• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 30 subscribers
  • Views 620 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.

TLV320AIC3104: noise on output

ajayt
Genius 9555 points
Part Number: TLV320AIC3104
Other Parts Discussed in Thread: TPA6211A1

Hi I have attached some scope shots of the output showing the noise when the output is enabled.  Can you give me some debug tips, and also should the MCLK signal be more of a digital signal than a sine wave- there is nothing I could see in teh data sheet about this.

Attached are three screen shots. 01-away-volume-level7.bmp and 02-away-volume-muted.bmp show the noise we are seeing at the output of the codec. 03-SC_SYSTEM_CLOCK.bmp is the system clock line. Put some cursors round this and it measures 12MHz, AC coupled.

 

w.r.t. image 01 and 02, the blue trace is on one of the speaker pins w.r.t. ground, scope DC coupled. The yellow trace is at TP106 on the output of the codec (note there is no LPF on the schematic but we have tacked one on to the circuit) so TP106 is before the filter, AC coupled. The pink trace is Vamp_C1, TP103 which is the audio amp enable signal, SHUTDOWN on IC25.

 

In 01-away-volume-level7.bmp the volume level is set to 7 and we can see the audio message on the blue trace, its about 0.5 seconds long. Just a short message that says “away”. The whole wav file plays for about 1.5 seconds which you can see on the pink trace as this is how long the audio amp is enabled. The noise is hard to see on the blue trace but its more obvious at the output of the codec pin on the yellow trace.

 

In 02-away-volume-muted.bmp, I muted the volume down to zero so all we see is an attempt to play the message but the pulse of noise is still there, lasts about 2 seconds and runs for about a second even after the message has stopped.

 

This seems to fit with the explanation that we got here http://e2e.ti.com/support/audio/f/6/t/465738

  • 0
    TI__Mastermind 30210 points

    Ajayt, 

    This looks like it may just be out of band noise from the DAC modulator.  can you do an FFT and see where they noise profile sits?  What did you do to make the noise stop?  when you muted the audio,  which mute did you use?

    Take a look at this application report regarding out of band noise.  If this is OOB nose then will provide some guidance.

    best regards,

    -Steve Wilson

  • 0 in reply to Steve-Wilson
    TI__Genius 9555 points

    HI Steve, The output does have this RC filter, unfortunately the noise is not out of band I suppose, it shows up in the audio spectrum as a mush type noise.  The have seen other posts that suggest the line outs should go through the mixer channel ( even though it has higher power consumption) because the mixer has a filter built in, would you concur.   Can you get noise like this if the master clock is jittery or needs improving- Is there a way to check if the clock settings are done properly?  The output amp is a class AB btw the TPA6211A1 driven single ended though.  Are there any other things we should be addressing or looking at a bit more closely, such as clock jitter?

    Thanks

  • 0 in reply to ajayt
    TI__Mastermind 30210 points

    Ajayt, 

    The Mixer channel will attenuate some of the out of band noise,  but if you have the filter already,  using the DAC_L2 path is actually the higher performance path.  It does have more out of band noise,  but the in band noise and distortion will be lower. 

    The Master clock should be a Square wave,  and having a sine wave would definitely cause more jitter, and lower performance.  

    Where is this MCLK coming from?  is the codec in Master mode?

    best regards,

    -Steve Wilson

  • 0 in reply to Steve-Wilson
    TI__Genius 9555 points

    HI Steve,  

    we will clean up clock to square wave and see if that improves the noise 

    1. I had increased the gain at the LEFT_LOP and RIGHT_LOP from 0dB to 6dB but this seems to add noise to the output. So I reduced these back to 0dB.
    2. I set register 81 and 91 to an arbitrary level in this case -27dB and this made a big difference to the noise level. As I went down the range of setting here there didn’t seem to be any really change w.r.t. noise levels but if these registers were set to zero the noise increased so setting it to something made a difference.
    3. Also set register 88 to be the same as register 81 and this set the DC output at SPKR_OUT to a fixed DC level.

     

    Below is the latest register settings for reference.

     

    [ 4033.268763] (0)0 : (1)128 : (2)68 : (3)16

    [ 4033.268792] (4)4 : (5)0 : (6)0 : (7)10

    [ 4033.268802] (8)0 : (9)0 : (10)0 : (11)1

    [ 4033.268811] (12)0 : (13)0 : (14)0 : (15)0

    [ 4033.268820] (16)0 : (17)143 : (18)248 : (19)120

    [ 4033.268830] (20)120 : (21)120 : (22)120 : (23)120

    [ 4033.268839] (24)120 : (25)0 : (26)112 : (27)254

    [ 4033.268848] (28)0 : (29)0 : (30)254 : (31)0

    [ 4033.268856] (32)24 : (33)24 : (34)0 : (35)0

    [ 4033.268865] (36)0 : (37)0 : (38)16 : (39)0

    [ 4033.268874] (40)0 : (41)80 : (42)0 : (43)129

    [ 4033.268883] (44)129 : (45)0 : (46)118 : (47)118

    [ 4033.268891] (48)0 : (49)118 : (50)118 : (51)4

    [ 4033.268900] (52)0 : (53)118 : (54)118 : (55)0

    [ 4033.268909] (56)118 : (57)118 : (58)4 : (59)0

    [ 4033.268918] (60)118 : (61)118 : (62)0 : (63)118

    [ 4033.268927] (64)118 : (65)4 : (66)0 : (67)118

    [ 4033.268936] (68)118 : (69)0 : (70)118 : (71)118

    [ 4033.268944] (72)4 : (73)0 : (74)0 : (75)0

    [ 4033.268953] (76)0 : (77)0 : (78)0 : (79)0

    [ 4033.268962] (80)0 : (81)182 : (82)118 : (83)0

    [ 4033.268971] (84)118 : (85)118 : (86)9 : (87)0

    [ 4033.268980] (88)246 : (89)118 : (90)0 : (91)54

    [ 4033.268988] (92)118 : (93)9 : (94)0 : (95)0

    [ 4033.268997] (96)0 : (97)0 : (98)0 : (99)0

  • 0 in reply to ajayt
    TI__Mastermind 30210 points

    Alan, 

    1.  yes,  gain at the output will add noise. I recommend only using gain there if you really need it. 

    2. ?  That is odd... There shouldn't be any interaction here... are they using the ADC?

    3. can you clarify? 

    best regards,

    -Steve Wilson