Has it ever occurred to you that some of these effects, artifacts, deviations or errors or whatever you choose to call them are unique to you ?
By all means ask all the questions you need but be aware that for a project to make it to fruition it has to driven to fruition with a large degree of single-mindedness more easily found in an individual than in a group. The successful projects are when an individual creates something and puts it out there effectively saying " Here is my widget it pleases me, try it and it might please you." It was certainly that way before the internet. You bought a magazine and it had projects in it and if you liked one you built it.
Last edited:
Fine. So who has referred to whom else as being "dead wrong?" Not me. In fact, every technical claim I made was technically correct to the best of my knowledge. I have also attempted to be respectful to people who disagree. Is your critical speech directed at bohrok, or to whom? What about yourself? Are you right without question?
Last edited:
I meant 'wrongly" in your objective sense, since I was speaking to you. You seem to feel that human perception can be objective. I am doubtful as to the extent human perception can be objective. To some extent, yes. Survival or the lack thereof determines that. If survival is not at stake then people can and will believe what they want. You too.
You are still dead wrong about your claim that WASAPI Exclusive does not use Windows UAC2 driver. Sorry, but your claim is technically incorrect. No need to get jumpy when somebody corrects you.
As you probably have already noticed this forum is not a place for community projects 😉
Just in case you think this is something where we just disagree read the following (should take only 2 minutes):
https://docs.microsoft.com/en-us/windows-hardware/drivers/audio/windows-audio-architecture
(hint: USB Audio driver in picture is the Windows UAC2 driver)
ASIO bypasses most (if not all) layers of Windows Audio Architecture so ASIO does not use Windows UAC2. And it is probably possible to create another Windows USB audio driver that could be used from WASAPI Exclusive but I'm not aware of such.
The practical effect is that mere UAC2 compliance does not guarantee access to WASAPI Exclusive Mode operation. Thesyscon gets paid extra if you want that feature. As you already described, without a WASAPI Exclusive Mode Driver you have set the default sample rate and bit depth every time a song in a playlist is different from the last song in either one of those regards. Another workaround you suggested was resample everything in Foobar2000 so that Windows won't resample it. Yet you keep insisting that mere UAC2 compliance is fine for serious hi-fi. Its not fine, its garbage.
Regarding ASIO, there is apparently still an entry point for Windows to access a PCM ASIO stream and resample it. It happens, and it has happened multiple times when Windows arbitrarily reassigns an ASIO device to "Default Sound Device" and or "Default Communication Device."
Only WASAPI Exclusive Mode drivers can solve those problems once and for all.
Regarding ASIO, there is apparently still an entry point for Windows to access a PCM ASIO stream and resample it. It happens, and it has happened multiple times when Windows arbitrarily reassigns an ASIO device to "Default Sound Device" and or "Default Communication Device."
Only WASAPI Exclusive Mode drivers can solve those problems once and for all.
Where have I said that? You seem to have serious trouble understanding the issue. Windows UAC2 is a USB audio device driver only, nothing else. As everything USB audio related goes through UAC2 driver it is fine for serious hifi.
Read this. It takes only 11 minutes.
https://docs.microsoft.com/en-us/windows-hardware/drivers/audio/usb-2-0-audio-drivers
The only issue is whether or not the audio stream gets corrupted or if corruption can be prevented. That's the issue for serious hi-fi, anyway.
Add fast start-up to your LT3042 regulator. I would move the regulator much closer to the dac chip. Maybe rotate it 90 degrees clockwise and flip over to the other side of dac chip (right).
Ty for the reply, I was thinking if I need the fast startup, why is it useful?
Is the LDO on the right side better placed because it's further away from the digital area of the IC ?
Is the LDO on the right side better placed because it's further away from the digital area of the IC ?
Fast startup does not impact LDO performance and it just adds 2 resistors. For 3V3 it may not be necessary but for higher voltages the output rampup may take seconds. I prefer to use the same schema and layout for all voltages.
The location may not be important but LDO needs to be closer. As there is not much experience of ES9033 it is hard to tell which power pins are most critical. I would place the regulator closest to the critical power pin.
The location may not be important but LDO needs to be closer. As there is not much experience of ES9033 it is hard to tell which power pins are most critical. I would place the regulator closest to the critical power pin.
Would you share your implementation with me? seems like you got some experience with that LDO
Fast startup is a definite requirement. I once had a LT3042 based 5V PSU without it and it took many seconds for it to ramp up. That caused strange behavior of the DAC and I think the reset circuit failed to work OK.
Implementation is datasheet wisdom. Just 2 resistors, no penalty.
Implementation is datasheet wisdom. Just 2 resistors, no penalty.
Maybe look at the ES9038Q2M evaluation board schematic I posted earlier. As can be seen, AVCC gets additional filtering and or buffering since it is a critical voltage reference for the dac output. In other words, the audio output of the dac is multiplied by the AVCC voltage. Any noise or distortion on the AVCC rail gets multiplied into your audio. Hopefully use linear caps for AVCC bypass and or filtering.
Another critical refernce signal is MCLK (either one or two master clocks, depending on implementation). Ideally clocks should be close to the dac chip and copies of MCLK needed by other equipment (e.g. a USB board) should be buffered by something like NB3L553-D. Both clocks and buffers benefit from clean dedicated power and linear type bypass caps. Bypass cap damping resistors if needed should also be linear types such as thin film. My opinion and the opinion of others such as @diyiggy. Other people may disagree, particularly people who haven't tried it.
Also, if using LT304x regulators, the Cset cap should be a linear type, not a class-2 ceramic which is piezoelectric and converts mechanical vibration to voltage noise (among other LT304x implementation considerations). Agree with others that sometimes fast startup can be necessary as some load circuits are sensitive to power supply rise time.
Regarding mode jumpers, if the jumper is removed when power is on then you have a floating input. Better to tie an input pin to Vcc or to ground through a 10k resistor, then allow for a jumper that can connect the pin to the opposite rail which will override the resistor programming. That way a missing jumper will not result in a floating input.
Another critical refernce signal is MCLK (either one or two master clocks, depending on implementation). Ideally clocks should be close to the dac chip and copies of MCLK needed by other equipment (e.g. a USB board) should be buffered by something like NB3L553-D. Both clocks and buffers benefit from clean dedicated power and linear type bypass caps. Bypass cap damping resistors if needed should also be linear types such as thin film. My opinion and the opinion of others such as @diyiggy. Other people may disagree, particularly people who haven't tried it.
Also, if using LT304x regulators, the Cset cap should be a linear type, not a class-2 ceramic which is piezoelectric and converts mechanical vibration to voltage noise (among other LT304x implementation considerations). Agree with others that sometimes fast startup can be necessary as some load circuits are sensitive to power supply rise time.
Regarding mode jumpers, if the jumper is removed when power is on then you have a floating input. Better to tie an input pin to Vcc or to ground through a 10k resistor, then allow for a jumper that can connect the pin to the opposite rail which will override the resistor programming. That way a missing jumper will not result in a floating input.
Last edited: