LTC4310 Over long distances

Dear Support,

There is a similar question here, but it does not address my question exactly.

The LTC4310 communicates in a pair with another LTC410 via 2 differential pairs. Isolation is achieved through the use of capacitors or ethernet transformer.

My question is whether these differential pairs are suitable for transmission over twisted pair cables for a distance of up to 5m if the isolation is located at one or both ends?

I have tried to illustrate the scenario below:



All application notes seem to indicate that the pair of LTC410 should be located close together.

I have reviewed a similar part LTC4331, as well as other isolation methods, but none are suitable at this stage.

Thank you kindly

  • +1
    •  Analog Employees 
    on Oct 15, 2021 11:53 AM

    Hi Liam, 

    You should have an FAE available to contact directly, and I would advise to reach out to them.

    The LTC4310 won't be for longer distances. The LTC4331 and the LT3960 would be the recommended options for sending an I2C signal over a 5m (and longer) cable. I would say a pair of either those devices in series with an ADuM1250 or ADM3260. What were the constraints that those devices couldn't be used?

    Regards,

    Jason   

  • Thank you Jason.

    In our current implementation we have a PCA9615 for differential transmission, which we've discovered needs to be isolated. The V_IL and V_OL don't quite work between ADM3260 and the PCA9615 (and same as the ADuM1250 and LT3960 respectively). V_OL for ADM3260 (side 1) is 0.6 - 0.9 V, but V_IL of PCA9615 and LT3960 is 0.4V.

    This could be solved with a buffer, but the solution starts to feel cumbersome. Which is why LTC4310 seemed appealing, due to its simplicity (if it was possible, that is).

    LTC4331 would actually be ideal, but part availability at the time of writing is poor, which is unfortunately driving design decisions. 

    In any case, thank you for clarification that LTC4310 is not for long distances.