Shunt capacitance spec for LFXO in nRF52840

[DBT]

The table under 5.4.4.3 in the datasheet for the nRF52840 lists the shunt capacitance for the LFXO as 2pF max

C0_LFXO (Shunt capacitance) 2pF max

What is the risk if the shunt capacitance of the crystal I'm using is larger than 2pF?  

Is there a risk that the crystal won't start up at all?

That it will take longer to start up?

That it won't be as consistent or accurate?

The crystal I'm using has a shunt capacitance of 1.2pF typical but 5pF max. 

I hadn't noticed before because this spec is not in the datasheet for the crystal. 

I only found out when I asked specifically for it.

[Andreas]

Hi,

 

Difficult to say, I do not believe it has been characterized too widely with higher than 2 pF shunt capacitance. What is the part number of the crystal and what is the ESR?

 

Best regards,

Andreas

[DBT]

The crystal is ABS05-32.768KHZ-T (Abracon) and the ESR is 90k ohm max.

[Andreas]

Hi,

 

The ESR spec also is on the edge of what is characterized. If outside spec there is an increasing risk that the start-up time will be prolonged and outside spec, or that the crystal might not start at all, but as this has not been characterized we do not have any fixed boundary value other than the 2pF spec.

I did look up the part number internally, there are a couple of hits but no reports that it has caused any failures. According to our designers also it is probably fine, but as it exceeds the spec we can not make any promises.

It is up to you whether or not you want to take the risk, unfortunately as this is outside spec we can not guarantee anything. If my somewhat limited understanding of crystal oscillators is correct you could do some experimenting by soldering a capacitor between the crystal leads. If the 32 kHz oscillator does not fail before >>5pF I would expect a worst-case crystal to also be fine.

 

Best regards,

Andreas

[DBT]

FYI per your suggestion I tried putting extra external capacitance in parallel with the crystal in order to model a higher shunt capacitance.

The limit appeared to be between 10pF and 20pF of extra capacitance.  

It seemed to work fine with 10pF but when I tried 20pF it looks like the LFXO initially started oscillating but didn't continue oscillating for more than a second or two.