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

Is Sabre ES9018 the best DAC chip right now?

1 2 3 4 5 6 7
9 10 11
  1. borizm
    Despite my previous mixed feeling about musicality of ES9018 based on the experience with entry-level Audio-GD DAC+headamp (no-loopback, no-opamp, A-class output), I must confess that I have recently bought old flagship-killer: Meizu MX4 Pro phablet/smartphon with mobile version of ES9018 and this small device have surprised me a lot in a positive way. Meizu MX4 Pro sounds really natural without any observed technical sounding artifacts. The sound is sometimes a little too bright, but in acceptable way (not an irritatning-brite). Sounds much better than ES9023 based, cheap, but quite good SMSL M2 (USB DAC) on the line-out (headamp in it sounds in little technical, dirty-like way and can't be fully accepted in the long term, so it's wise to skip it).
    Meizu MX4 Pro can drive various of the headphone types including including even 8 Ohm DT48, 50 Ohm HE-6 to quite enough loud level, and very easily all 250 and 600 Ohms. The SQ is really good - maybe not as good and realistic as in case in old multibit DACs and no-loopback discrete approach with deep A-class output stage, but good enough, I can risk that it's event can be described as a superior SQ in a category of such a small devices.
  2. tf10charged
    I guess thats why orpheus doesnt come with a dac and amp then.

    Oh wait. Why does it still comes with a dac and amp.....hmmmmm
  3. JIGF
    the TI PCM1704
  4. jusbe
    Sure. No, I meant which convertor? Sorry if I was not clear.
  5. Happy Camper
    I didn't read through all of this thread but we are told there is no bearing on sonics in the digital realm. It's in the analog area that you impact sonics. This is why the mention of implementation being the key. So please explain why the DAC impacts the sound? I've heard it said each maker has a house sound.
  6. borizm
    You probably right with "Tube DAC".
    I've already modified a few CDPs (all of them had 2x PCM56/61 or better) by removing all op-amps and providing some discrete I/V similar in sound signature to a tube. Of course doing something with a typical OS was a second crucial part of the modification - for me it was the best to remove and create the NOS or put some better OS IC. Sounds fantastic.
    About WM8741 I'm not sure - sounds too much technical to me, or it was a bad design which I've heard. WM8741 is a "voltage output" DAC, so it has 4 op-amps (or similar staff) build-in in it and there in not a possibility to improve it. Low voltage and "voltage output" makes me worry about possible SQ result even after the perfect modification.
    I wouldn't put TDA1541 and WM8741 to the same basked for the other reason.
    TDA1541 is a multibit R2R, NOS, Iout, 16bit (not very linear, but very natural), and WM8741 is a delta-sigma (less natural by the design), with some OS build-in (can have a signature determined by OS), 24 bit (more precise).
    omniweltall likes this.
  7. dharma
    would this and similar be somehow useful? https://hifiduino.wordpress.com/2014/10/12/r2r-for-the-rest-of-us/
  8. therealbene
    now that Sabre has released the new 9038 PRO series, the 9018 is no longer the best, by definition.
  9. FFBookman
    9038? Fuhgettaboutit!

  10. FFBookman
    Borizm thanks for the posts, I'm learning a lot and I really enjoy your point of view.
    You sound like you work in the industry so I'm wondering if you've heard Charlie Hansen from Ayre describe how he and his company design their audio products. It's fascinating to me as a developer/designer (not of audio gear) as well as a listener, since the SQ of the ponoplayer blows me away at that price point. 
    He said they have built their own extensive suite of listening tests that are not like typical AB or ABX tests, and demands all his engineers select components after spending hours if not days doing these long-term listening tests. The entire build is dictated by the results of these organic listening tests over specs, scopes, or outside recommendations/biases. 
    They basically work under the credo to trust your ears only, but only after an extensive and accurate listening to the component/circuit.
    The credo is played out in their finished products by removing most options and dials.  Many Ayre products will have 1 setting, or at minimum hide away the controls. 
    After some internal debate they even made a product that had 2 settings - the variables as decided by the ears, and the variables as decided by the spec sheet. A proof of concept.
    I can't afford any of their products except the ponoplayer, so this is philosophical for me. Since you work with these types of components so closely, whats your take on this Ayre style of circuit design?  It reminds me of instrument builders - they almost always use their ears and other senses to play the instrument as it's being created.
  11. Kalemkolik
    are which model better Hi-Fi DAC ES9028C2M (LG G5 HİFİ) and ESS Sabre 9018C2M (V10) ? 
  12. GUTB
    After a lengthy wait, I finally got my DIY builder's DAC from China based on the 9018. It's got dual transformers for analog and digital circuits, high-end op-amps, WIMA, ELNA and Nichicon caps, 0.5 ppm clock. XMOS board. Listening to it right now via WASAPI push through a Wyrd. My initial reaction: sound is COMPLETELY IDENTICAL to my Modi 2 Uber. I have a noisy Pioneer VSX-816 amp powering TH900s.
    Will need to go through a burn-in period. There should be some differences between a 9018 and an AK4396 that aren't being masked by a cheap amp...besides the periodic BEEP tones that are coming through.
  13. Ecoli
    So the question is which is currently the best dac chip?
  14. Soundsgoodtome
    For those that are interested in R2R, here are impressions from an ongoing tour I've arranged on a non-oversampling tube buffered R2R DAC. I find these DACs (Stockholm and Pagoda) to be a top performer in it's price range. DS d/a chips sound very sterile in comparison, not because of the tube but the details that's lost in the attack/decay due to DS filters. Impressions are on the very bottom and as the tour continues, more will show up.

    Just to add to the convo and something else for you guys to read:

  15. b0bb
    The ES9018 has the potential, depending on how it is implemented, including the surrounding support logic.
    I set out to understand this a little better and rebuilt and modded 2 ES9018 implementations.
    The first one is the Yulong DA8 MkI and the other a dual Sabre32 LKS MH-DA003
    The Yulong is a pretty close implementation to the reference design from ESS, out of the box, is a poster child for most of the common complaints directed at this DAC. I call this the "textbook" version of the ES9018, as I describe below, textbooks sometimes leave out some of the important stuff.
    I was curious to see if some of these issues can be fixed up and indeed it can.
    The main mod was to introduce a low phase noise clock from Crystek, this took care of most of the indistinct flabby bass issues and made the presentation less sterile.
    In order to improve the soundstage I added discrete voltage regulators from Belleson, this considerably improved the transient handling especially the definition of the  bass transients.
    It was interesting to note Yulong added a similar Crystek XO to the DA8 MkII, which came out about 8 months after I made the mod.
    The main limitation of the Yulong was the use of low cost SMT components for the analog filter, the ceramic caps and SMT resistors do not make for a good sounding DAC in my opinion.
    It leaves no room to make any improvements.
    This leads to my point of the dependence on implementation, overall pretty average DAC, very typical of the presentation of the  ES9018 and reinforces some of the negative impressions of DS DACs in general.
    More details here:
    The second attempt with the LKS produced much more desirable results, this is a Dual Sabre32 with a low phase-noise Crystek XO and a galvanically isolated USB input.
    Shunt regulators were used fro the analog side with the separate power supplies to the various digital parts of the DAC
    Out of the box it was sonically superior to the Yulong DA8  including the one I modded, quite untypical for a Sabre32 DAC, it shared more in common with my R2R DACS like the Schiit Yggy and my Esoteric D10 with Dual TDA1541A .
    It is about 10% more expensive than the Yulong when I bought it.
    This goes back to increased attention to implementation details yields improved performance with the ES9018.
    The basic LKS is overbuilt and the analog stage used quality components like Film Polyproplyene Caps and MELF metal film resistors, these are conventional resistors without the leads for SMT mounting.
    Rebuilding the DAC showed me the possibilities on how far the performance of  ES9018 can be taken to.
    About half of the DAC has been rebuilt starting with the analog stage, the critical I/V opamp was replaced with a discrete unit capable of handling the current from the ES9018 operating  in Class-A most of the time.
    The I/V resistor was replaced  with a precision Vishay metal foil resistor with a 0.05% tolerance, this improves the DC precision to 12bits, most of the common DACs use lower quality components with a tolerance of 1% or worse, this is critical in the ES9018 as the standard implementation requires 2 I/V converters per channel for a 4 in total.
    This was vast improvement and has convinced me to stump up for the cash to have of 0.001% resistors custom made, this brings up the precision to 16-17bits.
    The USB interface was next, both LKS and Yulong use the interface made by Amanero in Italy, however LKS made available an enhanced version with Crystek low phase noise XOs and a low noise power supply to keep the XO performing to their max potential.
    A simple upgrade that moved the sound stage a few steps forward, plus it eliminated most the inconsistent performance we sometimes get from USB Audio, especially the high frequency glare and sterility.
    This comes back to the point of implementation, the LKS was better implemented to start with and opened up greater possibilities to improve its default state of implementation.
    The heart of the Sabre32 is the Masterclock and PLL reclocker, as my modification of the Yulong showed, a good low phase noise XO will transform the pedestrian sound quality to something quite acceptable.
    LKS implemented the DAC in a similar fashion and used a Crystek XO as stock, however LKS implemented a dedicated low noise regulator for the clock and the stock performance was considerably better because of that.
    LKS used a regulator than can supply enough power for a miniature Ovenized XO (OCXO),  without this foresight on the part of LKS, I would not have been able to take the LKS to the next step.
    The Yggy struggles to keep up with the modified LKS. The modifications in themselves improves upon the existing work LKS has done.
    In summary it boils to attention to details of the implementation with the ES9018, and I can appreciate why Accuphase and Gryphon have chosen the ES9018 as the basis of their DACs.
    More details here:
    Traveller and Quadman like this.
1 2 3 4 5 6 7
9 10 11

Share This Page