How much does each timeslot period tier slow down BLE?

[Abraham Hamidi]

Using the period required for 2Mbps Gazell, 600us, how much does this slow down BLE (bandwidth/# packets)? Also, does the timeslots API work well with nrf52? If I modify this example for nrf52, will it work? github.com/.../wireless_debug.

[Håkon Alseth]

Hi,

 

The timeslot you request, and possibly extend, will be handled as per the specification:

https://infocenter.nordicsemi.com/topic/sds_s132/SDS/s1xx/multiprotocol_operation/multiprotocol_support.html?cp=4_7_3_0_8

I would strongly recommend that you read this to understand how the timeslot API works and behaves.

 

Using the period required for 2Mbps Gazell, 600us, how much does this slow down BLE (bandwidth/# packets)?

The problem is the other way around. Depending on when the BLE link needs to be serviced, your timeslot period will yield.

Please see this chapter: https://infocenter.nordicsemi.com/topic/sds_s132/SDS/s1xx/concurrent_multiprotocol_tsl_api/performance_considerations.html

 

Kind regards,

Håkon

[Abraham Hamidi]

Thank you. One more question, why does the timeslot period decrease with higher rate for gazell? From what I could find in the docs, it said >=60 us for 2Mbps, then I don't remember but something like >=90 us for 1Mbps. Why is higher timeslot period lower gazell rates? Shouldn't it be other way around?

[Håkon Alseth]

Hi,

 

Abraham Hamidi said:

One more question, why does the timeslot period decrease with higher rate for gazell?

Each "gazell period" can do one transaction. One transaction to send, and one to receive an ACK.

This how the gazell protocol is implemented for nRF5-series devices:

https://infocenter.nordicsemi.com/topic/sdk_nrf5_v17.1.0/gzll_02_user_guide.html?cp=8_1_5_4#timeslot 

 

Please note that gazell timeslots are different from the SoftDevice timeslot API, although they share the same name, which can be a bit confusing.

 

Kind regards,

Håkon