TAS5710 won't start above 12V. Locks up after shutdown.

Hi, Martin,

Sorry for the delay. I've asked my colleague to help you with this strange request. Something is wrong in your set-up, this should not be happening. We will try to help you debug your set-up remotely, but that is sometimes difficult.

-d2

Hi Martin,

As Don mentioned, this may be tough to debug remotely. I have a few questions to better understand the issue.

From the schematic side: Have you made any changes from the EVM schematic? If so, what are they/can you provide the schematic you are using?

How quickly are you sending these commands after each other? Are you making sure the I2S data is stable first? I have attached the shutdown sequence below from the d/s (page 34):

Shutdown Sequence

Enter:

1. Ensure I2S clocks have been stable and valid for at least 50ms.

2. Write 0x40 to register 0x05.

3. Wait at least 1ms+1.3 × tstop (where tstop is specified by register 0x1A).

4. Once in shutdown, stable clocks are not required while device remains idle.

5. If desired, reconfigure by ensuring that clocks have been stable and valid for at least 50ms before

returning to step 4 of initialization sequence.

Exit:

1. Ensure I2S clocks have been stable and valid for at least 50ms.

2. Write 0x00 to register 0x05 (exit shutdown command may not be serviced for as much as 240ms

after trim following AVDD/DVDD powerup ramp).

3. Wait at least 1ms+1.3 × tstart (where tstart is specified by register 0x1A).

4. Proceed with normal operation.

Not following this sequence may put the chip in an unstable state. Are you following this sequence when you shutdown the device as well?

In regards to the voltage difference, does the device not respond above 12.5V consistently? If you play audio at 12V and increase to 14V, does the audio stop playing?

Regards,

~John

Hi

Thanks for replying - I am sending the I2C commands by hand so there is plenty of time between steps.

3683.aap main v1.pdf

I have tried to attach a pdf of the relevant bit of the schematic with your web tool - let me know if you got it. If not let me have an email address so I can send as a regular email attachment. I am currently out of my office in China so please allow 24-48 hours to reply

I did briefly experiment before I left and found that once initialised I could raise the voltage - but I only tried as far as 16V. Also I increased the volume and the chip crashed during a loud passage of music. I don't think it was anywhere near the rated power and hadn't played loudly for more than a few seconds..

Martin

Hi Martin,

Thank you for the schematic, that was very helpful.

I have found an issue with your current setup that may be causing the TAS5710 to go into an unstable state. You have tied the MCLK and SCLK lines together and then are using an MCLK at 128x. The TAS5710 only supports an SCLK up to 64x (top of page 10 in the TAS5710 datasheet), so the I2S signal is not in a stable state, which could put the IC in an unstable state.

Try running MCLK at 64x and write 0x60 to register 0x00 (44kHz,64x MCLK).

Regards,

~John

Hi John

Sorry I forgot - I hand modified the PCB to separate MCLK and SCLK - so the schematic is out of date (and I cannot easily correct it remotely from China).

But I do have MCLK at 128x fs and SCLK at 64x fs - so unfortunately thats not the problem.

And remember - it does work at 12V - so I don't think its a logical error - its just not robust. 

Do you now if it is particularly sensitive to clock rise times, levels, power supplies, ground noise - I mean I've looked at all these. However I need a chip that "just works" not one that needs half a dozen other chips making the power supplies and signal levels "just so".

Martin

Hi Martin,

That makes sense, I was surprised how you would get any audio with that out of spec.

You mentioned before that you have the TAS5710 EVM, but couldn't getting the software working. Do you have pins on your PCB that allow access to the I2S and I2C lines? I understand you are in China and cannot do this now, but I would recommend trying to run these lines (and power) to the EVM directly (bypassing the MC57xxPSIA). This would help isolate where the problem is.

Another thing to note is that the TAS5710 is a high-performance amp with high current and switching frequencies, so this issue could be related to PCB layout. This IC doesn't have a sensitivity weakness, however PCB layout outside of the recommendations can introduce some unusual problems.

A couple questions related to board layout:

Is your board layout significantly different than the EVM in any way?

Does the PCB follow the decoupling capacitor guidelines from the "Power Supply" section on page 15 of the d/s?

Does the PCB take full advantage of the PowerPad? (For more information see http://www.ti.com/lit/an/slma002g/slma002g.pdf)

Regards,

~John

Hi John

That's a great idea - I can attach the EVM in about 15 minutes - I will do that on my return to UK and post my findings.

Martin 

Hi John

I have now tested the TAS5710EVM with my system and it works OK at higher voltages. 

I am assuming it is a layout problem and have just designed a new PCB that is pretty much a copy of the EVM layout. I should have that back next week.

On your schematic of the EVM - I cannot find any reference to the thermal pad - and the data sheet does not say anything about connecting to it electrically - so my original PCB had this pad floating.

I see from your Gerbers that you connect this pad to ground and can then route through this area on both sides of the PCB and that probably helps reduce impedance on ground.

Is an electrical connection to ground needed or just harmless? Maybe the schematic/datasheet could make this connection clearer.

I still have a remaining issue with the EVM when I shut down (write 0x40 to register 5) there is a loud bang through the loudspeakers. Is that expected? Is there a quiet way to turn off?

Martin

Hi Martin,

The thermal pad is mentioned (although briefly) after the Dissipation Ratings on page 7 of the TAS5710 d/s:

"The thermal pad must be soldered to a thermal land on the printed-circuit board."

This is very important for heat dissipation and better thermal performance.

As for the shutdown pop:

1 - X0E 92  ~increase the volume slew rate

2 - X07 FF  ~master volume mute

3 - X1A 14  ~increase the start/stop period to 538mS

4 - X06 07  ~soft mute

5 - X05 40 ~PWM stop

You may want to experiment with register 1A by increasing/decreasing the period to see if this helps.

Regards,

~John

Hi Martin,

I'm glad to hear that the code helped! The best practice is to follow our EVM layout and I apologize for not explicitly stating that it should be tied to ground. As I mentioned before, the PowerPad App Note has the most information about the PowerPad: http://www.ti.com/lit/an/slma002g/slma002g.pdf. On page 18, it shows the recommended 4 layer board connections.

Regards,

~John