I get new AD1865 NOS DAC working, I use for I/V pure 100 ohm resistor Halco H8, AD1865 outputs are paralleled, no any filter applied to output resistor.
Strange is that on 20 kHz I get drop in SPL for 0.7 dB.
On image 1 kHz level with 100 ohm is 53.9 dB, 200 Ohm load 59.5 dB, on 20 kHz levels are 53.2 dB and 58.8 dB.
Measured is sweep what is supplied to DAC by SPDIF from CD player. From same player analog output with same sweep, I get rise on 20 kHz by 0.2 dB. Measured with REW soft and EMU0204. I get same results with Arta software. Tested DAC also with sweep from DSP, same results.
I tested also singel (not parallelde connection) AD1865 ouput, same results.
I mesured also sweep with EMU0204 on 96 and 192 kHz sample rates with Arta, same results.
What can be wrong?
As when measured with 2 different loads there is no changes in high frequency drop, reason cannot be parasitic capacitance in measurement cables or circuit board, bad/wrong capacitors in Ad1865 analog power supply. Capacitors in supply are 10 uF OS-CON paralleled with 0.1 uF film capacitor.
DAC SPDIF receiver is AK4113, it had no filters, except de-emphasis what cause bigger SPL change than 0.7 dB, so it can obviously not be the reason.
I believe AD1865 chips are not fake as THD is less than 0.003% on lower than 1 kHz and little higher on higher frequencies.
Had anyone measured AD1865 with pure resistor I/V, what where the results, I not find any on internet?
Is drop on 20kHz property of AD1865 or...?
Strange is that on 20 kHz I get drop in SPL for 0.7 dB.
On image 1 kHz level with 100 ohm is 53.9 dB, 200 Ohm load 59.5 dB, on 20 kHz levels are 53.2 dB and 58.8 dB.
Measured is sweep what is supplied to DAC by SPDIF from CD player. From same player analog output with same sweep, I get rise on 20 kHz by 0.2 dB. Measured with REW soft and EMU0204. I get same results with Arta software. Tested DAC also with sweep from DSP, same results.
I tested also singel (not parallelde connection) AD1865 ouput, same results.
I mesured also sweep with EMU0204 on 96 and 192 kHz sample rates with Arta, same results.
What can be wrong?
As when measured with 2 different loads there is no changes in high frequency drop, reason cannot be parasitic capacitance in measurement cables or circuit board, bad/wrong capacitors in Ad1865 analog power supply. Capacitors in supply are 10 uF OS-CON paralleled with 0.1 uF film capacitor.
DAC SPDIF receiver is AK4113, it had no filters, except de-emphasis what cause bigger SPL change than 0.7 dB, so it can obviously not be the reason.
I believe AD1865 chips are not fake as THD is less than 0.003% on lower than 1 kHz and little higher on higher frequencies.
Had anyone measured AD1865 with pure resistor I/V, what where the results, I not find any on internet?
Is drop on 20kHz property of AD1865 or...?
Attachments
Could it be non-oversampling high frequency rolloff:
http://www.diyaudio.com/forums/digital-source/53611-dac-way-oversampling-non-oversampling.html
http://www.diyaudio.com/forums/digital-source/53611-dac-way-oversampling-non-oversampling.html
Yes, by itself a NOS DAC can't have any FR issues they are embedded in the data. Start here http://www.cypress.com/file/123171/download
Last edited:
All dacs have high frequency rolloff. The amount depends on the sampling frequency of the signal that is sent to it (nos)or has to convert (internal oversampling). The higher the sampling frequency, the less attenuation at a given signal frequency.
ATX7006: DAC frequency response
So a 44.1 kHz sampling frequency 20kHz signal has more attenuation than an oversampled ( 88.2, 176.4 kHz etc.) 20kHz signal.
ATX7006: DAC frequency response
So a 44.1 kHz sampling frequency 20kHz signal has more attenuation than an oversampled ( 88.2, 176.4 kHz etc.) 20kHz signal.
Last edited:
Sorry I worded that poorly, it's the sinc() rolloff from the zero order sample and hold. The same DAC will roll-off less at a higher sampling rate, it's just the math not the DAC.
This not fit with my measurements. I use sample rate 44.1 kHz, by calculator on link, I must have drop of 3.168 dB on 20 kHz, I had only 0.7 dB. Probably DAC in real not working on 44.1kHz sample rate and AK4113 or somthing else is making oversampling?
Last edited:
It would be worth exploring how your software is calculating the amplitude, particularly at higher frequencies. In the presence of fairly close-in image products its not exactly trivial to get an accurate measurement of the amplitude of the original signal.
I find my problem.
I used DSP for transmitting SPDIF and it was converting 44.1 to 96 KHz sample rate (ASRC).
With direct SPDIF from CD player with real 44.1 sample rate sweep, I get drop on 20 kHz about 3.3 dB. Seems now all fits together as with 96 kHz sample rate drop on 20k must be 0.629 dB.
And yes, NOS DAC, specially playing 44.1 kHz sample rate signal, needs compensation of high frequency drop. I will do it on DSP.
On image part of aliasing signal is also visible and indicates that filtering on analog side is also needed.
I used DSP for transmitting SPDIF and it was converting 44.1 to 96 KHz sample rate (ASRC).
With direct SPDIF from CD player with real 44.1 sample rate sweep, I get drop on 20 kHz about 3.3 dB. Seems now all fits together as with 96 kHz sample rate drop on 20k must be 0.629 dB.
And yes, NOS DAC, specially playing 44.1 kHz sample rate signal, needs compensation of high frequency drop. I will do it on DSP.
On image part of aliasing signal is also visible and indicates that filtering on analog side is also needed.
Attachments
Last edited:
- Status
- Not open for further replies.