TPS630242: TPS630242 Design Not Getting Up To 3.3V Output

Hi Kyle,

the schematic looks ok so far.

Do you start into a load?

regards,

Moritz

Hi Moritz,

I've tried starting into a load and starting with no load. Both give the same result.

Hi Kyle,

just wanted to make sure that you don't start into a short or some sort of overload condition.

Please also post your layout.

Is this the only device/PCB you tried this with? Do you have the option to try another unit?

regards,

Moritz

Hi Moritz,

I had 5 PCBs fabricated and the same issue exists on all 5.

Here are screenshots of the layout. There are 6mil vias in the pads of B1, B2, B3 and D1, D2, D3. I can provide Gerbers if that is easier for you. I just prefer not to post the complete Gerber's in public.

Thanks again for your assistance!

Layout Screenshots.zip

Hi Kyle,

ok, stupid question. is the orientation of the device correct?

regards,

Moritz

Hi Moritz,

Yes. I have confirmed that the orientation of the device is correct. The only other possible cause that I can think of would be an assembly issue with the BGA.

Thanks,

Kyle

Hi Kyle,

do you have an unassambled PCB at hand, which you could use to check all the traces from the device?

If not, I would suggest to unsolder one unit and to check all the ball pads for proper connection and shorts.

Regards,

Moritz

Hi Moritz,

I removed the switching regulator and ohm'd out the pads. All connections were as expected with no shorts.

Thanks,

Kyle

Hi Kyle,

ok, so far so good.

Can you verify that you are using a TPS630242, not a TPS63024?

Also can you please measure the signals at L1 and L2.

Thanks,

Moritz

Hi Moritz,

I have checked the spare components sent by the assembler and can confirm the part is a TPS630242YFFT.

Here are the scope captures at L1 and L2. There appears to be no signal at L2.

L1 Measurements.tif

L2 Measurements.tif

Thanks,

Kyle

Hi Kyle,

thank you for the scope plot. The device is switching and in buck mode, that is why you only see VOUT at L2.

However L1 is not correct. Where do you connect the GND of your probe?

How is your grounding done?

Can you please measure VIN as close as possible to the device.

regards,

Moritz

Hi Moritz,

The GND of my probe is connected to a common ground post for the entire board. The board does not have separate grounds (signal ground, power ground, etc.). The board has 6 layers and has a ground flood on the top and bottom layers. The board also has a dedicated ground plane on layers 2 and 5.

I measured VIN close to the device (at C39) and am getting ~4.0V, which matches the readout of my DC power supply.

Vin Measurement.tif

Thanks,

Kyle

Hi Kyle,

thanks for the scope plot.

in meas of C1 it shows 3.7 and 4.5V, which is not a stabel 4V supply.

Can you please redo the measurement with a better vertical and horizontal resolution, i.e. 1V/div and 10us. maybe even put a trigger at the falling edge with a level of 3.9V to see if you catch a dip.

thanks,

Moritz

Hi Moritz,

It appears you are correct. My circuit is designed so that the switching regulator can be jumpered in and out of the circuit.

When 4V is applied to the circuit and there is no power applied to the input of the 3.3V switching regulator, the power supply is very stable. I did not get any edge triggers when the trigger was set at 3.9V.

However, when 4V is applied to the circuit and there is power applied to the input of the 3.3V switching regulator, the power supply becomes unstable. I am constantly getting edge triggers when the trigger is set at 3.9V.

6330.Supply with Switcher Powered.tif

1588.Supply without Switcher Powered.tif

Hi Kyle,

thanks for the measurement.

Can you please remeasure L1 in combination with the input current (current probe on the input cable is fine, trigger on rising edge of L1 switch pulse), but the timebase should max be 1us and max 2V/div so we can see the switch pulses.

another question, you said you ohm'ed out the board. Did you measure L2 to GND with no device soldered?

Btw, do you have a current limit/clamp active in your power supply?

regards,

Moritz

Hi Moritz,

Here's a capture of the L1 in combination with the input current.

L1 with Current.tif

There is a current limit on the power supply, but the current limit is not being hit.

I believe you have solved it, though. I ohm'ed out the board with no switcher soldered and everything was as expected. However, when I ohm'ed out a board with the switcher assembled, I'm seeing a short between L2 and GND. I'm assuming this is due to an assembly issue at the board manufacturer.

Thanks for your help!

Kyle

Thanks for all of your help, Moritz!