We are using the ADS5294 chips within one of our PCB designs and we have recently discovered that some of the chips (~5% of the PCBs manufactured PCBs exhibit this issue) will not lock PLL state when clocked at 10MHz. The result is corrupted output data since the chip keeps toggling states and never settles, even over a ~10 second period.
In the faulty 5% of our PCB batch (there are 3 ADS5294 devices per PCB) 2 of the chips operate correctly, 1 does not. The faulty chip can be made stable if it is heated up to a temperature of ~ 60 degrees C however as soon as it cools down the PLL state toggling re-commences. Increasing the clock frequency of the chip will also stop the state toggling however the device remains temperature sensitive and if cooled down (using freezer spray) it will start state toggling again. All faulty chips behave in this way, the only difference between chips is at what temperature the state toggling occurs.
Fully functional chips do not exhibit any of this behaviour. We are investigating with the manufacturer if there are installation issues that are damaging the chips.
To fix the issue we have placed the PLL into a fixed state and this results in the chips no longer toggling state and providing uncorrupted data, however we are concerned that we are simply covering up an issue with the chip that may results in total chip failure sooner than expected.
Couple of questions:
1) Is this an issue that you have seen before?
2) Can you provide details on the soldering limits handling for this chip?
3) We assume this is associated with the frequency detection part of the PLL, can you provide any further information about how this works?
4) Can you provide a reason for the failure mode seen in these chips and confirm that by simply disabling the auto sensing of clock frequency the chip will still maintain its performance as defined in the data sheet?
Any help you can provide on this issue would be gratefully received.
Regards
Mark
