CODEWITHSUNDEEP
beagleboneblackmicrophonemicroprocessors
parags
parags

Reputation: 551

Interfacing an I2S microphone with BeagleBone Black

I am trying to interface an I2S microphone (https://www.adafruit.com/product/3421) with BeagleBone Black. I followed the article (http://www.ti.com/lit/an/sprac97/sprac97.pdf), and able to update the device tree, and the Linux kernel as suggested in the article.

The I2S component (for microphone) of the device tree is included as a dtsi in the main device tree source. The content of the dtsi is as below

&am33xx_pinmux {
        mcasp1_pins: mcasp1_pins {
             pinctrl-single,pins = <
                /* sink must enable receivers */
                0x1a0 0x23
                /* P9_42 mcasp1_aclkx - bit clock */
                0x1a4 0x23
                /* P9_27 mcasp1_fsx - frame sync */
                0x1a8 0x23
                /* P9_41 mcasp1_axr0 - i2s input */
             >;
        };
};

&mcasp1 {
        #sound-dai-cells = <0>;
        pinctrl-names = "default";
        pinctrl-0 = <&mcasp1_pins>;
        status = "okay";
        op-mode = <0>;
        tdm-slots = <2>;
        num-serializer = <4>;
        serial-dir = < /* 1 TX 2 RX 0 unused */
                2 0 0 0
                >;
        rx-num-evt = <1>;
        tx-num-evt = <1>;
};

/ {
        pcm5102a: pcm5102a {
                #sound-dai-cells = <0>;
                compatible = "ti,pcm5102a";
                status = "okay";
        };

        sound1: sound@1 {
                compatible = "simple-audio-card";
                simple-audio-card,name = "PCM5102a";
                simple-audio-card,format = "i2s";
                simple-audio-card,bitclock-master = <&sound1_master>;
                simple-audio-card,frame-master = <&sound1_master>;
                simple-audio-card,bitclock-inversion;

                simple-audio-card,cpu {
                        sound-dai = <&mcasp1>;
                };

                sound1_master: simple-audio-card,codec {
                #sound-dai-cells = <0>;
                sound-dai = <&pcm5102a>;
                clocks = <&mcasp1_fck>;
                clock-names = "mclk";
                };
        };
};

The final device tree (decompiled from am335x-boneblack.dtb) is attached here. McASP entries (mcasp1_pins), including pins as specified in the TI document above are on line no 1077.

I have also compiled the kernel with a new driver pcm5102 as suggested in the document. Finally, I see the driver listed in the output of the arecord command.

root@arm:/sys/class/gpio# arecord -l
**** List of CAPTURE Hardware Devices ****
card 0: PCM5102a [PCM5102a], device 0: davinci-mcasp.0-pcm5102a-hifi pcm5102a-hifi-0 []
  Subdevices: 1/1
  Subdevice #0: subdevice #0

However whenever I try to record audio, I am not getting any audio data. The audio file is formed, but the file size is always 44 bytes irrespective of how long I try to record audio for. Clearly no data is there in the file.

Recording using arecord command gives error as below

arecord -d 10 -Dhw:0,0 -f dat audio.wav
Recording WAVE 'audio.wav' : Signed 16 bit Little Endian, Rate 48000 Hz, Stereo
arecord: pcm_read:2032: read error: Input/output error

Moreover, whenever I try to connect the clock (MCASP1_ACLKR (Bit Clock) - P9_42) with my oscilloscope, I don't see any pulses. Looks like there is no proper clock signal on this pin. What can I do to debug, and fix this?

Any ideas?

Parag

Upvotes: 2

Views: 1951

Answers (1)

Simon De Meyere
Simon De Meyere

Reputation: 96

The PCM1864 board used in the link you provided generates its own clock and operates in master mode. The microphone you are using, however, requires the bus master to generate both the bit clock and frame sync signals to operate.

The McASP module of the am33xx processor can generate these signals, although some modifications to the driver/device tree might be necessary.

The mcasp signals can be mapped to the following pins on the BeagleBone board (generated by TI's pinmux tool). Note, the device used here is mcasp0, not mcasp1. I assume TI was using a different version of the beagle bone which had mcasp1 connected to the IO header. 

pinctrl-single,pins = <
        AM33XX_IOPAD(0x9ac, PIN_INPUT_PULLDOWN | MUX_MODE0) /* (A14) mcasp0_ahclkx.mcasp0_ahclkx */
        AM33XX_IOPAD(0x99c, PIN_INPUT_PULLDOWN | MUX_MODE0) /* (C12) mcasp0_ahclkr.mcasp0_ahclkr */
        AM33XX_IOPAD(0x990, PIN_INPUT_PULLDOWN | MUX_MODE0) /* (A13) mcasp0_aclkx.mcasp0_aclkx */
        AM33XX_IOPAD(0x994, PIN_INPUT_PULLDOWN | MUX_MODE0) /* (B13) mcasp0_fsx.mcasp0_fsx */
        AM33XX_IOPAD(0x9a0, PIN_INPUT_PULLDOWN | MUX_MODE0) /* (B12) mcasp0_aclkr.mcasp0_aclkr */
        AM33XX_IOPAD(0x9a4, PIN_INPUT_PULLDOWN | MUX_MODE0) /* (C13) mcasp0_fsr.mcasp0_fsr */
        AM33XX_IOPAD(0x998, PIN_INPUT_PULLDOWN | MUX_MODE0) /* (D12) mcasp0_axr0.mcasp0_axr0 */
        AM33XX_IOPAD(0x9a8, PIN_INPUT_PULLDOWN | MUX_MODE0) /* (D13) mcasp0_axr1.mcasp0_axr1 */
    >;

According to the datasheet of the processor, the clock signals of the receiver (ahclkr, aclkr, fsr) can be setup to run independently of of in sync with the clock signals of the transmitter and dividers can be specified as necessary. The pins which are used to provide the clock signals will need to be set to PIN_OUTPUT.

It seems to me that the overlay provided by beagle board here, when used in conjunction with the default clock device tree insert here, derives the clocks from the system clock. You might want to experiment with this.

Upvotes: 1

Related Questions