LM2842XMK-ADJEV: Component destroy after 1 to 5 power on.

Hello Sebastien,
This is very unusual, that device suddenly breaks after working for months.
Has anything changed between pre-production board and boards sued in production?

Looking at your schematic if you are using LM2842, which is 600mA device, your diode might be insufficient to handle inrush current during startup. The diode shown on the schematic is rated for 400mA and inrush current during startup might be higher. This could potentially explain why behavior is better when you add filter on shutdown pin, enabling device in more controlled manner.
To verify if this is the case please try following:
1. use diode rated for slightly higher current (for example 800mA), such as MA2YD26 (recommended in datasheet) or PMEG6010CEH
2. increase value for C1=220nF

If none of these helps, provide your board schematic and layout for further review.
Regards
Brani

Hello Brani,

thanks for your quick reply. I've ordered new diodes for testing (both reference are difficult  to find actually, I find this one with similar characteristic and same footprint than ZHCS400 : BSR106WS-TPMMSTR from Digikey.

I will receive them Friday.

Here the PCB around the LM2842 for information :

Regards,

Sebastien

Hello Sebastien,
After looking at the layout above i think the issue observed may be caused by less than optimal layout.
Following comments:
1. No input capacitor is visible . Proper placement of input capacitor (as close as possible to the IC pins 5&2) is absolutely essential. This device will experience large discontinuous currents on input and if input bypass capacitor is placed far way from the device it will negatively affect device behavior/performance. So placing small input bypass capacitor 100nF between pins 2&5 is very important for proper operation.
2. bootstrap capacitor C1 is not visible on your layout, meaning it is very far from device itself. C1 should be placed directly between pins 1&6 (between diode D7 and IC is enough room to place C1).
3. PCB trace that connects pin 6 to the diode D7 and inductor L1 is very thin. This trace needs to carry output current and from visual observation looks like it is insufficiently thick to do that. Suggest 2x or 3x of current trace thickness .
4. Suggest removing thin traces that connect ground side of D7 to ground and use solid copper fill for this connection.
5. suggest adding more ground vias.

For refernce please refer to suggested layout on page 16 of device datasheet.
www.ti.com/.../lm2842.pdf

Regards
Brani