On the ESS website: https://www.esstech.com/wp-content/uploads/2024/09/ES9039Q2M_Datasheet_v0.2.0.pdf
From scrolling a bit through my albums on tidal the most seems to be in the 44.1kHz family
From scrolling a bit through my albums on tidal the most seems to be in the 44.1kHz family
When using the ES9039Q2M the Cal. Res. is 50k and Rext 37k. It is wrong in your shematic.
I updated the schematic and added some features:
- Headphone Amplifier based on TPA6120
- Headphone Amplifier Class A Amp, based on the "Noir" design
- microcontroller ESP32
- protection circuit with relay
- USB to IIS with Amanero board
Where are the pullup resistors for I2C bus. 4.7k is a reasonable value for moderate bus speeds. Also possible to use lower resistance if higher speeds are desired.
Also, you might consider allowing for a T/H resistor to ground for dac AVCC signals. The regulator is capable of putting out more current than you might need. In some cases it seems that experimental loading of LDO regulators with resistors to ground may allow for some tuning of effective regulator loop gain. May or may not be of any benefit in a given situation. Only way to know for sure is to try it at some point and see if any effect. Same type of thing is possible for clock regulators.
If using an Amanero for USB it may be useful to try running it on clean +5v power rather than USB power. I know of one commercial dac designer who recommends that modification. Also, when I tried it the modification did seem to help the sound.
Also, you might consider allowing for a T/H resistor to ground for dac AVCC signals. The regulator is capable of putting out more current than you might need. In some cases it seems that experimental loading of LDO regulators with resistors to ground may allow for some tuning of effective regulator loop gain. May or may not be of any benefit in a given situation. Only way to know for sure is to try it at some point and see if any effect. Same type of thing is possible for clock regulators.
If using an Amanero for USB it may be useful to try running it on clean +5v power rather than USB power. I know of one commercial dac designer who recommends that modification. Also, when I tried it the modification did seem to help the sound.
Last edited:
You have differential DAC output, why not to make a balanced amplifier ?
And here some about about ESS regulators
there is a new update pdf at the ESS website, all ASYNC modes up to 50Mhz
Last edited:
@Markw4
Thanks, I indeed forgot the I2C pullups (most sensor boards etc I worked with in the past had them included), I added them. Also some (not populated) resistors to GND for the regulators.
I also have heard that using another power supply for the amanero could be beneficial, for this I have to modify the amanero PCB as far as I know, will try that when it arrives.
@Anatolii_A
Great, in the new datasheet also the question from above with the S/PDIF input was solved! For the TPA6120 I use the differential output of the DAC, main purpose of this PCB should be the comparison between the TPA and the Class A output, than I will decide and probably optimize.
Thanks, I indeed forgot the I2C pullups (most sensor boards etc I worked with in the past had them included), I added them. Also some (not populated) resistors to GND for the regulators.
I also have heard that using another power supply for the amanero could be beneficial, for this I have to modify the amanero PCB as far as I know, will try that when it arrives.
@Anatolii_A
Great, in the new datasheet also the question from above with the S/PDIF input was solved! For the TPA6120 I use the differential output of the DAC, main purpose of this PCB should be the comparison between the TPA and the Class A output, than I will decide and probably optimize.
For the 9312 regulators: I see there is a lot more experience with LT3045 etc (and these are better for hand soldering), but the ESS ones are dedicated for tehir DACs and a lot cheaper - so why not test them?
I mean fully differential amplifier with two TPA6120 for balanced headphones connection
It can be modified by removing L1 near the USB connector then applying +5v to the non-USB side of L1 pads. The 5v ground should connect to the bottom side ground plane. May have to scrape off some solder mask at a good spot to solder.
Other option would be to modify a USB cable to break the +5v power connection from the PC. Bring in your own +5v power on the dac side of the cable. Ground the 5v supply to USB ground in the cable.
Datasheet specs are for 24MHz version. In theory 50MHz version should around 6dB higher phase noise. The clock is rated down to 10Hz offset. 50MHz might be down to -95dBc at that offset. No data for 1Hz offset. Not bad for $4.33. That said, I wouldn't necessarily expect the sound to be as good as might be possible, given the sensitivity of some dac architectures.
Does frequency stability of oscillator matters a lot for DAC ? There are a lot of very stable XOs (1-10ppm) with not so good phase noise specs
Stability or accuracy? Accuracy in audio means nothing, even +/-10000 ppm could be used if such with low jitter existed.
(10000 ppm = 1%, the human ear hears deviations in pitch greater than 1.5%)
Alex.
Thank you ! Then, the optimal variant to try with lowest non audio frequency will be KC7050K27 27Mhz, just for 1.25.
