With two EndPoints and an Arduino Uno Qwiic Shield and all Qwiic cabling;
First test at 3.3V:
State of the Endpoints: The primary Endpoint has the 1-0 jumper is closed.
The terminating Endpoint has the 1-0 jumper **closed**.
Problem: If a 3 foot Ethernet cable is attached, the Qwiic OLED [Address 0X3D] or a Regular OLED [Address 0X3C] both work.
With a 18 foot Ethernet cable , which tested good, the OLED does not work.
While another 15 foot Ethernet cable , which tested good, **the OLEDs do not work.**
Second Test at 5V with a State of the Endpoints:
The primary has 1-0 jumper **open** and a regulated 5V on pin VCC1 and Ground on GND.
The terminating Endpoint is 1-0 **closed**.
Problem: If a 3 foot Ethernet cable is attached, the Qwiic OLED [Address 0X3D] or a Regular OLED [Address 0X3C] both work.
With a 8 foot Ethernet cable OR a 15 foot Ethernet cable, in which both cables tested good, **the OLEDs do not work**.
All Ethernet cabling use tested good.
Any one try something similar that worked and if so, what was done differently?
How were your non working ethernet cables tested?
Are all 8 wires present and connected in the non working cables and are they wired straight though? (Not crossover cables)
All of the cables were tested with a PDE33M Cable Tester. It appears the problem is the distance, as it works at 3 feet??? I2C at about 18 feet is what I need.
Must be just enough resistance in those longer cables to cause issues with i2c data transfer. Not really sure what is the problem could be otherwise.
Are the cables copper plated or solid copper wire? If you can try using the solid copper stuff. It might help. Id image the plated stuff will have more resistance and may pass basic cable testing tools.
To my knowledge testing OLED devices in this way has not occurred at SparkFun. It seems like a pretty specific use case.
Ethernet is the best(Cheapest / most practical) way to transfer signals through long distances.
Using a controller with a faster clock speed might help. Not sure exactly what speed would transfer data this far without issue but it would need to be better than an ATMega328.
