TAS2563: internal voltage converter is NOT working properly

Hi Tim,

One test that can help to understand if there is a limitation from the device configuration is to initialize the device in ROM mode, perhaps your tuning settings are limiting the output too much.
We can double check the power performance based on your configuration. There are some questions I'd like to ask so we can get similar test setup:

• Please share your configuration files. We need at least the initialization script you use on your system, it can be a cfg file or if you have the PPC3 used to generate the config files it would be good.
• Please share details about your speaker load, if there's a data sheet it would help. Did you characterize the speaker? Or are you using some default values?
• Do you have any oscilloscope or other measurement captures so we can understand the difference in performance between the two modes?
• If possible, please share your schematics to double check the passive components related to the boost.

If you need to share any of this information privately, please feel free to send an invite for private e2e message and I can also share my email address over there.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

I will gather the data, but I am not sure how to invite for private E2E messages. Could you assist, because I really would like to share in a private chat.

Best regards,

Tim

Dear Ivan,

I gathered some first information to describe the issue a little closer. I decided to post this non-private as it is simple a more detailed descriprition of the issue.

Best Regards,

Tim

The answer from our development partner:

Measurements taken from our target system and setup strongly suggest that it’s not the power stage or audio tuning that limits us in achieving maximum output power, rather it is the boost converter.

1.

We can see from TAS2563 status registers that the error condition observed typically sets the Overcurrent bit in INT_LIVE0 register. The moment at which this scenario happens (and thus maximum power achieved) changes with boost inductor value. Theory suggests that increasing inductance (while keeping its saturation current rating the same) would increase boost output power. In practice, the opposite appears to be true, with maximum performance achieved using an inductance value of as low as 0.47 uH. Still with this value, performance level was found to be lower than required.

Boost converter architecture is hardly documented at all, so we have gained limited insight only. Given an operating voltage of 3.7 to 4.2 V (single Li-Ion cell), we have characterized boost converter maximum output current and found it to be limited to about 300 mA at 11 V, which is rather low to support the output power values claimed. This measurement was taken with BST_ILIM set to default of 3.96 A. Given the step-up factor from 3.7 to 11.5 V, and comparing to similar step-up converter architectures, this default value appears low, but is prohibited to be set much higher. 

2.

The behavior was found to improve with speaker loads of 8 ohms (or higher) compared to 4 ohms.

Ultimately, stability with 4 ohms load was insufficient at all.

3.

Operating the TAS2563 with the same tuning settings in External PVDD mode drastically improves performance, pushing output power to extreme values with no signs of instability at all, neither at 8 ohms nor at 4 ohms.

4.

From our experience with PPC3 software, it turns out that the TAS2563 has a multitude of registers that are poorly documented (or not documented at all), so there possibly is a chance to tune the boost converter for better performance, given that someone knows about the internal architecture of the boost converter and its settings.

Hi Tim,

Thanks for the additional information.

Is there a chance you can share the schematic and PCB layout? There are some specific things we can double check on your design to better understand the observed boost limitations:

• Boost inductor is usually 1uH, and as VBAT is lower the saturation current is more relevant. The inductor should not reach saturation as it would degrade the inductance considerably.
• Just to double check, we can evaluate on our side using your specific configuration; if we find any limitations due to configuration, we could check if there's something we can tweak to improve the performance.
• PCB routing is very important around the VBAT, SW, GREG and PVDD/VBST nets. If problem comes from PCB routing unfortunately a software fix would not help, but we can help to identify the root cause for any future projects.

I sent you an e2e friendship request so you can send private messages, otherwise feel free to reach out to my email address.

Best regards,
-Ivan Salazar
Applications Engineer

Dear Ivan,

thanks for sharing contact details. I copied them, so please feel free to delete them again. I will send out some more info via E-Mail.

Best Regards,

Tim Mackerodt

Hi Tim,

Thanks for reaching out, I'll follow up on the email thread later today.

Best regards,
-Ivan Salazar
Applications Engineer