Coded PHY testing

[slowhand]

Hi,

I get the ble_app_att_mtu_throughput -example working with the coded PHY between two nRF52840 boards. Very nice, although I did not yet get the coded PHY advertising working. Just the coded PHY connection.

To understand how well the actual coded PHY works, I'd like to understand:

1. Does the example code fall back into 1Mbps mode in the middle of data exchange when the link quality drops - or does it go directly from coded PHY connection to DISCONNECT?

2. Are there other issues in the nRF52840 pdk that might cause the coded PHY to work in suboptimal way? Something that might be clearly better in the coming releases? I am not saying I am unhappy, I just want to get realistic idea how well the coded PHY performs.

Thanks!

[FormerMember]

The latest version of S140, S140 v.5.0.0-3.alpha supports advertising and connection establishment on coded PHY. However, the current version of the SDK does only support S140 v. 5.0.0-2.alpha. 2.alpha does not support advertising and connection establishment on coded PHY.

S140 v.5.0.0-3.alpha comes with an example that shows how to connect on coded PHY, the example is inside the softdevice download .zip file. I would recommend you to take a look at that example.

1) When connected on coded PHY, there will not be a change of phy unless it has been explicitly requested. (There is no automatic switching.)

2) Could you please elaborate?

[slowhand]

Thanks!

I actually already sent another question on the example code inside the softdevice .zip file (devzone.nordicsemi.com/.../).

1. Ok. I thought there might be reconnecting (via 1 Mbps mode) implemented, but if not, it is clear to me and I am happy.

2. Perhaps I ask this way: can you evaluate the performance of the nRF52840 PDK Coded PHY implementation? Is it going to be better in coming (non-preview) releases? This may be a difficult question and you are free to say "no comment" :-). Anyway, I am trying to evaluate the performance of the BLE coded PHY in general and I am currently totally dependent on Nordic's preview kit.

[FormerMember]

I have answered your other question/thread.

1) I have not tested the example that comes with 3.alpha, but it should be possible to double check the parameters that are being used. In your other question, I have uploaded another example as well that only uses coded phy.

**2)**We are aware that there are some performance issues with coded PHY. We are working on making a fix!

[slowhand]

Ok, thank you!