EVM-K for TLV320AIC3107 - No word or bit clock.

This is an app note by the Stellaris team. For any specific questions about this application you would have to ask the Stellaris team.

As far as the AIC3107EVM goes, it is design such that the BCLK and WCLK are supplied by the TAS1020 on the mother board. Check to see that SW2-4 and SW2-5 are to the Left (on) and SW2-8 is to the Right (OFF). Also check to see that the USB Active LED is illuminated. You should see MCLK, BCLK and WCLK (LRCK) on J14.

I don't want clock from the TAS1020 - I want to use a master clock from my MCU and have the CODECs PLL generate the Bit & Word Clock (Master). I have done this previously, and now it seems to be failing. What would the jumper settings be to do what I have just described ? 

The problem doesn't seem to have anything to do with Stellaris? I'm not actually using Tiva-C, just examined the jumper settings from that app note. 

in this case you just need to move SW2-4 to the Right to disable the TAS1020 clocks, If you are using an external MCLK you also need to move SW2-5 to the Right and SW2-8 to the Left.

Check the MCLK pin on the AIC3107 to insure that the external MCLK is making it to the device. Next it would come down to your I2C writes. I would verify what you are writing with an I2C sniffer at the I2C pins of the AIC3107.

If you have the register values, I can take a look at them.

I edited this down, here is the condensed version of what I am writing after a hard reset. I verified that the reset pin is being pulled down completely (approximately equal to 5mV). This can only be done with the reset switch (SW-7) in the "right" position, otherwise low is 1.8V. 

//w 30 07 00  C0 00 00

//w 30 65 00 02

//w 30 02 44

// w 30 03 91 20 1E 00

// w 30 0B 01

//w 30 08 C0 (REDUNDANT)

//w 30 11 0F FF

//w 30 0F 00

// w 30 10 F0

//w 30 19 80

//w 30 13 7C

All writes were verified with a read. The only value that returned differently was 0x19 reads out an additional 0x02 on top of the value, which one of you engineers previously verified for me was correct. 

My current dip settings are as follows (R being silk screen side of the dip, reading letters facing up)

1.A0(L); 2.A1(L); 3.A2(L); 4,USB i2S (R); 5.USB MCK (R); 6.USB SPI(L); 7.USB RST(R); 8. EXT MCK (R).

MCK appears on test points, no word clock or bit clock coming back out.

As far as I2C - the changes I make to the device show up in the EVM-K GUI, so I would assume on top of verifying my I2C writes with re-reads after the writes, triggering a scope and counting bits, and ATMEL status returning 0 for TWI_SUCCESS, that I2C is not the problem.

And here is a detailed run-down on switch combinations I have tried based on what you said . My orientation is silkscreen (A0, A1, MCLK, etc...) to the right: 

8(L), 7,6(L); 5,4(R); 1,2,3(R); ---> No MCLK on the test point

Set 8 to the Right, MCLK from my MCU appears on the test point. (Opposite of what you told me?)

SO : 8(R), 7,6(L); 5,4(R); 1,2,3(R); ---> MCLK on the test point, however TWI then fails.

SO : 8(R), 7(R),6(L); 5,4(R); 1,2,3(R); ---> MCLK on the test point, TWI good, still no bit / wrd clock.

I'm kind of at a loss here, because the behavior is not consistent with what you have told me to do. We have a presentation in about 6 hours, and I've been trying to verify this thing for a week. Is there number I can call to talk to an engineer in real time? 

Other possible Clues: Checked the bits in 0x6B (Register 107) and returns all 0's which to me would support the fact that this is not an I2C error. 

After all of my reads/writes, I checked Register 94, and even when L/R ADC appear to be powered in the GUI (I added right power on for the sake of my sanity) it returns 0's, i.e. ADC's are not powered on. 

As a side note, my GUI gain is backwards from the data sheet. When you write 0000, you get -12 db in the GUI, not 0. 1000 produces the 0db value.