TS5A3359: I3C Bus Switching

Hi Rami,

Thank you for your support.

I understand that I3C is push-pull protocol.

Are you aware of any I3C level-shifter or repeater solution which is able to change voltage between 1.8V and 3.3V?

Best regards,

Kazuki Itoh

Hi Rami,

• Looking at Figure 9-7 in the data sheet, the I3C signal is pulled up, is this OK?
  I think I3C will transition to Push-Pull mode when communicating, so I would like to ask if there is no problem by pulling up the Push-Pull type output like this.
• It also requires dynamic voltage switching between 1.8 V and 3.3 V.
  Is this addressed by connecting TSA5A4624DCKR to the power supply on the I3C Responder?

Best regards,

Kazuki Itoh

Hi Itoh-san,

I'm going to loop in the translation team to support the question on Figure 9-7 - TXP can you please support?

Are you using the TS5A4624 now? That device has a 55pF on capacitance, which would exceed the standard recommended max bus capacitance alone. Is the customer aware and okay with going with that device? See my recommendations about for a device with lower capacitance.
As for how you handle switching between 1.8V and 3.3V, you would need the supply to be 3.3V. If your I3C Responder has a supply of 3.3V then that will work to tie the supplies together

Thanks,
Rami

Hi Itoh-san, 

I will respond with an answer before noon CST today. 

Regards,

Tyler

Hi Itoh-san,

TCA39306 looks to be the best option here for 1.8V to 3.3V I3C level translation. 

Bus cap is 50 pF limited for I3C segments, however, I know some customers are operating I3C with cap loading greater than this. Please confirm if customer is okay going outside spec >50 pF for I3C. 

Regards,

Tyler

Hi Tyler, Rami,

I think there might be a misunderstanding. The response doesn't quite address the question.

In the customer's system, there is a need to dynamically switch a specific I3C bus between 3.3V and 1.8V at will, which goes beyond a basic level translation requirement.

Hi Itoh-san,

TCA39306 can operate in open-drain and push-pull mode. Its a passive device so during push-pull mode, the output will need to have a PU resistor. 

Regards,

Tyler

Hi Tyler,

Thank you. Could you elaborate on the datasheet description below? I think this is related to what my customer would like to ask.

> The TCA39306 has bandwidth to support the high speeds needed for I3C, but there are special considerations which are required. Since I3C uses both push-pull and open-drain, it may not be possible to support all I3C applications with a FET-based translator.

Also, is dynamic I3C bus voltage switching (1.8V <-> 3.3V) possible by connecting the TSA5A4624DCKR to the power supply of the I3C Responder side?

Best regards,

Kazuki Itoh

Itoh-san, 

Kazuki Itoh said:

> The TCA39306 has bandwidth to support the high speeds needed for I3C, but there are special considerations which are required. Since I3C uses both push-pull and open-drain, it may not be possible to support all I3C applications with a FET-based translator.

"All I3C applications" could refer HDR-DDR - double data rate modes where clocking samples on both the rising and falling edge of the clock. TCA39306 may not be the best for this use case due to passFET based level translator- i.e. it does not re-drive the I3C signal, the output is strictly reliant on the PU resistor for the HIGH logic which may or may not work for a push-pull use case at 25 MHz sampling. TCA39306 might be too slow for some applications of the full I3C spec. 

Kazuki Itoh said:

Also, is dynamic I3C bus voltage switching (1.8V <-> 3.3V) possible by connecting the TSA5A4624DCKR to the power supply of the I3C Responder side?

Can you describe this use case with a block diagram? I am not understanding how this would work. 

Regards,

TYler