nRF52832 Advertising Interval

RE: nRF52832 Advertising Interval

John — Fri, 17 Jun 2016 07:02:48 GMT

The 25ms that I'm referring to is the documenenation in your power calculator.

devzone.nordicsemi.com/.../

RE: nRF52832 Advertising Interval

Dave_couling — Mon, 13 Jun 2016 22:14:26 GMT

What do you have your Clock Accuracy set to? With a wide Rx windows combined with a very short Adv Interval, the relative variation of the intervals should be more noticeable.

RE: nRF52832 Advertising Interval

John — Mon, 13 Jun 2016 14:55:42 GMT

Agreed on the reading and the fact that the core specification is a bit inarticulate there.

RE: nRF52832 Advertising Interval

RichieJH — Mon, 13 Jun 2016 14:48:27 GMT

Thanks, I read that as a new random value for each advertisement, but could have been better put I think.

RE: nRF52832 Advertising Interval

John — Mon, 13 Jun 2016 14:37:37 GMT

This is from the 4.2 core specification: "The advDelay is a pseudo-random value with a range of 0 ms to 10 ms generated by the Link Layer for each advertising event."

RE: nRF52832 Advertising Interval

RichieJH — Mon, 13 Jun 2016 14:25:00 GMT

Thanks for the paper John. I'm too lazy to dig deeper and the paper doesn't mention it, but is that randomness set on the first advertisement and doesn't change or is it randomised for each and every advertisement?

RE: nRF52832 Advertising Interval

John — Mon, 13 Jun 2016 14:06:11 GMT

The stack is supposed to add a random delay between 0 and 10mS to each advertising interval to help minimize collisions. There is a good explanation of advertising here:

Advertising explained

RE: nRF52832 Advertising Interval

FormerMember — Mon, 13 Jun 2016 12:12:30 GMT

If the advertising interval is set to 20 ms, the device should advertise in the beginning of each interval each 20 ms.

How do you measure the 25 ms?

How do you set the advertising interval in your code?

RE: nRF52832 Advertising Interval

John — Sat, 11 Jun 2016 20:41:13 GMT

No I understand the "jitter" that you refer to. We have been working with TI and their timing is slightly different...not that it really matters. I'm just curious.

What they do is set the interval say 100ms, advertise at the start of the interval and deduct the time of the advertisement approx 4ms from the 100ms to calculate the sleep current which is then added to the total current over 100ms. What Nordic do is actually add the time of the event to the advertising interval so 100ms becomes 105ms. Just different..no problem working with it....I'm wondering if strictly speaking it conforms to the iBeacon specification with this very slightly different timing.