TCA39306: The Use of TCA39306 for I²C Communication over Long Distance

Hi Conor,

Conor said:

Q1. Signal Integrity over 5m Distance:
Does the TCA39306 maintain stable operation and signal integrity over such a long distance? Are there any additional considerations or recommended practices (e.g., pull-up resistor selection, cable shielding) to mitigate signal degradation?

5 meter = 500 cm, estimate roughly 1pF / 1 cm, 500 pF loading. 

The TCA39306 is a bi-directional I2C/I3C level shifter, however, this device does not buffer (aka re-drive) the I2C signal. 

Although it meets the speed requirements, the TCA39306 will struggle due to the high cap loading. The TCA39306 does not have an current sink to deal with the long cabling. 

Consider switching to a device such as our 1 MHz bi-directional I2C buffer (FM+ capable) - TCA9617B. It has current re-drive suitable for FM+ communication with current sink IOL specs up to 30mA. 

If I had to guess, the situation is probably...

host (1.1V) <--> 5 meter cabling <--> target (3.3V).

For best implementation of the TCA9617B I would place it on the host side. 

host (1.1V) <--> A-side of TCA9617B <--> B-side of TCA9617B <--> 5 meter cable <--> Target (3.3V). 

One question here, do you know the series resistance of the cable? 

Conor said:

Q2. Propagation Delay:
For a 1 MHz clock speed, does the TCA39306's propagation delay remain low enough to ensure reliable communication over 5 meters?

I would look at the buffer mentioned above. TCA39306 is not a re-driver. If estimating ~500pF loading, the TCA39306 would most likely struggle across 5 meters. 

TCA39306 maintains low prop delay, but the question is can parasitic cap produced by the cable going to cause a large enough rise-time that would push it above spec when operating at FM+. 

Conor said:

Q3. I²C Compatibility:
Are there any specific settings or conditions to consider when using the TCA39306 for I²C communication at this speed and distance?

Pull-up resistance, VOL / VIL, parasitic bus cap loading, and current drive of the host and target device (IOL). 

I don't think TCA39306 will work, again refer to buffer device above. 

If I had to guess, a host (1.1V) would struggle to sink current through the TCA39306 across 5 meters of cabling at 1 MHz, this is why I am suggesting the TCA9617B. 

Conor said:

Q4. Overall Suitability:
In your experience, is the TCA39306 the best choice for this use case, or would you recommend another TI product that might better suit these conditions?

Refer to above! 

Question from my side, what type of application is this where FM+ I2C is occurring across a long distance? 

is there any opportunity to use CAN / RS232 / RS485 to communicate across the cable? I2C is usually not recommended to communicate across long cabling, especially at FM+ speeds due to open-drain protocol. 

Regards,

Tyler

With the high capacitance of a 5 m bus, I would recommend a buffer with a current source, i.e., TCA9803. But it works only up to 400 kHz, and even that speed might not be possible.

There are buffers for larger capacitances, e.g., P82B96 or P82B715, but they do not go above 400 kHz either.