Question regarding nRF54L15 PHY and throughput

[Ziyao Zhou]

Hi, I have test the throughput using two nRF54L15 DKs and make them close to each other.

My result is:

1. My first question would be: the best througput for 2M phy is around 1300kbps, for 1M is around 800 kbps. If I have a TX divice move away from the RX, I first use 2M, at close distance I could get around 1300 kbps, after some distance the throghput goes down to 800 kbps. At this time, if I change to 1M phy, could I concluded and be confident that the throughput will perform well for further distance compared to fixed 2M. 

2. The second is that why coded s8 has higher througput compared to coded s2. I think I have one ticket about this but there is no feedback.

Thank you so much for your help!!!

[tesc]

Hi,

1. My first question would be: the best througput for 2M phy is around 1300kbps, for 1M is around 800 kbps. If I have a TX divice move away from the RX, I first use 2M, at close distance I could get around 1300 kbps, after some distance the throghput goes down to 800 kbps. At this time, if I change to 1M phy, could I concluded and be confident that the throughput will perform well for further distance compared to fixed 2M. 

While generally the link budget improves for lower baudrate, there is also an increased probability of packet loss when transmission time increases, due to packet collisions and other RF noise. Transmission time increases with longer packet length or lower baudrate. This means, at a certain range, which performs best of 2M and 1M, or for that matter Coded PHY, depend not only on the range but also on the RF environment. What you should expect, is for 1M to have less issues due to distance, but more issues due to collision and RF noise. When distance is the dominating factor, then 1M will perform better than 2M. The inflection point depends on the environment.

2. The second is that why coded s8 has higher througput compared to coded s2. I think I have one ticket about this but there is no feedback.

Throughput also depend on connection interval, packet length and event length, so differences there might impact the results. Depending on RF noise sources, you may be in an environment where the S2 packets are disproportionally affected, while the S8 packets are less impacted. I also see that all measurements for S8 are the same. I would definitely check if the values are correct, and if so why they all give the same value. For further discussion on this, if you have a different ticket for it, then please use that ticket. I assume that you refer to coded S8 is faster than coded S2.

Regards,
Terje