• The Vendor's Bazaar forum is for commercial offers and transactions. Only unmoderated members can post here.

    diyAudio provides this forum for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members. Use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

Reference DAC Module - Discrete R-2R Sign Magnitude 24 bit 384 KHz

Just tried it. Accessing the DAM1021 via tera term does not prevent the converter from showing the amplitude drift at 11025 Hz with the triangle wave changing to a "shaved top" triangle. It does however affect the current state of the triangle by setting it to a seemingly arbitrary value in the cycle.

I've attached scope readouts of the characteristic transition. The flatted out saw wave (5), not shot at the perfect position in my image, corresponds to the -3 dB below the initial level (which is given out correctly at the rest of the frequency spectrum).

It does not always cycle through from 1-8 (after which we are back to a clear triangle wave). Instead, it may also "reverse" from the low point (5), cycling through 4-1 backwards again. I've also managed to make it move the other way around altogether by logging into/exiting tera term.
 

Attachments

  • DAM1021_11025_cycle.jpg
    DAM1021_11025_cycle.jpg
    163.6 KB · Views: 349
Last edited:
I fed both recordings into DeltaWave and compared them to the unconverted original.

After processing with Deltawave the "bad" file nulls a little better than the "good" one. Phase response in the high end was certainly better in the "bad" file. Probably all related to drift rather than the 11025 triangle artifact.

What could be gleaned from the tests and anlysis is that drift is not correlated to the 11025 Hz sinwave issue. These are seperate phenomena.

It might all be a non-issue for purely auditioning purposes after all. Being unable to properly isolate the problems makes it hard to analyse.
 
And about the scope that you use to check for the converted 11025 Hz wave form - what is that scope sampling rate exactly? Even a toy $15 oscilloscope uses about 1 MHz (1 Msps) sampling rate. A better entry level scope is 1 Gsps. And that is exactly to avoid such artifacts.
 
It does not always cycle through from 1-8 (after which we are back to a clear triangle wave). Instead, it may also "reverse" from the low point (5), cycling through 4-1 backwards again. I've also managed to make it move the other way around altogether by logging into/exiting tera term.

Would you be able to superimpose the 8 graphs, and check if the zero crossings are aligned?

If the zero crossings are aligned, then it may not be the PLL driven clock shift, possibly something else contributing to the non-triangular shape.
 
And about the scope that you use to check for the converted 11025 Hz wave form - what is that scope sampling rate exactly? Even a toy $15 oscilloscope uses about 1 MHz (1 Msps) sampling rate. A better entry level scope is 1 Gsps. And that is exactly to avoid such artifacts.

This was a scope plugin analyzing the recording from the AD converter. So the sampling rate is 44,1 khz.

I've got an analog as well as a digital scope, but this artifact shouldn't happen in a DAAD chain.
 
This was a scope plugin analyzing the recording from the AD converter. So the sampling rate is 44,1 khz.

I've got an analog as well as a digital scope, but this artifact shouldn't happen in a DAAD chain.
This is the normal behavior of an ad converter. At 11025 Hz there are only 4 sample points and if dac and adc are not operated with the same master clock, there is drift. And by the way, the 11025 Hz triangle signal has a much higher frequency spectrum than can be reproduced with 44.1 kHz.