@jpk73 Do you think it will benefit from the pulse silence at all? Because without the Latch there is nothing on the output 
With stop-clock2 the clock is reduced three times, but it is still present at the same voltage level. It's different from reducing noise on an analog signal, where the noise is put to a lower level 🙂
With stop-clock2 the clock is reduced three times, but it is still present at the same voltage level. It's different from reducing noise on an analog signal, where the noise is put to a lower level 🙂
Hello,
I have used miro's AD1862 board 1 year now, I love it much. Thank You Miro again. Using I2Soverusb board with it to skip "shift registers" and it sounds good to myself.
planning to mod the I2soverusb board by removal of the oscillator and replace with better oscillator (e.g. neutron star from Newclassd).
May I ask kind people here who can share experience about it? Is that directly unplug the on-board oscillator and wire another clock?
Thank You
I have used miro's AD1862 board 1 year now, I love it much. Thank You Miro again. Using I2Soverusb board with it to skip "shift registers" and it sounds good to myself.
planning to mod the I2soverusb board by removal of the oscillator and replace with better oscillator (e.g. neutron star from Newclassd).
May I ask kind people here who can share experience about it? Is that directly unplug the on-board oscillator and wire another clock?
Thank You
Question for you guys who have tried a few different opamps and found them to sound different with the dac in this thread: Which, if any of them, sounds the most like the tape and or vinyl version of the same music? In other words, is any opamp producing close to be the real intended sound? How close is that?
Reason I ask is I am still working with AK4499, although maybe I will start on another dac before too long. So far the best I can do is not really superior to vinyl. Its pretty close, but not better in every single way. In other words, the theoretical promise of digital sound quality remains somewhat elusive in practice.
Measurement of course is another matter. The way we measure with an AP is what digital has been optimized for. No question digital is better if AP numbers were to be the ultimate goal. Don't see it as the goal myself.
Reason I ask is I am still working with AK4499, although maybe I will start on another dac before too long. So far the best I can do is not really superior to vinyl. Its pretty close, but not better in every single way. In other words, the theoretical promise of digital sound quality remains somewhat elusive in practice.
Measurement of course is another matter. The way we measure with an AP is what digital has been optimized for. No question digital is better if AP numbers were to be the ultimate goal. Don't see it as the goal myself.
You might want to look into the JLsound OEM board: smaller, more beautiful, more flexible, and free clock choice...
@jpk73 Do you think it will benefit from the pulse silence at all? Because without the Latch there is nothing on the output
With stop-clock2 the clock is reduced three times, but it is still present at the same voltage level. It's different from reducing noise on an analog signal, where the noise is put to a lower level 🙂
@miro1360, theoretically the moment the analog output of the DAC is set could benefit a lot from quiet digital lines. After all it's basically a transfer from serial to parallel with the output resistor ladder connected.
Many digital filters output stopped clock already so you don't see a lot stopped clock circuits in commercial products. The PMD200 does not output stopped clock, but the Naim ist the only PMD200 schematic I saw with a stopped clock circuit.
So yes, I am certain that it helps to get a more true analog output from the DAC.
BTW what is the deglitch output signal for, that some digital filters offer?
EDIT: just found a good explanation regarding the last question here.
Attachments
Last edited:
The normal board has the same feature (external Masterclock) but also Galvanic isolation ... and has NDK NZ2520SDA onboard ...
Thank you JPK73 and TJF for the hints,
There is an instruction on the I2Soverusb manual for using external master clock and its jumper setting. The manual mention the wiring of external master clock on H3.3 and its GND.
I cannot figure out why there is only one external master clock input header-H3.3, but normally there are 2 master clock frequencies (e.g. 45Mhz and 49 Mhz)
So the board can only use one master clock frequency if using external one?
Thank You
There is an instruction on the I2Soverusb manual for using external master clock and its jumper setting. The manual mention the wiring of external master clock on H3.3 and its GND.
I cannot figure out why there is only one external master clock input header-H3.3, but normally there are 2 master clock frequencies (e.g. 45Mhz and 49 Mhz)
So the board can only use one master clock frequency if using external one?
Thank You
Because the master clock input had been designed to work together with the JLS external oscillator board, which is pin compatible and executes the switching between the two, onboard Xtal oscillators.. (switch signal received from the appropriate output of the JLS 'motherboard', and gets galvanically isolated..
The JLS external clock board (that can be used with I2SoverUSB) is something I have some familiarity with. Sorry, although I do strongly recommend I2SoverUSB, I do not recommend the external clock board. Why? Its design is contrary in some ways to some things at least some people in the forum are staring to recognize. For instance, there is the issue of use of the film caps recommended by diyiggy for bypass instead of using X7R. There is the use of ferrites in series with clock power. There is failure to leave both clocks running at all times. There is uncertain optimization of clock power regulation. No clock output buffering...and so on. Time to consider a redesign IMHO. Sorry if its not what people want to hear. Maybe consider I2SoverUSB has gone through a few design iterations to make it what it is today. The external oscillator doesn't seem to have gotten the same attention.
Mark, I had described the reasoning behind the single external MCLK input.. was not expressing opinion on the modul. Yes, You are right in some points. Though the oscillators are Not continously ON, together. They are powered ON, but the En(able) pin is used to turn them off.
@Joseph K, Agree they are disabled when not in use. IME it is about as bad as powering them off in terms of settling time after being re-enabled.
@jpk73, I do think the internal I2SoverUSB clocks are very good. However I want better for my own use. Their external clock board didn't solve my clock quality goals. Had to design my own board with clocks, DSP, ASRC, etc., all the support functions to better assess the potential of a dac chip eval board. My board went through a couple of PCB iterations and a number of experiments were performed on each iteration. Jumpers and other features were designed in so as to better facilitate certain experiments. I can't share the design, but I have given a fair amount of hints about what sorts of things to consider. One feature I didn't mention that was built in was a port to connect another external clock. Experiments with external clocks included use of Andrea Mori clocks, which were the best. However I did find that a carefully optimized Crystek 957 design could work almost as well with my particular dac chip. In the end I decided there were other remaining issues with more audible effect than the Crystek implementation.
@jpk73, I do think the internal I2SoverUSB clocks are very good. However I want better for my own use. Their external clock board didn't solve my clock quality goals. Had to design my own board with clocks, DSP, ASRC, etc., all the support functions to better assess the potential of a dac chip eval board. My board went through a couple of PCB iterations and a number of experiments were performed on each iteration. Jumpers and other features were designed in so as to better facilitate certain experiments. I can't share the design, but I have given a fair amount of hints about what sorts of things to consider. One feature I didn't mention that was built in was a port to connect another external clock. Experiments with external clocks included use of Andrea Mori clocks, which were the best. However I did find that a carefully optimized Crystek 957 design could work almost as well with my particular dac chip. In the end I decided there were other remaining issues with more audible effect than the Crystek implementation.
Last edited:
Real intended sound is whatever the recording / mastering engineer produced. If any recording / mastering engineer produce music album to sound like tape and or vinyl version instead of live sound, run away. 🙄
Although stopping clocks etc., can have clear advantages it is questioned if this is the best approach, rather than converting a time multiplexed input to simultaneous mode as something that Audial and JLSounds does to implement the TDA1541a.
From the link to the Audial website below, in simultaneous mode "both master clock and bit clock, are twice lower at given sampling frequency". The conclusion being that a time multiplexed data stream with clocks stopped during its off time has equal clock cycles to a DAC operated in continuous mode at twice lower frequency.
In other words the simplest approach would be for JLSounds to create simultaneous mode signals for other DACS. The advantages of conversion to simultaneous mode (as Audial suggests happens below) could exceed those of stopping clocks
TDA1541A and Model S USB, part 2: Simultaneous data mode - Audial
From the link to the Audial website below, in simultaneous mode "both master clock and bit clock, are twice lower at given sampling frequency". The conclusion being that a time multiplexed data stream with clocks stopped during its off time has equal clock cycles to a DAC operated in continuous mode at twice lower frequency.
In other words the simplest approach would be for JLSounds to create simultaneous mode signals for other DACS. The advantages of conversion to simultaneous mode (as Audial suggests happens below) could exceed those of stopping clocks
TDA1541A and Model S USB, part 2: Simultaneous data mode - Audial
So could this be one of the biggest secrets in high end audio, the use of dual DAC's in simultaneous mode?
These ideas are similar in principle as I see it: the TDA1541A diagram for simultaneous mode shows the same pulse silence during analog update. Running the clocks at half the frequency also would reduce possible noise on digital lines during that moment. As does stopped clock operation. Nevertheless it would be interesting to see if the TDA is as clean with stopped clock as in simultaneous mode...
Last edited:
TDA1541A is cleaner in simultaneous mode and it doesn't have to be just because of the stopped BCK. The main reason can be, that the signal in simultaneous mode requires almost no further processing and can be used directly for R-2R buffer. Where on the other hand the I2S format must be digitally processed inside the chip and all this processing can easily contaminate the analogue part 🙂
@jpk73 ... the glitch depends from DAC to DAC. AD1862 has negligible glitch, but such a PCM53 may be significantly worse on it 😀
I have tried LM6171, AD797, ADA4627-1, LME49720HA (metal can), OPA1611A, NE5534 and Sparkolab SS3601. Of these I like the ss3601 and AD797 the best. I believe they are closest to the real sound. By real sound, I mean the sound I hear in real life with my bare ears. Sound that physically goes straight to my ears without having been recorded first by any audio devices.
I am using a headphone, so there are no room reflections to interfere with what I hear from the dac.
Incidentally, they also measure the best. The OPA1611A, though theoretically should measure the best, does not measure as well in this dac. Even LM6171 and ADA4627-1 can beat it.
Cheers!