https://devzone.nordicsemi.com/f/nordic-q-a/64834/softdevice-and-timer3nRF52840, SDK16.0.0, S132v7.0.1 I&#39;m checking the accuracy of an LFXO 32KHz oscillator for manufacturing test purposes. My rough plan is: Initialize and clear TIMER3 (HFXO) Set the TIMER3 comparison value to the duration of my test, converteden-USTelligent Community 13Mon, 17 Aug 2020 09:27:52 GMThttps://devzone.nordicsemi.com/thread/264806?ContentTypeID=1Mon, 17 Aug 2020 09:27:52 GMT137ad170-7792-4731-bb38-c0d22fbe4515:68a4f570-3c71-4d98-9e4f-3cf5e8f779cfEdvin<p>That is correct. The softdevice always uses the highest priority, so it may be that your TIMER3 events are delayed if the softdevice is doing something else.&nbsp;</p> <p>It is also a good idea to use PPI for timing critical measurements, as this doesn&#39;t require the CPU to run.</p> <p></p> <p>If you only intend to test the XTAL, the easiest way to do so would probably be to not use the softdevice at all. But you can use PPI. I don&#39;t know if you plan to do this on all devices during the runtime of your product, or only once (now) before production.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/264674?ContentTypeID=1Fri, 14 Aug 2020 19:18:15 GMT137ad170-7792-4731-bb38-c0d22fbe4515:85707b86-cd18-41a2-bf48-0b972b06962bDmitry<p>Hi Charles,</p> <p>SoftDevice has a highest priority level, so your interrupt handler can be delayed. The best way is to do sampling by hardware: set CC[0] and CC[1] of&nbsp;rtc to measure some distinct interval and connect COMPARE[0,1] events of rtc to CAPTURE[0,1] tasks of&nbsp;timer via PPI channels - you will get two time values with&nbsp;62.5 ns&nbsp;resolution.</p><div style="clear:both;"></div>https://devzone.nordicsemi.com/thread/264671?ContentTypeID=1Fri, 14 Aug 2020 18:35:58 GMT137ad170-7792-4731-bb38-c0d22fbe4515:ddf5d29c-8995-4a2a-8ea3-bc6e536c8e88charles<p>Alternately if there&#39;s a more stable way to do this, I&#39;d love any advice!</p><div style="clear:both;"></div>