Clarification on Errata [100]

[Ingfu]

Hi, 

I just read the nrf54l15 errata [100] of the nrf54l15. It's actually a huge issue, as we do have an interrupt line on one of the affected pins, that is active low, eg. default high.
Can you elaborate, under which circumstances the described behaviour occurs and how often? 

I think this errata should be on page 1 of the datasheet in bold letters and have an info box on your website, as it is absolutely not expected and can definetely destroy devices if it occurs.

[Carl Richard]

Hello Ingfu!

Thanks for reaching out. I need to ask a bit around internally to check on the frequency of this anomaly happening, but you should account for it regardless. As mentioned in the errata documentation it occurs during power-on-reset and only lasts for about 10 µs, given that nRESET is not asserted. This means that the anomaly is gone long before the MCU has booted properly and the interrupt is configured. The question is how reset is handled in your application as this can prolong the anomaly. Are you using the nRESET functionality? 

I also agree that this erratum should be more clearly communicated in our documentation and I will forward that request.

Best regards,
Carl Richard

[Ingfu]

Thanks for responding.

The issue is, that the interrupt line is actively driven high by a sensor immediately when our board is powered. If the pin on the nrf54l15 is shorted to ground during power-on reset, we have a short between the output and the pin, probably damaging both components permanently.
The datasheet states that all pins are high-z until configured otherwise, so it would be safe to directly connect the interrupt line to any pin.

[Carl Richard]

Hello again!

I understand. That makes it more problematic. I had a talk with some colleagues here and they suggested adding current-limiting resistors on the affected pin in question, given that it's not possible to move the interrupt to another pin. 

Where in the design cycle are you? Would it be possible for you change the hardware/layout?

Best reagrds,
Carl Richard

[hmolesworth]

"Shorted" is a wholly inadequate term for a pretty serious errata; "shorted" implies a direct zero Ohm resistance between pins or between pin and ground. Some designs hold nReset active during shelf storage or shipping to minimise power drain, and some designs have an analogue pin connected directly to a Lithium battery. During shipping with an extremely high ground temperature before takeoff such a design could cause the Lithium battery to catch fire in flight subsequently killing all on board; one hopes I'm exaggerating.

I see several possible explanations, maybe there are more:

1. Incorrect bias of schottky diode rail clamp (resistance unknown, phantom power issue)
2. Incorrect connection of analogue multiplexor (resistance unknown)
3. Incorrect activation of low-drive (Rdson say 130 Ohm)
4. Incorrect activation of high-drive (Rdson say 40 Ohm)
5. Multiple items 1-4 above in parallel (say 5 Ohm or less)

Please provide clarification of the actual mechanism of "shorted" pins, and an indication of whether any other nRF devices have a similar issue.

[Carl Richard]

Hello!

The affected devices are nRF54L05, nRF54L10 and nRF54L15. I will reach out internally to get more details on the nature of the "shorting".

Best regards,
Carl Richard