We have an application using 6 of the AD7794 converters. We have tied the clock pins of these devices together. Occasionally, we have been getting failures of one or two of these components during our board test. Replacing the component corrects the failure, but recently we discovered that lifting the clock pin can also correct the failure. We are not currently using clock synchronization but the board was designed with this capability (almost 7 years ago). Is there a failure mechanism at work here that isn't documented? Thanks for your help.