BLE TX weak signal in nRF Connect SDK

[ekidsalan123]

Hello, we're currently developing BLE devices with PCA10040 and nRF Conenct SDK 2.9.0.

During a series of experiments, we found that with the same hardware and the same BLE TX power setting (0dB), the firmware transmission performance using nRF Connect SDK is not as good as the old nRF5 SDK. The following are the experimental conditions:

Place a metal chair (made of iron) in an empty space, place the PCA10040 on it, and place a mobile phone about 50 cm away to check the RSSI.

At this time, both NCS2.9.0 and nRF5 SDK firmware can get a good signal (about -60dBm)

Now, we sit lightly on top of the PCA10040, wrapping our body around the entire development board (the metal plate is below and the human body is above).

The firmware developed using NCS2.9.0 will have no signal at all (no Bluetooth broadcasts can be received), but the firmware using nRF5 SDK can still get a weak but stable signal (about -80dBm ~ -90dBm).

We have tried adjusting CONFIG_BT_CTLR_TX_PWR_PLUS_4=y in NCS2.9.0, but it doesn't seem to help much. Can this be fixed?

[Turbo J]

ekidsalan123 said:

Place a metal chair (made of iron) in an empty space, place the PCA10040 on it,

Bad idea. The metal would either introduce a ton of directionality to the signal or could even be just a dead short. Bluetooth is not expected to work properly with large metal objects present near the antenna.

Using the human body as a testing tool is another bad idea, you also get an mountain worth of variance here.

[ekidsalan123]

Hi Turbo,

Absolutely I knew that's not a good idea for using BLE communication.

But this may happen in cases where our customers cannot control the usage.

What we are actually curious about is why the new SDK is worse than the old one? Under the same conditions, this should not happen, which will make us give up using the new SDK and new products.

This has been verified on more than one PCA10040 development board.

We only used human bodies for testing, but I think that maybe using other masks to reduce the signal would give the same results.

In summary, please focus on why the signal of the new SDK firmware is worse after being masked or interfered?