Riotboard startup issue

I have one board that didn't respond to this treatment like I would like.  Any ideas what would cause a non start?  When this happens, I am fairly confident there is no activity on the debug port.

Interestingly, on one board that did it here, whenever the ftdi debug adapter is connected, this won't happen, apparently there is enough current fed from the ftdi pullups that this doesn't occur?

1. Can you check the voltage on the board's UART RX pin? If the board's UART RX pin is not pulled high, then the CPU receives some garbage which might be taken as valid data ... Try to pull-up the UART RX pin (a 10K resistor should work fine)

2. If the above method still doesn't make any difference, could you try out a different power supply? I got weird things going on different Linux boards whenever the power supply couldn't meet the board's power needs. When I see these kind of things happening, I am always checking the power supply. If you have a good quality power supply (e.g. lab power supply) you can try powering your board with it and see if it makes any difference

#1.  Pulled up to what?  I tried grounding this pin without positive results, as pulling this up isn't really applicable, rx and tx are already pulled up to vdd_nvcc per schematic.  Like as mentioned, pulling up externally allows proper boot, but this is more likely due to external current feed rather than uart garbage.

#2.  I tried this on my bench supply with the same results.  The board boots at 3 volts if you go slow with it.

See these scope shots with the regular power supply(1.5 amp triad wsu050-1500 Blue is 5v power red is PORB Second image has power not completely drained

I feel like it is a power sequencing issue.  I also scoped the tx debug pin for action when the problem happens, and there is none.  Out of some 12 boards here, this one is fairly repeatable with the problem.  It only happens when the board has been powered down a few minutes.

Again, note that with the particular board described here, the 1uf cap on d40 fixes it.

I have another board offsite exhibiting some slightly different issues, perhaps related, where it doesn't boot correctly.

1. Sorry. My mistake. I forgot the UART pins are already pulled high to VDD_NVCC

2. Since the issue is the same when using the bench power supply, it seems that indeed there is a POR issue here.

One solution (that would somehow cope with the effect) would be to increase either the capacitance of the C331 or the resistance of R219. The root cause will still be present.

a. What about the main oscillator? Can you check if its start-up time between a good and a faulty board is somehow different?

b. Another thing to check might be the soldering for the decoupling caps.

If none of these applies, then it might be something faulty with the CPU...

I feel like I checked the oscillator when this thing was "browned" out and it was in operation, although this doesn't directly answer your question.

Adding a cap to D40 effectively adds capacitance to C331.

Is it possible for the chip to boot before the boot configuration pull ups are fully pulled up?  Note that putting the ftdi converter on would have the effect of pulling these up prior to boot.

 

Is it possible for the chip to boot before the boot configuration pull ups are fully pulled up?  Note that putting the ftdi converter on would have the effect of pulling these up prior to boot.

I wish I could help but this is a good question for Freescale support. They might have some ideas about these symptoms.

Who supports these boards, freescale or embest? 

The email on the riotboard site bounces..  RIoTboarders@riotboard.org

in this case "embest".

contact them here:

http://www.embest-tech.com/contact-us.html

in this case "embest".

contact them here:

http://www.embest-tech.com/contact-us.html