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

Thoughts about performance priority button-less dongle DAC.

Discussion in 'Sound Science' started by E1DA, Apr 13, 2019.
  1. E1DA
    I finished HW implementation and testing for the precise & tiny DAC+HPA based on ES9038Q2M+AD8397 fully differential current mode bridge, USB-C input, 2.5mm jack 4 wires outputs, over 300mW@32ohm, 24/192, 46x20mm PCB. The DAC IC offers some THD manipulation options and I wish to let users choose 3 modes:
    a) normal mode with instant PLL locking THD@1kHz@-3dbfs .0005% THD@1kHz@-10dbfs .0003%(measured on the real HFM HE400i load), S/N looks close to -126dbA but to be sure I need to arrange AP 555 and remeasure
    b) performance mode with excellent THD@1kHz@-3dbfs .0003% THD@1kHz@-10dbfs .00017%, however PLL not fast enough to lock data without loses at starting(up to half of second at the begin of the track is muted)
    c) SE emulation mode, the same as normal mode but 2nd harmonic domination at level .01% I think or any value up to .5% or so
    So far I want to use follow scenario to switch between modes but I opened for any opinions and ideas regarding that:
    1) Device powered On, MCU started and send initialization dump to the ES9038 for the mode determined variable MODE, which is = 0 that corresponds to the normal mode
    2) during the first 0..1000mS the MCU flashing by "Power" LED 4 times, if power will be interrupted that period, next power-up the MCU will start with MODE = 1 i.e. performance mode
    3) during next 1000..2000mS the MCU flashing by "Power" LED 8 times, and if power will be interrupted that period, next power-up the MCU will start with MODE = 2 i.e. SE mode
    4) if during first 4S power wasn't Off the device operating in normal mode, and generally looks like USB DAC+HPA with 4S blinking LED after turning it On.
    I understand that such an interface is kinda weird but I want to let try users all features available with no extra complexity with buttons. Any high-res players like Neutron, has a volume knob, so I believe dongle like DAC doesn't need any buttons at all. Any opinion?
    9038S_BOTTOM_VERTICAL_GLASS.jpg 2019-04-05_13-17-08.png 9038S_TOP_VERTICAL_WHITE.jpg
  2. StandsOnFeet
    That's an understatement.

    Are the three modes audibly different? DACs are supposed to be audibly transparent, so if those modes actually sound different in a blind test, then at least two of them aren't transparent. Why allow your users to choose among them? If they sound the same, then why allow your users to choose among them? If I were designing the device, I'd simplify my life and always go with one mode.

    Is there a way to tell which mode has been chosen?
  3. E1DA
    This is a stealth-interface, like gaming easter eggs. Regarding transparency and so on, due to my natural curiosity, I prefer to try by myself and decide what I'd like by myself too. I believe, most of potential users are similar to me, and they never like if someone will decide what is more transparent or less, curiosity is the dominant human instinct.
  4. bigshot
    There's no such thing as more transparent or less. Transparent is transparent. There can only be degrees of coloration.
  5. E1DA
    To me there's valid only psychological terminology, hence, no transparency, no coloration only differently described feelings. That's why I want to let a listener had the chance to test himself with my easter eggs :wink: To be honest, I can't hear the difference if THD .00015% or .05%(i.e. almost single 2nd harmonic there) but I don't consider myself as the best ears ever, and I believe if someone can. Another thing is a switching process, I mean from one mode to another one. If it not fast enough, like now when I should recompile MCU firmware to switch between modes, I simply can't remember the last sound signature precisely. A few seconds for the switch is ok, but <<10S as I feel. Much more easy to hear the difference related to the power supply(that affects dynamics, reshape the envelope of the signal) than THD levels or output impedance related difference, that's also quite easy(driver resonant less damped, and creates hump on frequency response, similar to EQ). However, I don't know why I can easy notice the difference with output impedances on my planars, which has perfectly linear imp. vs. freq. response.. 1.5ohm sounds more punchy vs 30mOhm.
  6. castleofargh Contributor
    does it mean that the extra options are only going to apply once, and next time we plug the device and leave it be it will revert to default setting?
  7. bigshot
    That's fine, but you'll have more variables to list, like what you had for lunch and whether you've been working too hard lately.

    Transparency usually means that in a controlled direct A/B switched line level matched blind comparison, two different sounds can't be distinguished from each other. That is a very objective thing. It has nothing to do with feelings or flowery language.
  8. E1DA
    does it mean that the extra options are only going to apply once, and next time we plug the device and leave it be it will revert to default setting?

    Yes, however, I like your idea better, thank you for the input! Indeed, If someone wants to use only SE mode, he'll have to use that stealth-interface every time, and finally, his USB-C connector will be broken, so I'll be damned )) I think, I could use only one LED flashing phase during startup and interrupt within that time will switch to the next mode next powerup, i.e. cycling Norm/Perf/SE/Norm/.. Only 2S LED's flashing instead of 4S, will reduce the probability of accidental modes switching. The drawback is only the Performance mode could be easily recognized(slow PLL lock mutes begin of tracks at Fs 44.1 & 48).
    Guys, another interesting question, how big 2nd harmonic I should add, homeopathic .01% or obvious .5-1% as real SE does?
  9. castleofargh Contributor
    someone is going to complain anyway, but I think that's better for the user, he'll run in circle a few times to try and figure out what is set up(and curse at you for not offering a clearer method), but once he's done he can relax and assume that his setting remains from then on. which is really all he will care about in the long run. unless of course you imagine that nobody will actually desire to remain on those settings, then having the proper option reverb back on its own is the right choice, or even better not having that cycle option at all. it's certainly tricky because by default I'm one to always favor more options when possible, but if those extra options have little interest yet end up affecting the normal use of the device, then it's not an option, it's a bother.
    while I'm in my role of worried "no happy" user, what are the chances for a computer boot to change the setting while the device is plugged? you know on some computers(or at least I've noticed it on some, maybe they all do? maybe it's just me ruining everything with my bios settings on my laptop? IDK) the motherboard would power everything, then cut power to stuff for a moment(sometimes even the fans) before resuming the boot process. I have no idea if it's a normal thing or how long that lasts, I'm a computer noob, but if that lasts a know amount of time, it could be good to have your device triggered with a different delay.

    no opinion on the harmonic thing, if it's not audible, then only a handful of people will dream about hearing a difference, and if it's audible it's going to be a subjective thing, just make sure it's sort of euphonic if you decide to add a lot, but I guess that's captain obvious talking.
    Last edited: Apr 15, 2019
  10. E1DA
    Haha, I begin worrying if the idea to curse me becomes popular )) Actually, I have a few points to avoid that prospect:
    a) As I know, nothing even close to my design performance does exist so far. I know some dongle USB DACs(Soundkey for example as I see there is about $4 BOM, $1 MCU, 1$ Ti's DAC, $1 Ti's HPA IC +surrounding parts) which are designed with the lowest cost priority, often there is not specified power, distortions, S/N(because it is the same as average smartphone with 3.5mm, i.e. 100.0% useless). Only the cost, size and color. My 9038S is the performance priority designed thing, and if it finally fitted into a dongle form-factor, I should thank electronics technology which offered me today 1000uF/6.3V just 2mm tall! Totally I put there 4000uF of capacitors that's perhaps 1000 times more than others USB dongle DAC+HPA.
    b) BOM of my beast is not cheap(one of the top DAC ES9038Q2M $8, AD8397x2 $6, CM6642 $2.2, those 1000uF $1.5 per each), however, It will be inexpensive after all due to no greedy brand or distributors. In fact, the project is DIY, I did that by myself and for myself but the place where I working last 12 years is very convenient for production(south of China) and electronics mass production is very well familiar to me. I can arrange 1KU production of 9038S and sell for $70/pcs or 10KU and $50. So, I guess I would be damned by lazy and greedy competitors rather than users, huh? :wink:
    Back to the interface topic, I started another DAC+HPA project on the SE9038Q2M DAC, it is bigger(I gonna use the same case as my PowerDAC V2 see attached pics pdac2.jpg pdac2_top.jpg ) and I have a chance to add there BLE( Bluetooth Low Energy) interface to control parameters by IOS/Android app. I have named that project 9038X, USB interface implemented on XMOS with 32/384 and DSD 11.2MHz. Less compromise, more expensive. I think BLE is an ideal interface, your smartphone is always in your pocket, and I need to prepare only 2 apps, one for IOS, another for Android. Maybe such a rich interface is too much for simple modes switching yet, I can add there volume control and filter types switcher as well. For the 9038S I simply have no room to add BLE interface, that's why I thinking about the easter-eggs interface.
  11. E1DA
    And finally I did that, SE mode got .25% of 2nd harmonic, and I can hear, or rather feel that as a more body in female vocals, like if a mic become closer to the singer, and a bit darker sibilants. Details are simplified and magnified at the same time pretty much in impressionists manner :wink:
  12. E1DA
    I've added visual differs to Performance and SE modes, now LED flashing 2 times/S if no audio data detected in the Performance mode, and 1 time/S in SE mode. There is video aboust it
    Also I want to show THD performance in all 3 modes(HFM HE400i load, 1kHz -10dbfs i.e. half of the windows volume slider), see attached screenshots. THD and THD+N tags need to divide by 100 due to -40db notch filter + 20db linear gain.
    2019-04-18_18-57-32.png 2019-04-18_18-56-11.png 2019-04-18_18-55-28.png

Share This Page