LF/HF XTAL (ppm) tolerances vs performance

[dav01]

Hi, would like to understand what scenarios a clock stability better than +/-40ppm would be needed in a BLE setup.

i.e. channel sounding, and similar. I mean you can get a pretty reasonable price for a 10ppm or even down to 5ppm XTAL or TCXO nowdays.

[Asbjørn]

Hello David,

You're right, the clock accuracy of +/-40ppm is good enough for most applications and usually also come with a built in margin. A more accurate crystal might give you some benefits in noisy environment as out of band noise could become less. Although for BLE you often only have control of one side of the link, so you don't necessarily know what the other end have installed.

There's drift and degradation across the lifetime of the crystal and related components and this is usually the deciding factor. Some products are designed to meet larger temperature variations and life. For these products it can be useful to have a tighter spec crystal that across time and temperature still is within the requirement.

The channel sounding profile could benefit from a tighter crystal, but if that level of accuracy is desired, you often end up with an application with more HW as well (antenna and supporting circuitry) which makes the additional cost of a tighter crystal far smaller in the overall cost of the application.

Best regards

Asbjørn

[dav01]

hi asbjörn, can you put a number on "tighter crystal" for channel sounding

[Asbjørn]

Hello David,

I was referring to the extra cost of the 10pm and 5ppm  crystal you mentioned, that the extra cost of these types would be easier to handle for a full channel sounding type application as the cost for the additional HW would be higher so the overall extra cost of a more expensive crystal would be smaller percantage wise in a BOM.

That said,there's no need for the 10ppm or 5ppm for the channel sounding profiles to work, so the 40ppm will be ok.

Best regards

Asbjørn

[dav01]

how would the timing difference between two nrf54's be affected by worst case tolerance at 2x 40ppm vs 2x 10ppm (channel sounding or aoa that is)

[Asbjørn]

Hi David,

The maximum timing deviation between them for a 32MHz crystal would be along the lines of

40ppm is 40us deviation or 1280 clock cycles

10ppm is 10us deviation or 320 clock cycles

What sort of distance measurements and accuracy are you looking?

The channel sounding and direction finding are combining different techniques, I don't know of you had seen the video from Bluetooth SIG on this topic: https://www.youtube.com/watch?v=U9GAK2AsJi0 it might be interesting for you to look at.

Best regards

Asbjørn