https://devzone.nordicsemi.com/f/nordic-q-a/10016/no-hard-fault-with-softdevice-and-ram-off-after-wfiHi Folks, i have implementet a sleep mode by using only the RTC timer with the 32KHz OSC and the WFI instruction. The current consumtion is around 1.7&#181;A (all peripherals including the RAM and HSXOSC is turned off). On timeout, the chip wake up anden-USTelligent Community 13Wed, 11 Nov 2015 09:28:10 GMThttps://devzone.nordicsemi.com/thread/37112?ContentTypeID=1Wed, 11 Nov 2015 09:28:10 GMT137ad170-7792-4731-bb38-c0d22fbe4515:8f295e3a-1778-48d0-8c6a-7ca0c4a059c4AD1170<p>Meanwhile i have fond a solution using the watch dog instead of RTC timer. This is not a pretty solution but it works for the first time.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37111?ContentTypeID=1Fri, 06 Nov 2015 06:54:12 GMT137ad170-7792-4731-bb38-c0d22fbe4515:d1b8cbbe-78e6-49ab-8f35-1ffab3eb899dMartijn<p>correct, but to be able to forward any interrupts it needs some ram to do so.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37110?ContentTypeID=1Thu, 05 Nov 2015 07:46:47 GMT137ad170-7792-4731-bb38-c0d22fbe4515:5fc3b65b-d747-4826-9e37-f03f11e46720AD1170<p>But the softdevice is disabled and the interrupts should forwarded to the application.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37109?ContentTypeID=1Thu, 05 Nov 2015 06:59:40 GMT137ad170-7792-4731-bb38-c0d22fbe4515:e7f50acc-6678-4989-b461-f4902261cbb5Martijn<p>Then you answered your own question, right? For the softdevice to forward any irq handler (including hardfault) you need stack space.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37108?ContentTypeID=1Wed, 04 Nov 2015 21:20:43 GMT137ad170-7792-4731-bb38-c0d22fbe4515:50dc222e-53bb-4d95-802e-b38eab0a5f2cAD1170<p>If the controller awake (from WFI), the next statement from flash is executed or the interrupt handler address, in this case the RTC1 interrupt vector, is fetched from the vector table. In this moment, you are right, there is no (stack-) RAM to save the PC. So i think the hard fault is the last way out and is called without saving any data on stack. I think, the problem is the changed vector table when soft device is flashed on the chip.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37107?ContentTypeID=1Wed, 04 Nov 2015 20:54:31 GMT137ad170-7792-4731-bb38-c0d22fbe4515:a424a8ae-68ed-4665-860e-b317baa2de57AD1170<p>... sorry, it was a mistake. Can you undo it?</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37106?ContentTypeID=1Wed, 04 Nov 2015 20:40:55 GMT137ad170-7792-4731-bb38-c0d22fbe4515:512c31ed-b121-4c05-bb33-a91d1e84f268Martijn<p>How is it possible to call hardfault handler without ram enabled? You need stack space, right?</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37105?ContentTypeID=1Tue, 03 Nov 2015 11:19:42 GMT137ad170-7792-4731-bb38-c0d22fbe4515:3a7f6cc1-dc72-4297-8d6d-6ae9ce1f884eSusheel Nuguru<p>Hardfault is first caught inside the softdevice and then forwarded to the app if they were not caused by softdevice. I am not sure why the generated hardfault in your app is not forwarded.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37104?ContentTypeID=1Sat, 31 Oct 2015 22:57:20 GMT137ad170-7792-4731-bb38-c0d22fbe4515:6dae41c0-6a0f-44e6-8646-b8074f8d9c3bAD1170<p>I think, the hardfault is a result of the &quot;un-powered&quot; RAM after the wake up from WFI. So the hardfault is not desired but an secure entry point after the sleep mode.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/37103?ContentTypeID=1Sat, 31 Oct 2015 21:47:21 GMT137ad170-7792-4731-bb38-c0d22fbe4515:540a8a40-5d21-4690-9939-5b3ebba8ae66Susheel Nuguru<p>why do you need to generate a hardfault to call NVIC_SystemReset? How do you generate hardfault as there are many ways to do it?</p><div style="clear:both;"></div>