LMZ10504: Cause of LMZ10504 failure

Hi,

Yes, the shorting of Vout to GND is one failure mode of a buck converter.  It usually comes from an overvoltage at the input or poor PCB layout.

Can you share the part numbers of C408 and C415?  Can you share the PCB layout?

The TPSM82816 is a newer, smaller device, and is recommended.

Chris

Thank you for your continued support.

Thank you for contacting us.
The part numbers for C408 and C415 are as follows.

・C408
Model number: GRM31CR61E476ME44K
Manufacturer: MURATA

・C415
Model number: GRM21BR61E106KA73K
Manufacturer: MURATA

I am attaching the PCB layout around U117.
The layer on which the U117 components are mounted is layer 1.
In order to show the location of the components, silk is shown even for the inner layers (layers 2 to 7).

I would appreciate it if you could explain the problem.LMZ10504_LAYOUT_20241224.pdf

Thank you for providing the part numbers and layout.  The capacitors and the layout look ok.

Is only 1 device failed or have several failed?  What testing was going on when the failure was detected?

How is the 5V input supply provided?

Chris

Thank you for your continued support.

Thank you for checking the part number and layout.

Only one device is faulty.
When a fault is detected, the device is constantly powered (for more than 24 hours).

Approximately two months later, I confirmed that the U117-VOUT pin was shorted.

I am attaching the overall circuit diagram around U117.

The +5V power supply is supplied from the following pins of the connector (J1) shown as Power Supply Connector in the attached circuit diagram.

・J1-1 pin (+5V supply)

・J1-2 pin (GND)

The following power supply unit is used to supply +5V.

・Model number: LFS50A-5

・Manufacturer: Daitron

I would appreciate it if you could explain the problem.U117_entire_peripheral_circuit.pdf

Thank you for explaining.  

Were there any power surges or power outages during that 2 months?

From the specification of that power supply, it looks like its 5V output voltage can go up to 6V, even in a normal condition.  This is above the 5.5V maximum operating range of the LMZ10504 and is therefore not recommended.  https://daitronglobal.com/products/low-noise-power-supplies/lfs50a/

Chris

Thank you for your continued support.
Thank you for your opinion.

I contacted the manufacturer (Daitron) about the LFS50A-5, and they replied that when used with +5V output, the output voltage may change by 10mV even if there is a change in the input voltage or load fluctuation, but it will not exceed +5.5V. The 6V specification is the maximum value that can be changed using the included volume.

Also, when I first asked the question, I stated that the condition was 1A, but when I actually measured the current at the U117-VOUT pin on a board with the same design as the faulty board, it was 0.226A. I know from the datasheet that DC-DC converters become less efficient when the output current is low, but is there any relationship between this failure and the low output current (current from the U117-VOUT pin)?

I would appreciate it if you could explain the problem.

Hi,

I don't understand their statement that "The 6V specification is the maximum value that can be changed using the included volume."  Their spec is 6V maximum.

No, a lower load current is not an issue.

Failures of a device during its operation in the field are very difficult to debug, since we don't know when they failed or what was going on when they failed.

Chris

Thank you for your continued support.

Thank you for your advice on load current.

I performed aging while monitoring the VIN pin of U117 using an oscilloscope.

As a result, I observed a spike waveform of over 6V.

I investigated the source of this high spike voltage and confirmed that a voltage higher than the voltage monitored at the VIN pin of U117 appeared at the same time at pin 1 (+5V power input) of the J1 connector.

Please see the attached file.

So, please confirm.

As you can see in the attached file, the waveform occurs in a very short period of time, but could the U117 be damaged by such a spike voltage being input to the VIN pin of U117?

I would appreciate it if you could explain the problem.U117_VIN_pin_overvoltage_check_20250109.pdf

Thank you for sending the waveforms.

I'm surprised something this short in duration gets through the FL5 filtering device.

Were these waveforms measured with a low inductance probe technique, like this? 

Whether you are able to capture an over-voltage waveform or not, a Vin over-voltage will always be a possible source of the failure.  We don't know what happened to the failed device while it was in the field.

Chris

Thank you for your continued support.

Thank you for your insight regarding the possible malfunction.

There was a mistake in the description of probe Ch2.

The description of probe Ch2 in the waveform data (pdf) I sent is actually as follows.

・Error: Between pins 1 and 2 of J1 (+5V power input)

・Correct: Between pin 1 of J1 and TP15 (GND) (+5V power input)

I apologize for the inconvenience.

I have attached the waveform data with the corrected description.

I have attached the details of the probe connection in this case.

I would appreciate it if you could explain the problem.U117_VIN_pin_overvoltage_check_20250110.pdfU117_VIN_pin_overvoltage_check(prove).pdf

That you for sharing the probe method used.  That is not a low inductance method.  It will pick up noise from the surrounding environment.  A low inductance probe tip looks like this and wraps around the barrel of the probe: 

Chris