I made a PCB using an LTM4646 chip to generate 1.5V and 3.3V from a 12V supply.
All three of the power nets, 12V, 3.3V and 1.5V are shorted to the GND net in the finished board.The PCB manufacturer checked them out with an X-Ray to verify the issue was not with the manufacturing of the board, so I wanted to see if my design looked correct to this community.
Here is my setup in LTSPice, which is the design I used on my PCB. It simulated the correct voltages just fine.
This LTspice circuit looks fine. The schematic and layout files should be checked to see if there is issue causing the short.
Thanks for taking a look.
Believe me, we have checked the schematic and layout several times. =)
There are a couple of things I had suspected might be wrong with this setup, so I'd like to run them past you.
(We didn't notice these things in the datasheet until after we started having problems.)
Do you think either of those issues might cause this GND short we are seeing?
You are right about these two items. Missing DRVcc cap should not cause the short. While RUN pin max is 6V, connecting it to 12V will possibly damage the internal IC and it might cause the short.
I just wanted to follow up on this.
The reason the nets were shorted was that R3 and R6 on my PCB were milliohm values. They were defined with the lower-case 'm' in the LTSPICE demo file, and so I ordered them that way. After I realized those resistors were creating a rather effective short, I removed them from the PCB. LTSpice lets you use a lowercase 'm' for a Megohm resistor. Lesson learned.
To avoid damaging the chip due to the 6V RUN pins, I also lowered the input voltage of the shown setup to 5.5 VDC. At 12VDC I was blowing a 2 amp fuse on my PCB.