LM5116 PREMATURE FAILURE

Attached are the waveforms you requested

Hi 

The waveform clearly shows high-side switching is not proper. And also I could see your input voltage is >80V. 

Absolute max of SW pin is 100V. Please check if the voltage spike at SW pin is greater than 100V. 

LM5116 need a high quality bypass capacitor close to the VIN pin. Please populate one of the ceramic input capacitor close to the IC.

Schematic wise, I could not find any problem. The issue might be related with your layout or just a wrong component population. 

Regards 

1.per the spec sheet VCCX should be grounded if not used, how critical is this ?

2. we are thinking about adding a voltage clamping device to VIN, due you have any suggested devices that will react fast enough ?

Hi 

If you would like to ask a Failure Analysis. Please contact your local FAE and ask him to send PQA. 

Regards,

Do you have any suggestions on the following 2 questions?

1.per the spec sheet VCCX should be grounded if not used, how critical is this, as we currently do not have it connected?

2. we are thinking about adding a voltage clamping device to VIN, due you have any suggested devices that will react fast enough ?

Thanks

Hello Shannon and Eric.

I have also seen this failure mode before, in fact, I'm investigating an instance of it right now. Have you discovered the root cause of the problem? In my application, several devices are powered through electromechanical relays. Typically, the failure occurs when the relays are opened. Turning off a relay does not produce any observed voltage transients on the supply output or switch node. Is your experience/configuration similar? Eric, can you provide comments on Shannon's last post? It would be helpful to information to anyone designing with the LM5116 controller at higher voltages.

Thank you.

Hello Larry,

I believe at this time I have corrected my issue but it was multiply changes that I made and I'm not sure which of the changes fixed the problem. But if you refer to the schematic that I posted at the top of this chain (4104_001.pdf) I will run through the changes made. First thing that was added was the addition of a .1uF cap on the VIN (pin 1) places just above the driver package and grounded to the ground plane that I have under the driver IC. Second change was the snubber resistors (R110,R111) for the low side FET (Q9), I moved the resistor traces from the source of Q9 and connected them to the ground side of the current sense resistor (R109). Third change was I had the DEMB (pin 11) connected to the ground plane under the driver IC, I have now connected that pin directly to the CSG (pin 13) which is dedicated trace running to the ground side of the current sense resistor (R109). Forth change is I have connected VCCX (pin 17) to the ground plane. Fifth change was to slow down the startup of the driver IC, this includes changing the SS (pin 7) cap to a .1uF from the .022uF, adding a 47uf cap from the UVLO (pin 2) to ground, adding a S1M diode from the UVLO (pin 2) to the +74V net (VIN pin 1) anode on pin 2 cathode on pin 1. I did start with just adding the .1uF cap to the VIN pin but that did not fix the problem. All of the other changes I did at the same time so again not sure which change fixed the problem or if all were needed to fix the problem. That's the run down of what I did so hopefully you find something in here that will help with your problem. Best of luck to you.

Shannon