Hello, first off, thanks for submitting this great project and build files.
I have built this (and using the initial bit file as in post #1) and am using it together with the fab RTZ DAC, with a Squeezebox receiver as source. It sounds great, however the output seems to be lower than the same files played using an Amanero clone, which is odd.
I wanted to ask what format it is expecting Left Justified or Philips format with the 1 bit offset? (Sorry if it's been answered on the thread but it's 30 pages and I couldn't spot it)
I've had an issue in the past when interfacing with Logitech Transporter in that they used left justified format and I needed Philips I2S since I was feeding a TDA1541A. ie the MSB was lost on the TDA1541A since it waits for 1 x BCLK after the WCLK transition. On that one I had to add a delay to the data to get bit 15 back.
I fairly sure the SB receiver is outputting Philips version with the 1 bit MSB delay since the Wolfson WM8501 only supports that and right justified and the format pin is definitely set to the left justified mode.
Many thanks,
Tony
I have built this (and using the initial bit file as in post #1) and am using it together with the fab RTZ DAC, with a Squeezebox receiver as source. It sounds great, however the output seems to be lower than the same files played using an Amanero clone, which is odd.
I wanted to ask what format it is expecting Left Justified or Philips format with the 1 bit offset? (Sorry if it's been answered on the thread but it's 30 pages and I couldn't spot it)
I've had an issue in the past when interfacing with Logitech Transporter in that they used left justified format and I needed Philips I2S since I was feeding a TDA1541A. ie the MSB was lost on the TDA1541A since it waits for 1 x BCLK after the WCLK transition. On that one I had to add a delay to the data to get bit 15 back.
I fairly sure the SB receiver is outputting Philips version with the 1 bit MSB delay since the Wolfson WM8501 only supports that and right justified and the format pin is definitely set to the left justified mode.
Many thanks,
Tony
Hi Tony,
Not sure if you are comparing PCM to DSD, or if you have two different sources of DSD that have different levels. In the former case it is normal for DSD to be lower level that the equivalent PCM. In the latter case how can it be that the 'files' be the same if you are comparing PCM->DSD converter output to a preexisting DSD file?
Mark
Not sure if you are comparing PCM to DSD, or if you have two different sources of DSD that have different levels. In the former case it is normal for DSD to be lower level that the equivalent PCM. In the latter case how can it be that the 'files' be the same if you are comparing PCM->DSD converter output to a preexisting DSD file?
Mark
Hi Mark,
No, I was comparing the PCM output from Amanero into the PCM2DSD and the PCM output from the SB receiver playing the same file. I checked with a scope so I know it was definitely PCM coming out of the Amanero rather than DSD. The Amanero is the source that is louder.
I've had this same thing before on the Transporter and in that case I had to use a couple of D flip-flops and inverters to delay the data line by one BCLK cycle and that fixed the issue.
If the PCM2DSD is using left-justified without the 1 bit offset to the MSB then effectively it'll think my MSB is always zero and that my MSB is in fact the next lower bit, hence lower output. I just wanted to have it confirmed as my scope is playing up and I can't get it to trigger properly on the word clock so that I can compare the waveforms of the Amanero and the Squeezebox.
Thanks,
Tony
No, I was comparing the PCM output from Amanero into the PCM2DSD and the PCM output from the SB receiver playing the same file. I checked with a scope so I know it was definitely PCM coming out of the Amanero rather than DSD. The Amanero is the source that is louder.
I've had this same thing before on the Transporter and in that case I had to use a couple of D flip-flops and inverters to delay the data line by one BCLK cycle and that fixed the issue.
If the PCM2DSD is using left-justified without the 1 bit offset to the MSB then effectively it'll think my MSB is always zero and that my MSB is in fact the next lower bit, hence lower output. I just wanted to have it confirmed as my scope is playing up and I can't get it to trigger properly on the word clock so that I can compare the waveforms of the Amanero and the Squeezebox.
Thanks,
Tony
@PJotr25 Thanks for confirming that for me. I can be certain now that there is no format mismatch between the SB Receiver output and the PCM2DSD.
@Markw4
Sorry Mark, I probably didn't describe my tests very well. I was testing two "I2S" sources to the PCM2DSD board, one from Amanero and the other from the SB Receiver (tapped from it's CPLD outputs). The Amanero was connected to PC playing FLAC file using FooBar2000. The SB Receiver playing the same FLAC file but via the Logitech Media Server. Replay gain adjustments were off. So in theory the output level from the DAC should've been the same.
However the Amanero was a lot louder.
I'd assumed that the difference was due to the I2S format from the SB Receiver not matching that expected by the PCM2DSD. However, now it is confirmed that the PCM2DSD is expecting the Philips variant with the 1 BCLK delay in the MSB after word clock then I can discount that being an issue.
So the only other possibility is that the Amanero is outputting left-justified instead of Philips. That would presumably mean that (for 16bit) the MSB bit is shifted 1 left and the bit 14 becomes the MSB. This would certainly increase the output would it not? I guess it's irrelevant now anyway as my plan was always to use the SB Receiver, the Amanero was just a means to prototype and test the RTZ DAC.
I will do more tests to try to calibrate the DAC so that I can be sure I've not made some error in the build that could reduce the gain/output.
@Markw4
Sorry Mark, I probably didn't describe my tests very well. I was testing two "I2S" sources to the PCM2DSD board, one from Amanero and the other from the SB Receiver (tapped from it's CPLD outputs). The Amanero was connected to PC playing FLAC file using FooBar2000. The SB Receiver playing the same FLAC file but via the Logitech Media Server. Replay gain adjustments were off. So in theory the output level from the DAC should've been the same.
However the Amanero was a lot louder.
I'd assumed that the difference was due to the I2S format from the SB Receiver not matching that expected by the PCM2DSD. However, now it is confirmed that the PCM2DSD is expecting the Philips variant with the 1 BCLK delay in the MSB after word clock then I can discount that being an issue.
So the only other possibility is that the Amanero is outputting left-justified instead of Philips. That would presumably mean that (for 16bit) the MSB bit is shifted 1 left and the bit 14 becomes the MSB. This would certainly increase the output would it not? I guess it's irrelevant now anyway as my plan was always to use the SB Receiver, the Amanero was just a means to prototype and test the RTZ DAC.
I will do more tests to try to calibrate the DAC so that I can be sure I've not made some error in the build that could reduce the gain/output.
@Cestrian,
IIRC you can check and or set the Amanero output format using their maintenance tool. Their main download page is at: https://amanero.com/combo384_firmware.htm
Also, application notes and various other things can be found at: https://amanero.com/drivers/
IIRC you can check and or set the Amanero output format using their maintenance tool. Their main download page is at: https://amanero.com/combo384_firmware.htm
Also, application notes and various other things can be found at: https://amanero.com/drivers/
Just found a problem with Simple DSD Converter: When Native DSD is being played from a DSD file, then when there is a clock family change the dac produces a loud pop. However if Simple DSD converter is removed from the signal chain, and Native DSD is played from a file using only the USB board, then no pop noise occurs during clock family changes.