LM63635-Q1: ICs doesn't always start

Hello

I assume you mean that 1 out of 30 don't power up ??

What is you input voltage ??

What is the input voltage slew rate in your application ??

What kind of load are you using ??  Constant current ??  Resistive ?? 

How much load current during start-up ??

Can we see some images of your PCB layout ??

I will assign this post to the engineer in charge of E2E for this month.

Thanks

Hi,

-yes the failure to start is rare .

-I use the LM63635-Q1 to convert 15V to 5V.

-input starts in a few ms (under 5ms):

here picture when ok

picture when nok:

-it supplies a board with different kind of ICs, max few hundred mA (<300mA)

- I was thinking about a bad soldering on the gnd pad under the chip, but it seems to be ok.

Thanks

Hello,

Thanks for sharing this information. 

Is the switching frequency 315kHz when the device is able to properly regulate the output to 5V? Can you share the same start up images with SW voltage included as well? I'd also like to see zoomed in SW node voltage waveforms (including Vout) similar to Figure 8-7 in the datasheet. 

Does the PG5V rail assert high when the device output only goes to 0.985V?

In terms of the board layout, what are these components (below)? I don't see them in the schematic you sent:

Regards,

Harrison

Hi,

1/Is the switching frequency 315kHz when the device is able to properly regulate the output to 5V: yes, when the 5V is 0.985V, the switching doesn't start, I think the 0.985V is supplied by the FB pin.

2/Can you share the same start up images with SW voltage included as well: unfortunately I don't have the board with me, next week maybe.

3/Does the PG5V rail assert high when the device output only goes to 0.985V?:  0V when fault

4/In terms of the board layout, what are these components (below)? I don't see them in the schematic you sent:: 649R + led

Thanks

Hello,

In terms of the switching frequency, the lower frequency you are measuring may be due to the device operating in PFM which is expected considering your maximum load current. Please share the switching waveforms so I can confirm. 

Have you performed an ABA swap by swapping a 'faulty' IC onto a known good board and a known good IC onto the 'faulty' board? 

Performing this swap will help determine if the device is at fault or if it is something else on the board. 

Regards,

Harrison Overturf 

Hello,

I don't think it was working in PFM mode, the switching was constant, you can see on the picture, it's a zoom, f=314khz, duty cycle 1/3 ( 15V to 5V). Other board ok have 413khz frequency.

The swap will be done next week, waiting new ICs.

I suspect a bad batch or some boards were impacted by ESD.

Regards,

Clement

Hello Clement,

It looks like the part is operating in PWM, thanks for confirming that. Please confirm my  understanding:

Some boards switch at 315kHz and are able to regulate the output to 5V. Some boards switch at 413kHz and also are able to regulate the output to 5V. For the boards whose output voltage only comes up to 0.985V, these parts never begin switching. Can you please confirm? Also I would still like to see the switch node of the 0.985V output devices at start up. 

Please perform the part swap and let me know the outcome of testing. 

If you have access to the device EVM you can also try to place one of the faulty parts on the EVM and see whether the same behavior follows the device. 

Regards,

Harrison Overturf 

Hi Harrison,

Most of the boards regulate 5V and switch at 413kHz.

Some boards regulate the 5V but switch at 315khz but sometimes these boards don't even start to switch and 5V increase slowly to 0.985V. When it doesn't start the switching node has the same shape than the output .

No I don't have the EVM but the swapping is in progress.

Regards,

Clement

Hi Clement, 

Ok thanks for confirming. When the fault occurs Vcc=5V correct? How does BOOT to GND voltage compare to SW to GND voltage during the fault condition? 

Please let me know the outcome of the swapping testing. 

Regards,

Harrison Overturf 

Hello Harrison,

After swapping with a new IC ,switching frequency is 406khz, and it works fine. So I don't know if it's the batch which has a problem or something went wrong during fabrication/assembly.

Regards,

Clement

Hello Clement, 

You can try testing the bad devices with a multimeter set to the diode mode to see if there is possible ESD damage to some pins. Place the red positive lead on the IC PGND pin then use the black negative lead to probe the other pins. Each pin (besides the NC pin) should have a diode drop when measured this way. Try this with a good IC and a bad IC to compare the results. 

Regards,

Harrison Overturf