Questions about Direction finding connectionless beacon setting parameters

[Sho_Ri]

Hello,

I am using the "direction finding connectionless beacon sample" to test various settings of the beacon.
About setting parameters, I have the following two questions:

1.

For the [PER_ADV_EVENT_CTE_COUNT] parameter, according to the description of "BLUETOOTH CORE SPECIFICATION Version 5.1" on P.1475: "The CTE_Count parameter specifies how many packets with a Constant Tone Extension are to be transmitted in each periodic advertising event. If the number of packets that would otherwise be transmitted is less than this, the Controller shall transmit sufficient AUX_CHAIN_IND PDUs with no AdvData to make up the number". In addition, on P.2753, the periodic advertising event is shown in the figure below.

So, when setting PER_ADV_EVENT_CTE_COUNT = 5, will the CTE packets after the first CTE packet be transmitted through [AUX CHAIN IND PDUs]? In addition, is there a fixed time interval between [AUX SYNC IND] and [AUX CHAIN IND], or is there no time interval?

2.

For the parameters "bt_le_adv_param" and "bt_le_per_adv_param", there are items to set the advertising interval. I think "bt_le_adv_param" is to control the extended advertising interval, and "bt_le_per_adv_param" is to control the periodic advertising interval. Will they affect each other?
     EX, set the extended advertising interval to 1s and the periodic advertising interval to 4s. Or set the extended advertising interval to 4s and the periodic advertising interval to 1s.

Also, do you have any recommended settings for the interval about "bt_le_adv_param" and "bt_le_per_adv_param"?

Best regards,

Sho

[Elfving]

Hello Sho, and thank you for your patience. 

Both the 17th and Pentecost Monday are public holidays in Norway, which has unfortunately lead to some delays, as well as a rather high workload at the moment. As always, please contact your local RSM if you require more urgent assistance than what is provided.

I've forwarded your question to the relevant R&D team, and will let you know once I hear from them.

Regards,

Elfving

[Sho_Ri]

Hi, Elfving

Thank you for your contact and look forward to replying.

Regards,

Sho

[Elfving]

Hello again,

For the first question, if CTE_COUNT is set to 5, the AUX_SYNC_IND and the first four AUX_CHAIN_IND packets in each periodic advertising event will contain CTE. The frame space between AUX_SYNC_IND and AUX_CHAIN_IN is defined by T_MAFS and it is implementation-specific. By core spec, the minimum T_MAFS is 300 us (Core spec Vol 6, Part B, 4.1.2 Minimum AUX Frame Space), and for the softdevice controller which we use in NCS, this value is 330 us.

For the second question, the two intervals should be independent, and depending on the data length and which PHY it is on, there could be scheduling conflicts. Is that what you are asking about how they affect each other?

Regards,

Elfving

[Sho_Ri]

Hi, Elfving

Thanks for your reply, it's very useful to me.
For the first question, I tested it under the following conditions:
     1. Use two nrf52833-DKs, one as a Beacon tag and the other as a locator and connected to a 4X4 antenna array.
     2. Use the direction_finding_connectionless_tx and direction_finding_connectionless_rx examples, and use the default settings.
     3. In the electromagnetic anechoic chamber, set up Beacon tag in 9 different positions with a distance of 1 to 1.5 meters from the locator. Each position stayed for more than 2 minutes.
According to the default settings of the Beacon example, the intervals of periodic advertising is 1 second, and each periodic advertisement carries 5 CTE packets. Therefore, I assume that in one minute, 300 CTE packets are sent.
When I counted the logs of the locator, I found that as shown in the figure below, about 120 CTE data packets can be received in every locations, and the reception rate was about 40%.
I would like to ask, why is the reception rate in the electromagnetic anechoic chamber only 40%? Are there many CTE packets being ignored during the reception process?

For the second question, I would also like to know how the settings of the extended advertising interval and periodic advertising interval will affect the AoA accuracy calculation？ In addition, do you have any relevant suggestions on how to avoid scheduling conflicts between the two intervals?

Regards,

Elfving

[Elfving]

In general I think it helps to transmit on 2M PHY instead of 1M, and choose an extended adv interval larger than the periodic one so that the periodic events won't be kicked out. However, in this case scheduling conflict typically won't be an issue. As long as the interval duration has sufficient airtime for all the chained PDUs and the two interval are multiples of each other then there should not any problem.

Sho_Ri said:

According to the default settings of the Beacon example, the intervals of periodic advertising is 1 second, and each periodic advertisement carries 5 CTE packets. Therefore, I assume that in one minute, 300 CTE packets are sent.

The periodic adv interval is set between [BT_GAP_ADV_SLOW_INT_MIN, BT_GAP_ADV_SLOW_INT_MAX], and if you look up those values in gap.h, it says they are 1s and 1.2s respectively. However, those values are for extended advertising with a bit interval of 0.625ms. For periodic advertising, the bit interval is 1.25ms. That is why the actual periodic adv interval becomes 2.4s and why you are seeing 125 packets with CTE per minute (60s / 2.4s * 5 = 125). To fix this, use the values starting with BT_GAP_PER_ADV_ as they are calculated with the 1.25ms bit interval.

In addition, is there a fixed time interval between [AUX SYNC IND] and [AUX CHAIN IND], or is there no time interval?

yes, CONFIG_BT_CTLR_ADV_SYNC_PDU_BACK2BACK_AFS=300 the chain PDUs are placed with 300 us aux_offset

Regards,

Elfving