LTM4653 doesn't start (again)

When it's connected to Vin the module starts (as shown in the osci picture) but I can't turn the module off, so I don't know. But I changed the 1Meg pull up to 100k without any noticable improvement.

It does turn off here since the output starts to go down when Vrun level is 0.
Is your area of concern after trying to turn the umodule on again after the initial Vrun becomes 0?

In this picture I control Vrun via the power switch of my device. So yeah, the issue is, that it does not turn on again, once I turned it off although nothing else changed. Unsually it works fine (in 147/150 PCBs).

Have you tried changing the IC? It might be a faulty IC since 147 boards are working properly.

I know it's faulty, since it's not working properly. I was rather interested in what happens or happened to the device so I can prevent it from happening to the other boards. And 2% failure rate is a rather bad quota for a >30€ uModule. Changing the LTM does resolve the malfunction but I can't send the customer a device back with a hand soldered BGA.

Can you give more context then on how it was used? What are the parameters and testing condition? What was the kind of test when it failed?

It is used as a 52V-to-12V converter with 2-3A of load current. It sits in a 19" rack device in a server rack with aroud 35°C ambient temperature, has a heat sink and forced cooling with a fan. The customer just reported that they the device did not turn on anymore (after about 1 year of normal operation), no more data on that end. The other two failed devices did not work out of the box. What additional parameters or testing conditions do you want to know? On the same PCB are three more LTM4653 which work as 52V-48V converters and they never showed that particular issue.