Post Go back to editing

ad7768 eval start_n pin

The AD7768 manual says that the start_n pin needs to be tied high if not used.

The AD7768EVAL schematic shows that start_n is not tied high by default.

The ad7768 HDL reference design brings this pin into the Zynq's GPIO interface

The ad7768 linux driver, and the DTS file, don't reference GPIO anywhere.  Is the this reference design pulling this pin high somewhere else that I don't see?

  • Hello Jay,

    This seems to be a bug. Thank you for mentioning it. This GPIO should be driven from software. We'll fix this soon.

    Regards,

    Stanca

  • Thanks Stanca

    I have verified that this design does work in Linux, any idea how that is possible?

    As I mentioned before, the Linux design doesn't use any gpio for the ad7768 (as there is no gpio listed in the DTS file).  Does this seem correct?  Looking at the ad7768 datasheet it does appear that all control, other than some PCB level pin strapping, can be controlled via SPI, so I think it's fine.

  • Hello,

    I will keep investigating, but right now it seems that the neither the software, nor the hardware are pulling the START pin as instructed in datasheet. My theory is that the pin is floating in a high impedance state since it's routed to a PS GPIO that has no particular drivers enabled. That means that a (very weak) pull-up may exist on it since the IOVDD trace would be relatively close.

    I think this can cause a slippery bug where, depending on external factors more often than not, the board works well some of the times, but sees glitches other times, or might even not function.

    This is what I'm going with for now. But I haven't tested all possibilities yet.

    Regards,
    Andrei