• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 30 subscribers
  • Views 1218 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

PCM5310 codec: Very high output noise level

Darko Obretan
Prodigy 85 points
Other Parts Discussed in Thread: PCM5310, TPS65270

Hi,

I'm using PCM5310 for 4CH audio IN and OUT. I have external out of band noise filters (active 2nd order, fc=24kHz and pasive RC fc~30kHz). Power supply is clean, I have decoupling capacitor pairs (2u2 and 100nF) as close as possible on every power supply pin, and 1uF ceramic on VCOM pins (AD, DA...). Ground is connected under device as suggested. I'm using 9V for line output drivers.

When I route PCM5310 LINEIN12->LINEOUT12 then there is no audible noise present (~ -90dB).

As soon as I configure codec and unmute it (and without any digital data or clocks present on any digital ports), there is a constant white noise present at around -74dB measured with RME fireface audio card mixer (with input gain set exactly, so that maximum signal level reaches around 0dB).

If I provide some clock input and zero DATA, then noise rises for another 3dB to -71dB.

Codec is configured and running as 48kHz, 24bit I2S, all CLK ports are in SLAVE mode configured as inputs, and ALL ADC & DACs uses same slock source from digital port 2. Codec is used without deemphasis, in 64x oversampling mode, and there are no audible pops and clicks due to some clock settings mismatch, and audio sounds OK otherwise.

Noise is independent on digital attenuation level setting for DAC12 and DAC34, so it must come from DAC section or after.

What could be causing so much noise ?

Regards, Darko Obretan

I'm attaching image of recorded audio signal levels and its spectrum.

  • 0
    TI__Genius 16270 points

    Darko,

    Would you be able to attach a schematic, and a BOM/layout?  Are all of the grounds tied under the chip, or just some of them?  Also, can you notice any improvement when increasing the value of the decoupling capacitors to 4.7uF or greater?

  • 0 in reply to Patrick Galvin
    Prodigy 85 points

    Hi Patrick,

    I tried to do all of those things, but noise level was exactly the same:

    - Added some decoupling capacitors (I already had combination of 100nF+4u7 on every power supply pin)

    - Stopped all digital clocks (LRCK, BCK, DATA) except system clock

    - Provided external system clock (not correct 12.288MHz, but 10,9,8,7 and 20MHz), but noise was everytime the same

    - Disconnected my TPS65270 based DC-DC power supply and provided external 3.3V laboratory power supply

    - Added/removed 470uF low ESR tantalum bypass capacitor on 3.3V supply

    - Replaced ferrite chip beads with coil or straight connection without ferrite chip bead

    - Replaced external 9V power supply with 9V battery


    Noise level on headphone output is lower, but signal levels are also lower in ~same relation.
    All of the ground pins are connected directly under the chip.

    I'll post layout and schematic as soon as I get approval, but is it possible to send them just to your email? (this I can do immediately)

    Best regards, Darko Obretan

  • 0 in reply to Darko Obretan
    Prodigy 85 points

    Hi Patrick,

    I'm posting PCB layout:

    Any ideas appreciated.

    Best regards, Darko Obretan

  • 0 in reply to Darko Obretan
    Prodigy 85 points

    Hi,

    I can get only ~3.2Vpp (4.5V centered) out of the codec with digitally generated sinus signal on ADAU1442 DSP at maximum 24bit digital level (if I raise level more, it is clipped).

    I set DAC attenuation to 0dB (0xFF).

    Could someone explain how to get claimed 2V RMS or even 2.4V RMS out of PCM5310 if analog section of the DAC is running only on 3.3V single supply?

    Are the DA converter and high level output buffers AC coupled internally? (I don't think so)

    So: 2V RMS = 2V * 2 * square_root(2) = 5,66V peak to peak

    Analog section is also running on 3.3V and claim is the same, that it should take 2V RMS input, which should mean 5.66Vpp, but I think this is stated wrong.

    I think it is only taking 2.83Vpp centered around 3.3V/2, so from 0.235V to 3.065V. Is this true?

    I don't see sense here... so I'm thinking where I'm wrong or the datasheet is?

    Regards, Darko Obretan

  • 0 in reply to Darko Obretan
    Prodigy 85 points

    Hi,

    My new findings are:

    - I can get 5.66V from line_out_34, but not from line out 12 ( I'm using same digital data and clocks on both DAC inputs, i mean I'm using same PCM5310 physical pins routed to DAC12 and DAC34)

    - On line_out_34 (routed to DAC34) is much better S/N ratio, noise level is down to -93dB at same conditions, and MAX signal reaches -1dB

    - If I route DAC34 to line output 12 then signal and noise is OK on line out 12

    - I'm using same attenuation levels & configuration flags for both DACs...

    It is getting strange...

    Regards, Darko Obretan