TLV70433DBVEVM-712: TLV70433DBVR

Hi Mattia,

I don't think this is a documented failure. This kind of fault should be protected by this device. Have you repeated this failure more than once? 

Something I noticed is that the output isn't settling to 4.96V; only the highest peaks are getting to 4.96V, but the set point looks like it's about 4.3V, which is a little more dropout than the datasheet indicates for 40mA load, so I wonder if the pass device is damaged or if there's something else going on. Have you brought the input voltage above 5V after this fault to see if the output follows? It also looks like after the device stops regulating to 3.3V, the output is still recovering and overshooting and subsequently settling to ~4.3V, which seems to indicate that the control loop is still trying to regulate. 

Can you share what the layout looks like? My thinking is that the DBV package has pretty poor thermal performance, so if the device is increasingly hot during these repeated shorting events, eventually it could be getting damaged by the high power dissipation when the short initially happens because the headroom is maximized since VOUT goes to 0V and the current spikes before it can be limited. The last current spike is a pretty big one, so I wonder if that was enough to cause the failure. 

Regards,

Nick

Hi Nick,

Yes we have seen that we can repeat this failre more than once.

As you can see from the following picture if I increase the input voltage, the output follows, so it is no longer regulating.

Then, if I reset the power supply the output does not go to 4.3V, but it goes to 4.8V.

Talking about the layout it is a 4 layers board.

TLV70433 (U3 in the pictures below) input, output and GND are connected to different inner layers through some vias. This inner layers have extended copper areas.

Regards,

Mattia

Hi Mattia,

To me this seems like a thermal issue based on your layout. Have you tried putting the device into thermal shutdown by leaving it in current limit? If it does reach thermal shutdown in this scenario, I would expect repeated short conditions to eventually get it there as well, and as I described above, the initial power surge when a short is applied may be enough to push its junction temperature high enough to damage it since the junction temperature between shorting events would already be close to the thermal shutdown point. Do you the option of revising your board to add more GND copper around the device? I see a lot of empty space that could be filled in with more copper. 

Regards,

Nick

Hi Mattia,

Sure thing. If you can afford the space, it would be good to also add more thermal vias to the new top-layer copper. 

Regards,

Nick