Hi All,
small update on the progress over here.
After changing the op-amp to OPA2211 and removing the caps everything works - more or less.
2 Issues though:
I still get a faint hiss, similar to the one before changing the op-amp, the level is much lower though. Should I try to tame it fitting those caps back or would anyone suggest another path to follow ?
Second, more annoying issue:
When hooking up an amplifier with DC blocking cap everything works fine.
When I try to put a DC blocking cap (4,7uF MKP) ahead of my R/C lowpass in line with the signal, I can't get any output.
The RC is a 1K resistor in series and 3300pF to ground.
Any thoughts on that matter ?
Cheers,
Max
small update on the progress over here.
After changing the op-amp to OPA2211 and removing the caps everything works - more or less.
2 Issues though:
I still get a faint hiss, similar to the one before changing the op-amp, the level is much lower though. Should I try to tame it fitting those caps back or would anyone suggest another path to follow ?
Second, more annoying issue:
When hooking up an amplifier with DC blocking cap everything works fine.
When I try to put a DC blocking cap (4,7uF MKP) ahead of my R/C lowpass in line with the signal, I can't get any output.
The RC is a 1K resistor in series and 3300pF to ground.
Any thoughts on that matter ?
Cheers,
Max
Some update,
in finally I managed to get the DAC working byloading the output (positive to ground) with a 10K resistor.
Then I have my 1K/3300pF LP filter and finally the 4,7uF DC blocking cap, which I will skip later as I have one in each amp I am using.
This is until now the only way to get rid of the high pitched noise on the output, turn on/off thump has been reduced to acceptale levels as well.
I admit than 10K is a lot, I am loosing to much output level, but at least it works.
What's the minimun load I can apply without getting into current out mode ?
in finally I managed to get the DAC working byloading the output (positive to ground) with a 10K resistor.
Then I have my 1K/3300pF LP filter and finally the 4,7uF DC blocking cap, which I will skip later as I have one in each amp I am using.
This is until now the only way to get rid of the high pitched noise on the output, turn on/off thump has been reduced to acceptale levels as well.
I admit than 10K is a lot, I am loosing to much output level, but at least it works.
What's the minimun load I can apply without getting into current out mode ?
Has someone noticed huge noise after the filter rolloff frequency with I2S source?
http://www.diyaudio.com/forums/twis...embedded-audio-applicance-13.html#post3859330
I might have something incorrectly configured, but the I2S timing seems to be alright:
http://www.diyaudio.com/forums/twis...embedded-audio-applicance-13.html#post3860103
http://www.diyaudio.com/forums/twis...embedded-audio-applicance-13.html#post3859330
I might have something incorrectly configured, but the I2S timing seems to be alright:
http://www.diyaudio.com/forums/twis...embedded-audio-applicance-13.html#post3860103
Notice: The I2S source is BeagleBone Black with internal low quality oscillator and the Linux with sound driver I've wrote -- http://bbb.ieero.com/ ... so there might be problem with my driver also.
I've been doing more research on the roll off noise issue with ES9018 described above.
Recent findings:
- the white noise in SoX->BBB->I2S->Buffalo is louder by 6dB than ARTA->PC->SPDIF->Buffalo
- when playing via I2S with volume reduced by 6dB the roll off noise is actually better than with SPDIF
- white noise with volume reduced by less than 6dB has roll off issues
$ play -r 48000 -c 1 -b 24 -n synth white
The stats of "full volume" white noise generated by SoX:
DC offset 0.000745
Min level -0.999998
Max level 0.999999
Pk lev dB -0.00
RMS lev dB -4.77
RMS Pk dB -4.58
RMS Tr dB -4.98
Crest factor 1.73
Flat factor 0.00
play -r 48000 -c 1 -b 24 -n synth white vol -6dB
The stats of "-6dB" white noise generated by SoX:
DC offset -0.000290
Min level -0.501187
Max level 0.501187
Pk lev dB -6.00
RMS lev dB -10.77
RMS Pk dB -10.55
RMS Tr dB -10.99
Crest factor 1.73
Flat factor 0.00
From the stats I think that full volume white noise should not cause the issues. Or am I wrong and the volume must be really reduced?
I've been doing more research on the roll off noise issue with ES9018 described above.
Recent findings:
- the white noise in SoX->BBB->I2S->Buffalo is louder by 6dB than ARTA->PC->SPDIF->Buffalo
- when playing via I2S with volume reduced by 6dB the roll off noise is actually better than with SPDIF
- white noise with volume reduced by less than 6dB has roll off issues
$ play -r 48000 -c 1 -b 24 -n synth white
The stats of "full volume" white noise generated by SoX:
DC offset 0.000745
Min level -0.999998
Max level 0.999999
Pk lev dB -0.00
RMS lev dB -4.77
RMS Pk dB -4.58
RMS Tr dB -4.98
Crest factor 1.73
Flat factor 0.00
play -r 48000 -c 1 -b 24 -n synth white vol -6dB
The stats of "-6dB" white noise generated by SoX:
DC offset -0.000290
Min level -0.501187
Max level 0.501187
Pk lev dB -6.00
RMS lev dB -10.77
RMS Pk dB -10.55
RMS Tr dB -10.99
Crest factor 1.73
Flat factor 0.00
From the stats I think that full volume white noise should not cause the issues. Or am I wrong and the volume must be really reduced?
Attachments
The same situation happens with SPDIF input.
Measurement setup (Linux):
SoX -> Int. soundcard -> SPDIF -> Buffalo 3 -> Preamp -> Juli@ -> ARTA
But now, the SPDIF white noise -6dB is correctly filtered. The measurement in previous post was done on different PC with Windows.
Is the white noise with peaks at -0dB a valid test signal?
Measurement setup (Linux):
SoX -> Int. soundcard -> SPDIF -> Buffalo 3 -> Preamp -> Juli@ -> ARTA
But now, the SPDIF white noise -6dB is correctly filtered. The measurement in previous post was done on different PC with Windows.
Is the white noise with peaks at -0dB a valid test signal?
Attachments
You are right, here is my B3 settings & wiring:
- SPDIF autodetect OFF
- Oversampling & reclocking ON
- Fast FIR roll-off
- DPLL Lowest, 128x multiplier (BBB has low-quality onboard clock)
- Quantizer 6, True differential, outputs joined
- "Buffalo II compatible input wiring configuration" for both PCM & DSD stereo input
- Input remapping ON
Nevertheless, I've been testing the most extreme cases when data gets clipped during oversampling. In the normal conditions it works perfect. Sorry for confusion if it was not clear.
- SPDIF autodetect OFF
- Oversampling & reclocking ON
- Fast FIR roll-off
- DPLL Lowest, 128x multiplier (BBB has low-quality onboard clock)
- Quantizer 6, True differential, outputs joined
- "Buffalo II compatible input wiring configuration" for both PCM & DSD stereo input
- Input remapping ON
Nevertheless, I've been testing the most extreme cases when data gets clipped during oversampling. In the normal conditions it works perfect. Sorry for confusion if it was not clear.
OK, so the DAC doesn't have enough headroom built-in... at least it clips cleanly and doesn't wrap around like some buggy ones do !
Since it is just a setting (not a hardware bug) it will go away if you reduce the digital volume a little.
The spectrums you get are normal, if the oversampled signal clips, it will generate out of band noise.
Since it is just a setting (not a hardware bug) it will go away if you reduce the digital volume a little.
The spectrums you get are normal, if the oversampled signal clips, it will generate out of band noise.
Almost none have. Oversampling DACs cannot handle artificially generated signals like 0dBFS square in principle (oversampling + interpolation principle).