MK66 unusual power-up issue

[emir1396]

Hello, I'm using a board with MK66FX1M0VLQ18 on it and I've encountered a weird issue.
If I use the MCU, and then cycle the power OFF-ON it sometimes just doesn't run.

I've managed to conclude that this happens only in the time frame of about 0.8s - 2s after OFF.
If I cycle it "instantly" OFF-ON it works, if I give it >= 3 seconds like OFF-3s-ON it works.
If I cycle the power like OFF-1s-ON it NEVER runs.

The EzPort function is disabled, the NMI interrupts are disabled. I've checked the supply and reset pins with the oscilloscope and nothing is different than in the case when it works correctly. I've tested it with a code that does nothing except turn on a LED and it's still the same.

Can someone point me in the right direction of trying to figure this out?

Thanks.

[vicentegomez]

is this also happed on the freedom board ?

Please check if the reset pin and all the VDD went down 

[emir1396]

I don't have a freedom board to test this on.

As I've written in the original post, I haven't seen anything unusual on the reset and supply pins when probing them with the oscilloscope.
It makes it even weirder that the immediate reset doesn't produce a bug. So what you're asking, if the pins went down, doesn't produce a bug. Only when you wait a little longer, but not too much, if that makes sense.

[RaRo]

Hello @emir1396,

How are you powering your board?

Some solutions or ideas: You could try the SDK's examples, to see if it is not code related. Also, it might be useful try with a mass erase before going on with another program or example.

While that behavior is unexpected, as you are dealing with a custom board, we cannot assure the problem could be related to the MCU itself. That's why we asked you to try with the FRDM-K66, to compare the behaviors both boards could have.

Best regards, Raul.

[emir1396]

Hi @RaRo,

The board is powered via USB, with an on-board buck converter generating the 3.3V supply for the MCU (VDD on the screenshots).

I’ve tried multiple versions of the code, with and without using external crystal, doing nothing but turning on an LED, disabling EzPORT and NMI, all with the same results. I feel quite certain that this bug is not code related.

I will include oscilloscope screenshots now that illustrate the bug. There are 5 screenshots of the whole cycle with various power cycle delays, and also three that are zoomed into the rising edge of VDD and RESET while powering up in 3 cases of the delay value. XTAL is used as a visual indicator of the MCU running, but as I’ve said, with or without using it, the results are the same.

If it helps, I’ve found that adding a dummy load to the VDD line makes reproducing the bug a lot harder, the delay in which it happens shortens. That makes me think that the delay time is not what is relevant here but the voltage levels on some of the pins (example: the MCU doesn't run every time the VDD start rising from ~450mV level, lower than that and it's fine, higher than that and it's fine)? If you think I can capture something else that would be useful to you let me know.

Thanks,

Emir.

[RaRo]

Hello @emir1396,

It might be good to check the voltages, because as you mention, that could be the reason behind the issue. Please check the VDD and Reset voltages and compare them to the Kinetis K66 datasheet specifications.

Also, just for double check, please take a look to your buck converter, here you may find some cues worth reviewing.

Best regards, Raul.