Moved to DLL
I would say go with a 9028, since AVCC is much easier to drive well. Also OPA1611 for output stage.
Personally happen to like LT1763 for the digital and clock use, at least if run at a sufficiently high current. To keep them cool, best to limit input power to 5v in that case.
LT1963 is one I don't like much for anything. For opamps rails dual LT1083 is fine if used with an R-core transformer with separate windings (no center tap windings). LT1083 Adjustable HIFI Linear Regulated Power Supply Board Dual Output | eBay ...Then tie regulator grounds at the load only.
EDIT: AVCC is the trickiest power supply to figure out. For 9028 it should be doable with a couple of high current opamp buffers. The opamps may need some big-ish power rail filter caps near them to more easily deliver bursts of current for good bass. Opamp outputs should be close to AVCC inputs. The idea is to keep trace/wire resistances, and ground path resistance as low as possible to help minimize distortions (linear and or nonlinear). Other good options might include discrete regulators, maybe Nazar type shunts, or some other choices. IME and IMHO probably best to avoid LDOs for that, up to you though.
Personally happen to like LT1763 for the digital and clock use, at least if run at a sufficiently high current. To keep them cool, best to limit input power to 5v in that case.
LT1963 is one I don't like much for anything. For opamps rails dual LT1083 is fine if used with an R-core transformer with separate windings (no center tap windings). LT1083 Adjustable HIFI Linear Regulated Power Supply Board Dual Output | eBay ...Then tie regulator grounds at the load only.
EDIT: AVCC is the trickiest power supply to figure out. For 9028 it should be doable with a couple of high current opamp buffers. The opamps may need some big-ish power rail filter caps near them to more easily deliver bursts of current for good bass. Opamp outputs should be close to AVCC inputs. The idea is to keep trace/wire resistances, and ground path resistance as low as possible to help minimize distortions (linear and or nonlinear). Other good options might include discrete regulators, maybe Nazar type shunts, or some other choices. IME and IMHO probably best to avoid LDOs for that, up to you though.
Last edited:
if these are your first experiments with ES90X8 DACs, then ES9028 ES9028PRO /ES9018 ES9018S Q8 HiFi Audio DAC Decoder DIY KIT | eBay kit might be the best bang for the buck to start with, IMHO
Yes, with sabre this is first experience but I was make about five dac's (tda1541, ak4493, pcm1794).
I am already have power transformers, lt3042 power boards and amanero. I need only dac board.
Also want to ask you about 1.2v 9028 power supply: they are analog and digital, necessary they must be separately or they all can be connected to one stabilizer?
I am already have power transformers, lt3042 power boards and amanero. I need only dac board.
Also want to ask you about 1.2v 9028 power supply: they are analog and digital, necessary they must be separately or they all can be connected to one stabilizer?
IIRC the 1.2v pins (or 1.3v, for high sample rate DSD at high MCLK frequencies) are all tied together inside the chip, and ESS evaluation boards only use one voltage regulator for them.
IME, LT3042 makes a rather unsatisfactory AVCC regulator. Especially the cheap Chinese versions. The cheap class II ceramic caps are nonlinear and somewhat piezoelectric. Those LT3042 boards are good for the digital stuff though. Even when well implemented I would have doubts about the sound of the LT3042 error amplifier, which if it has any sound at all will be audible at the dac analog outputs. Much better to opamp buffers to drive AVCC, or maybe a Jung regulator. Something that is known to sound good for analog audio, could be shunt too.
Also, Amanero should be a real one, not a Chinese clone. With the isolated power supply mod it sounds *much* better. Or, just go with I2SoverUSB which is cheaper and better.
For +-15v opamp rails, one can get quite good hi fidelity sound using a dual LT1083 board along with an R-core transformer having a separate windings for each rail (center tapped windings won't work for that). LT1083 HIFI Linear Power Supply Dual Output High Adjustable Power Supply Board | eBay
The way to wire them up to the dac board is explained in another thread: ES9038Q2M Board ...Posts #7158 and #7162
Also, Amanero should be a real one, not a Chinese clone. With the isolated power supply mod it sounds *much* better. Or, just go with I2SoverUSB which is cheaper and better.
For +-15v opamp rails, one can get quite good hi fidelity sound using a dual LT1083 board along with an R-core transformer having a separate windings for each rail (center tapped windings won't work for that). LT1083 HIFI Linear Power Supply Dual Output High Adjustable Power Supply Board | eBay
The way to wire them up to the dac board is explained in another thread: ES9038Q2M Board ...Posts #7158 and #7162
Last edited:
What a pile of rehashed unsubstantiated opinions. I know asking you to substantiate the "sound good/bad" argument is a futile exercise, you have zero proof of that, but at least from an engineering perspective:
- How are you proposing to supply >120mA from an op amp buffer to the ES9038PRO AVCC?
- In general, using any DAC board requiring external power regulators is a clear invitation to troubles, that's more critical than any opinions about the quality of a certain regulator topology.
- I know you are following my DAC thread, didn't you note it uses exclusively non linear and piezoelectric 0603 ceramic caps for decoupling (essentially all caps except those involved in the output filter, which are all C0G) and the performance is at the limit of what can be measured today. A slap in the face of those claiming that ceramic caps are not good for decoupling, mind you.
- Amananeros are digital devices, what would be the shortcoming of Chinese clones, assuming they use the same parts as the original (which they do, or even better, mind you). There are Chinese crap clones for everything, the right advice would be check it before you buy, not avoid.
- What's the problem with center taped transformers, do they "sound bad" even for passing the result through a >60dB line regulation and >100dB op amp PSRR? And you have the nerve to suggest an off DAC board for this power supply, do you realize that the required wiring, whatever way you are doing it, is going to be the limiting performance factor?
Last edited:
I think an opamp buffer might work for ES9028, not for ES9038. Was that not clear?
Regarding Amanero, the Chinese were able to use early firmware made available by Amanero for its customers. Amanero later changed the firmware so the Chinese units would be bricked if a firmware upgrade was attempted. Thus the clones are not able to use Amanero ASIO drivers.
Center tapped doesn't work when using positive regulators for both the + and - 15v rails since one regulator has to be used 'upside down.'
Many people have tried non class II ceramics for certain subsystems in dacs. It only makes a difference for listening, not for FFT measurements since it probably affects something like noise modulation. One can deny it of course, but people can tell the difference if they listen.
Regarding Amanero, the Chinese were able to use early firmware made available by Amanero for its customers. Amanero later changed the firmware so the Chinese units would be bricked if a firmware upgrade was attempted. Thus the clones are not able to use Amanero ASIO drivers.
Center tapped doesn't work when using positive regulators for both the + and - 15v rails since one regulator has to be used 'upside down.'
Many people have tried non class II ceramics for certain subsystems in dacs. It only makes a difference for listening, not for FFT measurements since it probably affects something like noise modulation. One can deny it of course, but people can tell the difference if they listen.
Last edited:
You pulled again the "noise modulation" argument out of your hat; you were asked several times, by several competent members, what is your understanding of "noise modulation" in this context and you were always unable to produce a rational response. Diagnosis: a made up argument, intended to impress the less knowledgeable, AKA a "red herring".
You claimed:
and NOT an usability/functional issue with the Chinese clones.
Last edited:
I wonder if I'm the only guy who bothered to measure how the power supply current of ES9018 varies depending on signal in order to design a proper power supply for it
