Differential i2c dropouts on pi?

Aloha, I’ve been working a project that uses the Sparkfun Differential i2c (DI2C https://www.sparkfun.com/products/14589) with a raspberry pi 3B+. The board is used to connect several i2c devices to the pi via a 5’ Cat5/RJ45 cable. Over the year or so of experimenting with this, we have had a number of failures to communicate with the remote devices. The system has been reconfigured several times, using different DI2C boards. The current setup has only the Di2C board on the pi-side (no other i2c devices) and two devices on far end - enviromment sensor and an AD (16bit) board. 5vdc is provided to the Di2c on the pi side and used to power devices on far end. i2c on pi end is still 3v (jumper traces cut).

The system runs fine for a while (assuming cables arent bumped) but then, in middle of night, the application can no longer talk to its 2 devices on i2c. This causes an E-Stop condition and the app halts. Last test run (expected to take 2days) died about 3hr after starting. The next day, the i2c seemed to be working just fine.

Any ideas why the devices stop communicating? Could the pullup resistors on far end devices/di2c be causing issues? I need to double check but both devices’ pullup traces should be cut.

Also i see the DI2C board is out of stock and may perhaps be headed for Discontinued status. Is there an update on the horizon?

If the Di2c continues to fail, i may have to put a pi out on far end and just do comms via classic serial or conventional IP/ethernet.

Sounds like you might have a lose connection somewhere and that would definitely cause a dropout.

I don’t see any plans to discontinue these, but it looks like there may be a revision coming on the PCB and that would explain why these are out of stock right now.