UCD90160: fail write on 0xfb register

Hi

Please capture the I2C waveform of both FA and FB command with the failing case

Regards

Yihe

Hello Yihe,

Per checked with customer, we found out there is register map which mentioned as below. May we know if the ignore meaning is to send NACK to host or still send ACK but register will not update the value?  And also, will UCD90160 clear the Enable bit automatically? Customer will check if they clear the bit before they send 0x7 to FBh register, if they sent the 0x4 first then this bit should be ignored. Thanks!

Best regards,

Ann Lien

Hello Yihe,

After checked with customer, UCD90160 works normally with the GPIO function even they saw write fail on this case. So they would like to know below questions. Could you kindly help on it? Thanks!

1. Any timing requirement on this access of 0xfb, since there is no issue on 0xfa?
2. Any process requirement to repeat the GPIO setting? No specific description in document.
3. Is it possible UCD is busy to set the configuration? I’m saying that because we see the repeat write failure message on single unit.
4. What is the behavior of UCD90160 if we saw this failure.?

Best regards,

Ann Lien

Hi

Please send the scope capturing waveforms. From the device perspective, it shallnever send out NACK for the 0xFB command.

So i need check the waveform to see whether there is any interference from outside.

Regards

Yihe

Hello Yihe,

Do you have strong evidence to prove it is related to noise? I've already asked customer to capture waveform but they think 0xFA command has no issue and host didn't change, that should not have problem on HW. Thanks!

Best regards,

Ann Lien

Hi

Customer need provides waveform so that we can review and help debug.

Regards

Yihe

HI, Yihe

please see below customer's reply. Would you please answer below behavior on UCD? Thanks!

Hi, Ann

You still can’t catch my point.

For 0xfa, it select the corresponding GPIO

For 0xfb, it enable this GPIO

Let’s change the way, please fill the possible behavior from TI’s point of view.

We absolutely will check how UCD90160 responds, but we have to know what the expected behavior is.

TI should have more concept than us.

Thanks.

#wedge_ucd_gpio_set 22 1
root@bmc:~# i2cset -f -y 2 0x34 0xfa 22       => select GPIO13 successfully.
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x7
Error: Write failed                                        => Possible behavior when writing failed????????                                    
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x4
Error: Write failed                                            => Possible behavior when writing failed????????

#wedge_ucd_gpio_set 12 1
root@bmc:~# i2cset -f -y 2 0x34 0xfa 12       => select GPIO14 successfully.
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x7
Error: Write failed                                            => Possible behavior when writing failed????????
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x4
Error: Write failed                                            => Possible behavior when writing failed????????

#wedge_ucd_gpio_set 13 1
root@bmc:~# i2cset -f -y 2 0x34 0xfa 13        => select GPIO15 successfully.
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x7
Error: Write failed                                            => Possible behavior when writing failed????????
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x4
Error: Write failed                                            => Possible behavior when writing failed????????

#wedge_ucd_gpio_set 20 1
root@bmc:~# i2cset -f -y 2 0x34 0xfa 20             => select GPIO3 successfully.
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x7
Error: Write failed                                            => Possible behavior when writing failed????????
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x4
Error: Write failed                                            => Possible behavior when writing failed????????

#wedge_ucd_gpio_set 21 1
root@bmc:~# i2cset -f -y 2 0x34 0xfa 21           => select GPIO4 successfully.
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x7
Error: Write failed                                            => Possible behavior when writing failed????????
root@bmc:~# i2cset -f -y 2 0x34 0xfb 0x4
Error: Write failed                                                    => Possible behavior when writing failed????????

Best regards,

Ann Lien

Hi 

Please ask them to provide waveforms if they need TI's help to debug the issue.

As explained in the previous email, device shall not NACK the 0xFB command at all.

Regards

Yihe

We can discuss more tonight Taiwan time.