TPS26750: MCU to PD I2C patch download

Russell DeGuzman

Part Number: TPS26750
Other Parts Discussed in Thread: TPS25751

Tool/software:

Hi there, I'm having trouble loading a patch from MCU to the PD controller over I2Ct as described in section 6.2 of the reference manual.

In particular, I'm confused as to the transaction structure and I2C addressing for sending the burst transfer.

I'm setting the target address in PBMs and getting a task success return value -

[I2Ct addr, DATA reg, data len, patch sz 0, patch sz 1, patch sz 2, patch sz 3, patch target address, 0x32]

[0x41, 0x09, 0x06, 0x80, 0x33, 0x0, 0x0, 0x50, 0x32]

When I try to send the burst on I2ct as described in figure 6.2, I get a task rejected (0x03) code

[i2ct addr, i2c patch address, data0, data1...]

[0x41, 0x50, 0xff, 0xff ...]

I have also tried sending the patch address directly as the I2c address, but the address is not ACK'd

[0x50, 0xff, 0xff ...]

What's the correct burst structure and how does it interact with the patch address specified in PBMs?

Thanks,

Russ

2 months ago

0 Chris Sterzik 2 months ago

TI__Guru 63181 points

Hello,

I am working to revise this document, but believe it is a good place to start.

https://www.ti.com/lit/pdf/slvafv8

Also, here is an example with the TPS25751 for reference.

/cfs-file/__key/communityserver-discussions-components-files/196/tps25751_5F00_PMBs_5F00_DBFg.csv

Regards,

Chris

0 Russell DeGuzman 2 months ago in reply to Chris Sterzik

Prodigy 10 points

Thanks Chris!
Those were very helpful.
One question I have is that the PBMc command looks to take on the order of 6ish seconds to complete (able to see in the logs you shared and also the same behavior on my device).
Is that normal? Can you give some insight into why this process takes so long? Is the PD loading the patch into non-volatile memory?

0 Chris Sterzik 2 months ago in reply to Russell DeGuzman

TI__Guru 63181 points

Hello,

Russell DeGuzman said:
Is that normal?

Yes.

Russell DeGuzman said:
Can you give some insight into why this process takes so long?

I am looking into this. In the example there are several cases where the MCU is polling the PD until a command is received/completed. The poling consumes CPU cycles of the PD controller making the response slower as it attempts act upon commands and respond to I2C polling. My goal is to move to an interrupt based method which will improve time and efficiency. The flow for this operation is not entirely clear because the new image can potentially change which interrupts are masked. If you get to a place where you want to attempt this, then let me know and we can compare notes.

Russell DeGuzman said:
Is the PD loading the patch into non-volatile memory?

No. The data is in volatile memory and is lost on a power cycle/cold boot.

Regards,

Chris

Power management

Power management forum

TPS26750: MCU to PD I2C patch download