We are trying to use the AES-FMC-HDMI-CAM on a Xilinx KCU105 development board.
We have drivers for the ADI ADV7511 on the KCU105 itself that run on a MicroBlaze in the design and have successfully configured that device.
We took that driver and modified it to deal with the differences in the I2C subsystem on the AES-FMC-HDMI-CAM and the slightly different video interface that the FMC has (embedded video syncs for example).
Unfortunately the driver for the FMC locks up when the MicroBlaze tries to configure the TCA9548 I2C switch on the FMC. We're reasonably sure that the code is correct since the driver for the ADV7511 on the KCU105 also has to configure a TCA9548 to talk to the ADV7511 on the KCU105. We just updated the switch port number and the I2C addresses in the FMC version of the driver.
While double checking everything one thing popped out when we were looking at the AES-FMC-HDMI-CAM schematics was that the TCA9548 is powered by the FMC VADJ feed. That's a problem when trying to use the FMC on a KCU105 since the supply range for the TCA9548 is 2.5 to 3.3V and the maximum voltage that VADJ can be set to on the KCU105 is 1.8V.
So we have been thinking of modding the FMC to connect the VCC pin on the TCA9548 to connect it to the 3.3V supply on the FMC rather than VADJ. We still have to check to see if the I2C inputs on the TCA9548 will still function correctly with a 1.8V swing and if the ADI ADV7511 will function correctly with 1.8V inputs. I don't see a problem with the ADV7611 HDMI receiver since it has level shifters on the signals driving the FMC pins.
Does anyone have any experience with using the AES-FMC-HDMI-CAM on more recent Xilinx development boards that do not support running VADJ at 3.3V?