MCU and USB Section Powering

[rho-bot]

Looks pretty good, I think you will pass ESD now :thumbsup:

One last hint you might think about:
The 5V BuckSMPS is connected to GND
In HOST mode it sources USB, but the backcurrent has to go via CPU starpoint 8-O ...
Maybe not the best idea in this mode. I would suggest to add an external starpoint below the CPU, maybe configurable as 0R.

 

[tiwari.sachin]

I have connected output of LM3525 to USBVss (ie output cap at OUT Pin) while the control signal ST1 and EN to GND (pull up and pulldown to MCU power)

There is a possibility that I might use the ID pin to detect USB/HOST and run code either for HOST or device depending of ID pin status

- - - Updated - - -

I have used a 0Ohms. Thanks for pointing that out.



I have kept this 0R close to PIN 3 of U16

 

[rho-bot]

:thumbsup:
maybe a little academic (don't know your layout), but maybe R61 is better located at the other starpoint (CPU), to prevent GND-loops.

Good luck & give us a feedback if the design passed the ESD ;-)

 

[tiwari.sachin]

I have a GND layer and I decided to connect the 0R to that GND via via directly. This shouldnt create any ground loop. Else I really had to take that track a long way and that might have created that loop.

I will surely post the results. I have sent the board for fab.

 

[tiwari.sachin]

rhoo-bot

Just a thought

In case i use a varistor also in parallel to the TVS, Would it by any ways help.



It could be USB lines (VBus only) or for that matter adaptor Vcc

 

[rho-bot]

I don't see much benefit adding a varistor in parallel.
Varistors:
+ may consume more energy
- slower response time
- less precise voltage threshold
- bigger capacitance
- higher aging effect
So Varistor may be added in parallel to supply lines, if you expect surges with higher energy than TVS can handle.
For Data lines it's a NoGo, due to the capacitance (but you already stated not to use them on the USB data lines