, 2004
    There are subtleties. I personally go for "good enough for me" which is typically fairly exacting, but nothing like the audiophiles out there. Sometimes it's nice to try to squeeze performance out of more general purpose parts as a design goal. Other times, it's fun to design something with higher end parts and brute force results. I find it's best to stay in the realm of good solid parts selected for the task at hand. I don't go cheap, and I don't get out into snake-oil land. There are plenty of good parts from good companies that perform well for a decent price. As far as tubes, I don't dabble in them myself, but have heard some pretty impressive results. It's all subjective though. If something sounds good to you, then use that, whether it cost $75 or $7500. (or even $75,000 if you're nuts) Some people like to get lost in the minutiae. That's fine too as long as they've got the bank account for it :p It all does reach a point though where the word "asinine" comes to mind. :D

    Also, I like vinyl, own quite a few records actually, (have a lot of DJ friends from way back) but I don't actually own anything to play them on now. :D I would never get into a debate about what formats of audio sound better. I kind of like it all. There's a place for everything.
     
  15. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,753
    Joined:
    Jun 25, 2004
    I can't remember the part number, but ESS put out some REALLY nice ones recently (last few years). They're also very inexpensive (in relative terms).

    Interesting side note: ESS was formerly Electronic Speech Systems, the people that did the digitized voices for the game Impossible Mission (C64 etc.) :D Pretty cool!
     
    Deleted member 93354 and nyt like this.
  16. Come and stay a while. Come and stay forever. Mu Ha Ha Ha

    Teh SABRE DACS are da' bomb though.
     
    tungt88, nyt and J3RK like this.
  17. nyt

    nyt Limp Gawd

    Messages:
    373
    Joined:
    Jan 28, 2008
    I have a motu 8a in one of my systems that I used for some extra IO. It uses a lower end 9016 multichannel dac. I measured it and distortion was non existent until signal levels were pushed real high. Something like 0.00002% THD on the bench. When pushed to right below clipping it 'skyrocketed' to 0.00009%. THD+N for those measurements was something like 0.0006% and 0.0002% respectively. Their high end dacs like the 9038pro are even more ridiculous.

    edit: found the measurements

    deQZNoX.png

    9uFmAM3.png
     
    pfc_m_drake and J3RK like this.
  18. clockdogg

    clockdogg Gawd

    Messages:
    695
    Joined:
    Dec 12, 2007
    You know that's BS, right. A well-designed tube circuit has plenty of bandwidth for audio. And let's not get into the THD single figure of merit debate - it's nice and simple for a spec sheet, but is more for marketing than insight. Even still, a good triode like a the ECC40/E80CC can have extremely low THD - but more importantly, tube distortion characteristics (in low or zero feedback designs) are more benign that SS devices with predominant 2nd order distortion vs odd-order distortion characteristics of SS that generally use copious amount of negative feedback to achieve gain and distortion targets.

    The hipster vinyl movement got to the party late. Pretty sure they're not spending $20k+ on the their turntable rigs to enjoy that amazing 26dB of channel separation at 1khz and mono bass. Although, vinyl is not subject to the cruelty of the Loudness Wars that despite the promise of 16bit 96dB dynamic range, we got 6-8dB crest factors and headaches.
     
  19. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,753
    Joined:
    Jun 25, 2004
    Love the term "skyrocketed" there. ;) I don't have enough instruments right now to need a big interface, so I'm just running a smaller Focusrite. I'd like to pick up another big interface at some point though. Just need to build some more synths to make it worthwhile.
     
  20. clockdogg

    clockdogg Gawd

    Messages:
    695
    Joined:
    Dec 12, 2007
    Yes, but if they're inexpensive, they're not High End. However there a few High End DACs that stoop to using ESS chips that cost several thousand dollars and don't sound much better than the harsh $100 ESS DAC models. And Dammit! That's the point. High End Audio is about over-spending for the last 1% of perceived enhancement!

    High End Audio is not about value-for-money. It's about exclusivity, wine-snobishness-with-ears-on-it and 2" thick hand-milled aluminum faceplates made in the USA!
     
  21. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,753
    Joined:
    Jun 25, 2004
    Hey! Don't knock milled aluminum faceplates!! (I guess mine aren't 2" thick though, and aren't milled by hand...) :D

    Obligatory synth shot:

    1931_silversynth004_1.jpg
     
    clockdogg likes this.
  22. Dullard

    Dullard 2[H]4U

    Messages:
    2,997
    Joined:
    Jun 28, 2012
    This ^

    There are tiny little coupling capacitors connected to the DAC chip, and the value is pretty important for each DAC/circuit combo. You might get lucky, or you might spend a nice chunk o green and not hear any improvement or even sound worse.
     
    Darunion likes this.
  23. clockdogg

    clockdogg Gawd

    Messages:
    695
    Joined:
    Dec 12, 2007
    Oh...very pretty. Getting some Buchla flashbacks. Is that your own design?
     
  24. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,753
    Joined:
    Jun 25, 2004
    Thanks! Yeah, it's a bit Buchla / Serge / Wiard -esque in nature. Low Pass Gates, Complex VCOs, Linear Slopes, etc. Those guys tend to inspire my designs the most. My dad made the cherry-wood case. I've slowed down my building recently to spend more time with my family, but still do a good amount from time to time. I've laid out a few new PCBs that I want to try out sometime, so I feel another system coming on. I sold the one in the pic recently.
     
  25. clockdogg

    clockdogg Gawd

    Messages:
    695
    Joined:
    Dec 12, 2007
    That's awesome! Great work! Does it have an analog sequencer module? Will you be releasing some recordings from it? Targeting R2R tube DACs of course. ;-)
     
  26. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,753
    Joined:
    Jun 25, 2004
    Thanks again! The bottom panel is an 8 step analog sequencer, (the knobs and buttons on the sides are manual CV and gate buttons, for triggering additional events on the fly). It can also be clocked up into the audio range, to be used as a wave shaper. The module above it with the 8 LEDs, is a Linear Feedback Shift Register based pseudo-random voltage source, that either samples white noise for data, or an eternal input, or Linear Feeback. I use a Bourns R/2R in that for 8 bits worth of voltages. ;)

    I'm just now starting to get back to actually writing some music. For the longest time I had just been building synths, and making little demo clips and stuff like that. You can hear some of demo sounds on my Soundcloud page. The "Stroh Modular Synth" playlist is mainly just demoing some of the features. Not terribly musical, but kind of fun sounds anyway. Most of the rhythms are generated with Boolean logic functions pre and post sequencer. I find that XORing two clocks makes interesting beats. :D

    https://soundcloud.com/drj3rk

    Sorry for straying off topic. :D
     
    gtrguy likes this.
  27. clockdogg

    clockdogg Gawd

    Messages:
    695
    Joined:
    Dec 12, 2007
    I'm not. Far more interesting than the D10 topic.

    Will checkout your soundcloud demos. Thanks!
     
  28. Nukester

    Nukester [H]ard|Gawd

    Messages:
    1,398
    Joined:
    Mar 21, 2016
    Whew, reading through this I realized just how out of my knowledge this all is. Just give me my airpods and let me remain oblivious.
     
  29. nyt

    nyt Limp Gawd

    Messages:
    373
    Joined:
    Jan 28, 2008
    lol harsh $100 ess dac? I assume you mean the 9038 which is so far beyond transparent it's absurd. As absurd as calling it harsh.
     
  30. nyt

    nyt Limp Gawd

    Messages:
    373
    Joined:
    Jan 28, 2008
    I use it for IO for processing in my theater. It's chained off of an RME UFX+ via adat.
     
  31. JargonGR

    JargonGR Limp Gawd

    Messages:
    440
    Joined:
    Dec 16, 2006
    Once upon a time I spent a shitload of money on Audio & Video gear. I spent countless hours studying deep into the subject in order to understand the differences between speakers, amps and room acoustics. My first big sound system comprised of B&W 800Ds Speakers ($20,000), HTM1D Center Speaker ($8,000), 4X B&W 803Ds for surrounds and 2X more B&Ws for extra height channels. For amps I have used Brystons & Electrocompaniet Nemos.

    I also tested several high end dacs ($0,000+) and played with very expensive 2 channel systems at shows and audiophile houses. Yet, after participating in countless discussions on each and every audio forum, reading white papers, research papers and testing I finally realized what is going on.

    There is a lot of snake oil in the industry and A LOT of people that get carried away through placebo on high end gear.

    - Cables? Ridiculous pricing for cables that either don't do jack to your audio chain or alter the sound in the wrong way acting like an equalizer something that they should never do. I tested expensive cables (like $5-6,000 per L&R pair) but I never bought them. NEVER. I stick to Belden for my cables (large manufacturer) and I would only carefully buy based on the electrical characteristics of them in relation to the audio application (i.e. resistance, capacitance).

    - Speakers. There are certainly differences there and in order to enjoy full range and high dynamics you need to pay some "euros" period. I don't mean high-end brands only of course but you won't reach the limit of performance in a

    - Room Acoustic. Huge topic and of extreme importance. After all it is the room that you hear in a way since soundwaves from your speakers get reflected off the walls as well before reaching your ears and this can make or break the sound fidelity. It is as critical as selecting the right speaker for your room. I am a big advocate of room correction but done properly through measurements, appropriately and custom designed diffusers/absorbers and bass management.

    - Amps - another snake oil pit but there are exceptions of course. I am perfectly fine with my Brystons or even some high end professional amps like MC2 audio especially for high output home theater systems.

    In the end I sold my B&Ws after experiencing a JBL synthesis system that knocked my socks-off in both movies and music and realized what I want. Although owning 2X big JBL Everests DD6700 or K2S9900 would be a dream come true the price of those babies for a multichannel system is simply prohibitve ($000,000) so I settled for something less. SH50s from Danley Sound labs and I called it a day. Sometime in the future, when kids education expenses are over, I know I will get a pair (used) of at least K2S9900 or equivalent along with some big bad Bryston monoblocks but till then I am fine.

    The only thing I might go for sooner rather than later would be a Metric Halo ULN-8 external audio interface ($3-4,500) and a Lynx AES-16e or an RME card on my PC and lots of measurements for room treatments.
     
    GoodBoy and Meeho like this.
  32. clockdogg

    clockdogg Gawd

    Messages:
    695
    Joined:
    Dec 12, 2007
    No more absurd than calling it 'far beyond transparent' - it's still an oversampling sigma delta DAC - which is a fine technique for budget-constricted designs. It's not a 'perfect' method.
     
  33. nyt

    nyt Limp Gawd

    Messages:
    373
    Joined:
    Jan 28, 2008
    You can put together some JBL stuff for cheaper than the SH50s that will best them if you're ever interested. I've converted a few people who had SH50s. I demoed them in my room at one point many years ago and they just didn't cut it. The detail was just lacking.

    I'm using some of the JBL 2452H-SL compression drivers with Be diaphragms on 2384 waveguides over a 2x15 lf section running fully active with some fancy processing. The CDs are the only expensive bit, which run about 1k a pop. 4" compression drivers are awesome when not using crappy materials. The 4367's are also good speakers and not stupid expensive, but the D2430K they use, while a good CD, just doesn't compare to 4" models imo.

    fr at the listening position... distortion is < 0.1% THD+N across the bandwidth of the speaker at 85dbspl at the mlp, most of which coming from the mic and noise floor anyway.
    tDHrMe6.png
     
  34. nyt

    nyt Limp Gawd

    Messages:
    373
    Joined:
    Jan 28, 2008
    If you think you're hearing any artifacts from dacs at these levels you're nuts.
     
  35. clockdogg

    clockdogg Gawd

    Messages:
    695
    Joined:
    Dec 12, 2007
    Yes! That's the whole point. Getting carried away. Probably to the poor house and divorce court. :ROFLMAO:
     
  36. JargonGR

    JargonGR Limp Gawd

    Messages:
    440
    Joined:
    Dec 16, 2006
    It sounds great and might explore the option but my SH50s are not "Stock" and I have 3 of them (LCR) since they are mostly used for Home Cinema (they are really powerful MOFs). They are actually custom made (wood) and have undergone some fine tuning. They can also be further improved by using active crossovers and this is straight from Tom Danley that can also provide assistance if needed. Now a Metric Halo could be certainly of assistance here.

    Yet, I have made custom stands for them that cost quite a lot so I don't intend to sell them. I really like them to be honest.

    Here are the stands (Custom designed, 75KG each, spring loaded to assist telescopic height adjustments, tilt, rotation and hidden steel ball wheels). IMHO the stands are crucial for how those Danley's sound and since their main application is not intended to be the Home Environment they are mostly omitted from the chain.

    IMG_5486.JPG IMG_5489.JPG IMG_5483w.jpg


    Anyway, regarding detail I don't see them lacking compared to my B&W 800Ds but that is a totally different speaker. They extend to 18KHz that is more than enough (who can hear above that after the age of 5?) and I do find the resolving not to mention how dynamic they are. Anyway, a lot depends on the room too.

    Regarding your JBL suggestion I would love to play with that option and I certainly love JBL speakers, but it won't happen soon since I am really at a time deficit most of the time. Ideally a used pair of KS???? at about $15K in the future should be feasible (there are used Everests for $25-30K) but I am not even thinking about it now since kids have lots of needs and the economy in my country (Greece) needs some serious fixing before I even thing of doing such things. LOL - First world problems those are......
     
  37. nyt

    nyt Limp Gawd

    Messages:
    373
    Joined:
    Jan 28, 2008
    Anything compared to b&w will sound good ;)
     
  38. You do know B&W's sit in like 90% of recording studios in the world?

    Skywalker_Sound_Scoring_updated.jpg
     
  39. Capacitors are just a cheap way to filter out of noise. But they introduce phase shift in the process.