OMAP and TPS65950

Probably beyond the scope of this list and primarily directed to Gerald
Cooley, but here is the question/problem.

We have used the Beagle Board schematics as a reference and designed and
build our own OMAP based board. We are having a problem with the
TPS65950 powering up it's on board power supplies. VIO, VDD2 and VPLL1
all come up fine and then as the first pulse on VDD1 is sent, the 65950
resets and starts over again. We're not looking for an answer and with
such little information in the email, we don't expect one, but, any
clues or possible direction to take in diagnosing this problem would be
greatly appreciated.

Thanks,
Ross

I have never seen that before. Could be a layout issue or a soldermask issue under the processor. Send me you schematic and I can take a quick look from that standpoint to see if anything looks off.

Gerald

I have never seen that before. Could be a layout issue or a soldermask issue under the processor. Send me you schematic and I can take a quick look from that standpoint to see if anything looks off.

We have had a very similar issue in our first GTA04 boards and had to do a new layout.

Carefully study and follow the layout guide of the TPS65950!

The key issue is if the traces are too small for e.g. VDD1 the rise time is too
slow and the TPS65950 thinks this must be a short circuit, although it isn’t
(but you may still have one! please check).

It was proposed on e2e to attach an external 1.2V supply to “bootstrap”
the converter. This worked for us. Therefore, we did attach a 220 Ohm
resistor (you may need to experiment) between VDD1 and VDD2 to have
at least one board that worked - until the new layout was done (after
really carefully studying and following the layout guide).

Here are other hints:

http://e2e.ti.com/support/power_management/pmu/f/43/p/849/3027.aspx

Thanks Dr. Schaller.

Using your suggestion, we were able to bootstrap VDD1 with a 33 ohm
resistor from VDD2 into working.

We'll now dig into the layout and verify that it is a trace width issue.
We have a 5mil trace from C14 to D15, a 5mil trace from D15 to D16 and a
10mil trace (approx .2 inches in length) to the 1uH coil.

Thanks again for your suggestion.

Ross