i take out 2 first resistors from chip output and measure right there, signal drop 0.03V RMS only (from 1.06V@1kHz to 1.03V@20kHz), so i think problem not came from internal filters
After that, i resolding and measure without 2 wima capacitors, the freq response still have little drop but not too much at 20kHz like before
After that, i resolding and measure without 2 wima capacitors, the freq response still have little drop but not too much at 20kHz like before
Last edited:
The output stage on your dac has clearly the same topology as in the schematic of the original board in this thread (voltage output). Same schematic you showed here:
https://www.diyaudio.com/community/threads/es9038q2m-board.314935/post-7240000
Only the resistor and capacitor values are different. Resistor values can be seen in the picture but have you measured the Wimas and the SMD C0G capacitors next to the opamp?
What the other 2 opamps are for is unclear.
https://www.diyaudio.com/community/threads/es9038q2m-board.314935/post-7240000
Only the resistor and capacitor values are different. Resistor values can be seen in the picture but have you measured the Wimas and the SMD C0G capacitors next to the opamp?
What the other 2 opamps are for is unclear.
-0,2 dB in output @ 20kHz it's good.
It remains to simulate the analog output circuit with SPICE as I had proposed above
I think it is also necessary to take into account the details of bohrok2610 which we welcome and thank in passing
It is possible that the 'additional' OPAMPS can form a high pass and a low pass before the audio output: the best would be to bypass them
Less often equals better in audio...
It remains to simulate the analog output circuit with SPICE as I had proposed above
I think it is also necessary to take into account the details of bohrok2610 which we welcome and thank in passing
It is possible that the 'additional' OPAMPS can form a high pass and a low pass before the audio output: the best would be to bypass them
Less often equals better in audio...
Last edited:
2 wima are 220/100v. smd c0g are 10p, i also removed them on 1 channel and compare with second channel but nothing different
2 additional opamps, one i think preamplier with gain high, last opamp gain ratio is 1, i guest it for more stable. This DAC has 2 mode output, one direct to 1st opamp, one is through all opamps to get louder signal
Last edited:
Not sure that Audience is related to the long-est'd Audience AV (San Marcos, CA).
Audience
120 N. Pacific St., K-9
San Marcos, CA 92069
I've been using their excellent Auricaps for decades:
https://audience-av.com/
Hello trungdtmc,
it seems to me that this is the first hypothesis I mentioned...
Let's go back to your table with some of my 'additions' ->
Shape of the graphs according to FIR used for the D50s TOPPING DAC (same ES9038Q2M chip) ->
Short reminders on the 'notion' of THD ->
Even at -100dB, the distortion is inaudible to the human ear, it's a fact
Now, if we focus on the values mentioned in your table, we can say in a way that only 3 FIRs (*) are in line with the 'hi-fi standard' (20Hz to 20kHz) ->
Now to reproduce the frequencies at the extreme bands (20Hz and 20kHz), I wish you to have speakers but also and above all a room capable of reproducing them correctly but that's another story...
Regards
(*) = Linear phase/fast roll off, Minimum phase/fast roll off and Brickwall up to 19889,1Hz 'only')
PS: note that the Minimum phase / fast roll off will have a slight tendency to 'inflate' the low register while the Linear phase fast roll off will give a more 'punchy' sound reproduction.
it seems to me that this is the first hypothesis I mentioned...
Let's go back to your table with some of my 'additions' ->
Shape of the graphs according to FIR used for the D50s TOPPING DAC (same ES9038Q2M chip) ->
Short reminders on the 'notion' of THD ->
Even at -100dB, the distortion is inaudible to the human ear, it's a fact
Now, if we focus on the values mentioned in your table, we can say in a way that only 3 FIRs (*) are in line with the 'hi-fi standard' (20Hz to 20kHz) ->
Now to reproduce the frequencies at the extreme bands (20Hz and 20kHz), I wish you to have speakers but also and above all a room capable of reproducing them correctly but that's another story...
Regards
(*) = Linear phase/fast roll off, Minimum phase/fast roll off and Brickwall up to 19889,1Hz 'only')
PS: note that the Minimum phase / fast roll off will have a slight tendency to 'inflate' the low register while the Linear phase fast roll off will give a more 'punchy' sound reproduction.
Not so convinced on that point, myself. In audio forums there has sometimes been some misunderstanding as to what 'thresholds of audibility' represent. In particular they are not hard limits. They are actually estimates of average thresholds for a population, meaning it is estimated that 50% of the population can't hear below the threshold level, and the other 50% of the population can still hear below the threshold.
Hello Markw4
For this you would need:
It doesn't make a lot of people anymore, it seems...
As long as we use the FIR 'Linear phase / Fast roll off' THD @ -120dB at least we will be sure !
For this you would need:
- that this 'population' be passionate with a 'trained ear'
- that it is in perfect health and in good conditions for listening
- that she is under 30 years old
It doesn't make a lot of people anymore, it seems...
As long as we use the FIR 'Linear phase / Fast roll off' THD @ -120dB at least we will be sure !
Last edited:
Again, not so convinced about some of that myself. The ear is most sensitive at frequencies well below 10kHz. Also, thresholds don't necessarily mean someone cannot hear a particular frequency, it may be that simply turning up the volume level some can make a frequency once again audible.
Another factor is that the ear is more sensitive to higher order nonlinear distortion, and IMD tends to be more objectionable than HD. In addition, IME most people don't recognize the sound of low level IMD as constituting a 'distortion.' Again IME it does have a sound signature that can be recognized in at least some cases, vocal harmonies being one common example.
Another factor is that the ear is more sensitive to higher order nonlinear distortion, and IMD tends to be more objectionable than HD. In addition, IME most people don't recognize the sound of low level IMD as constituting a 'distortion.' Again IME it does have a sound signature that can be recognized in at least some cases, vocal harmonies being one common example.
With voltage output such as in the dac @trungdtmc is using THD of ES9038Q2M is probably above -100dB (see e.g. here).
Markw4 wrote: 'it may be that simply turning up the volume level some can make a frequency once again audible'
My question: what amplifier do you use to make a THD at -120dB 'audible', for example only from -120dB to -100dB (+20dB difference !)
I don't have it...
Knowing that while you do this, the other frequencies also increase: you need super powerful speakers then
My question: what amplifier do you use to make a THD at -120dB 'audible', for example only from -120dB to -100dB (+20dB difference !)
I don't have it...
Knowing that while you do this, the other frequencies also increase: you need super powerful speakers then
I have used various amplifiers including Benchmark AHB2 and various HPAs including Neurochrome HP-1 and HP-2. Remember, the question is not if we go from zero IMD to some non-zero value. The question is more to the effect of, is there any recognizable change in the distortion even if it is very small?
BTW, I know some people get tired of me mentioning this story so I will try to keep it brief: I once sorted recordings of some unity gain non-inverting audio opamp buffers in order of distortion in a blind listening test. The recordings and measurements of the distortion levels were all done by PMA. As I have explained before, it was very hard to do but it wasn't impossible in DBT.
BTW, I know some people get tired of me mentioning this story so I will try to keep it brief: I once sorted recordings of some unity gain non-inverting audio opamp buffers in order of distortion in a blind listening test. The recordings and measurements of the distortion levels were all done by PMA. As I have explained before, it was very hard to do but it wasn't impossible in DBT.