frequency response reuqirement on ADAU1701
Hello All,
I got one question about the frequency response on ADAU1701. In my project, there is a functional requirement that the frequency response between 100Hz to 6.5kHz should less than ±0.5dB relative to 1kHz.
I have implemented all other functions except this one. attached is the project.AFILS.dspproj
with the project i have measured the frequency response based on the pink noise signal. but the result was far away from the goal.
below was my measurement process acc.to the IEC 60118-4 :
1. adjust the output gain to make sure the field strength is in the range of -1dB to +1dB using 1kHz sinewave signal.
2. change to signal to pink noise and measure the frequency response at 100Hz, the result was :
3. keep using pink noise and measure the frequency response at 5kHz,the result was :
I measured the input singal (the blue one) to ADAU1701 and the output (the yelow) to ADAU1701.
so far i didn't know what should i do or which module to use to implement this function.
Could anyone give some suggestions or instructions? Thanks
Best regards
Eason
Audio Signal Processors
software
SigmaDSP Audio Processors
adau1701
Hello Eason,
I do not see a problem with your schematic. I suggest you use a sine wave not pink noise for this.
The pink noise will be pseudo-random in frequency and level so being precise at one certain frequency is tough and then it will be dependent on the bandwidth of the measurement filter. The scope screenshot shows how the noise out of the generator is very broadband and has a lot of high frequencies and that will be attenuated once it goes through the ADC and DAC. You are running at 48kHz fs so you will not get anything over 24kHz.
The difference on the scope does not alarm me and would be expected.
I suggest you use the sine wave function. This NTI meter is a very good meter. I do not know how flat its output is over the frequency range you need. All oscillators are not perfect. Audio Precision has done such a great job with their products so I am biased towards their devices but I think this NTI is good. You can always test itself, route the output back to the input and sweep the frequency and see what it says about its own oscillator output. That will give you confidence to believe the readings.
0.5dB is rather tight but I think the DSP will do that. It has been a while since I have tested it that closely.
Dave T
Hello Dave,
Appreciate for your quick response and instructions.
About the pink noise singal, this is the condition must be met acc.to the IEC60118-4, shown as bolow:
Besides, the frequency response test device only support pink noise
yesterday, we found an hardware fault and test again after repaired. the frequency response performance is ok at 100Hz,but still terrible at 5kHz. we consider that the filter compoment on the circuit board is not available. we will replace it with another one and verify again.
Thanks again for your help.
Best regards,
Eason
Please log in to reply.