Why nRF52832 and nRF52833 as center connect peripherals often fails?

[Jing Luo]

Hi, Nordic

This is the test result that I use nRF52833 as the host to connect the slave,  slave txpower = 0dbm.

And use Wireshart Sniffer to grab the air packet when it fails.

It looks like update MTU failed. 

We cannot accept the failure rate of connections within 5 meters. Is there any way to reduce the failure rate？

Thank you at first.

[Hung Bui]

Hi Luojing, 

What you showed in the sniffer trace suggest that the CONNECT_REQ was not received by the peripheral, hence the connection was not established. 

There could be many factors that can cause such result: 
- Not optimized hardware. Please make sure the hardware board is tuned. You can use the nRF52 DK as the reference for testing. 

- Collision in the air traffic. Please check if there is any other 2.4GHz activity in the area, for example wifi, other advertisers. 

- Is there any other activity on the peripheral ? If the peripheral would need to do advertising, it won't be able to scan at the same time. 

- It's impossible to have 0 error with RF on an open environment. Theoretically, you can only achieve success rate of 100% when you test in a RF chamber or using a wired antenna. 

You can consider increase the TX Power of the central to +4dBm or +8dBm if needed, this would increase the signal strength and reduce the error rate. 

Note that in indoor environment, it's quite normal to see packet drops at 10m.