The effect of HPF changes

[tjk]

I am using the Cirrus Logic/Audio card for a sub-sonic audio project and was disappointed that the card was abandoned. I received two of the AudioInjector Stereo Cards and have tested them as potential replacements. They use a different Wolfson chip than the Cirrus Logic card. The chip used is still supported by Cirrus Logic. I did extensive testing with both noise and square waves and they seem to have a low 3dB point out of the box (no modifications) of 0.3 Hz when proper configurations are set. The hardest thing to decipher was how to turn off the digital High Pass Filter used for DC blocking. Matt provided the key -- type "m" to toggle the on/off control using the alsamixer controls. The chip documents show a 3dB point of about 3.5Hz with the filter on, but gave no data for filter off.

Testing setup:

A Raspberry Pi 3, with 2.5 amp power supply
Linux kernel: 4.9.30+ #1001 Fri May 26 16:03:39 BST 2017 armv6l GNU/Linux
Cables for BNC to RCA
BCN T adapter
Ethernet Cable
Syscomp Circuit Gear CG101 with Version 1.24 software
MacBook Pro laptop with Audacity software

Procedure:

The CG101 was set up to drive a T adapter with BNC to RCA cables plugged into both R and L channel inputs. A stereo sound system was also used on the output channels to confirm audio record, but was not used in the testing. The CG101 was set to "noise" source for spectral testing with a few seconds of recorded data. I used "alsamixer' to view and change the HPF on/off switch. For record I used:

arecord -f S32_LE -r 48000 -c 2 noise_HPF_Off.wav . (HPF Off setting)


arecord -f S32_LE -r 48000 -c 2 noise_HPF_On.wav (HPF On setting)

Set the CG101 generator to 0.2 Hz Square Wave and repeat the above commands to record a HPF On and Off waveform.

For the random noise tests, I used Audacity to plot a 65k FFT spectrum. For the square wave test analysis, I used a Python program to read the .wav files and generate a .csv file for entering into a spread sheet in Libre Office.


Random Input Tests:

Shows much lower f3dB frequencies when the HPF is off. From the Spectrograms, I estimate well under 1 Hertz verses 3 to 5 Hertz for the filter On. Using a 0.2 Hertz square wave and then entering the data into LibreOffice Spreadsheet, I used "add trendline, exponential" curve fits for the exponential decay for both cases. The HPF On curve is not really a first-order system (slight undershoot), but the first-order exponential does fit it pretty well, leading to a calculated f3dB of 5.9 Hz. The HPF Off plots show a much longer time constant, leading to a calculated f3dB of 0.195 Hz.

Conclusion:

This card can be used for sub-sonic recording.

Images:

Noise_Spectrum_HPF_Off

HPF_Off.jpeg (114.34 KiB) Viewed 1795 times

Noise_Spectrum_HPF_On

HPF_On.jpeg (115.91 KiB) Viewed 1795 times

Libre_Office_Plot

HP_On-Off_Plots.jpeg (118.85 KiB) Viewed 1795 times

[flatmax]

Thanks for informing us here. It is great to know that your card modifications allow it to reach such low frequencies !

Matt