nRF21540 antenna diversity for BLE

[akif]

Hi,

We are working with nRF21540DK and developing a BLE application. nRF21540DK with nRF Connect SDK 2.5.0 . I have couple of questions.

1. In nRF21540DB product page says "The two antenna ports on the nRF21540 DK are useful for antenna diversity scenarios in 802.15.4 based protocols (i.e. Thread or Zigbee) to reduce multipath propagation effects.". I couldn't see any explanation on the datasheet regarding two antenna and BLE.
   
   Does BLE protocol use the both antennas?
2. If BLE doesn't use both antennas, which antenna it use, POUTA or POUTB?
3. POUTA set to 20dB and POUTB set to 10dB as default. Is there any reason you have selected these values as default? For example why it is not 21db which is the maximum value.
4. Why you haven't selected 20dB for both antenna? Why the POUTB is exactly the half value of the POUTA ?
5. In datasheet page 25, CONFREG0 TX_GAIN TX gain control (0: minimum, 31: maximum). What is the unit of this value, obviously it is not dB, because maximum is 21dB. 
6. What is the minimum and maximum dB value we can set for POUTA and POUTAB individually?
   

Thanks.

[JONATHAN LL]

Hi,

akif said:

Could you be more specific? How we can use antenna diversity scenario on BLE? Does it a part of configuration of MPSL on SDK 2.5.0? I think the quote from the product page I mentioned above only mentions 802.15.4 based protocols not BLE for a reason. We possibly make a our custom board with nRF52840 and nRF21540 to use with BLE Mesh. I want to understand if we can have any benefit from the 2nd antenna.

There could be scenarios where using two antennas would be beneficial, but it often just complicates design and introduces extra components. So for BLE, you could have a scenario where you have a directional antenna and a omnidirectional antenna that you can switch between depending on what you want to achieve. So there is not neacsaeraly a dedicated use case for antenna diversity but it does not mean it cant be. It could be used for localization if an antenna array is used. So there are antenna diversity options for BLE as well. But this would have to handled in the application as and not in the softdevice/BLE stack.  So there are better diversity options for other protocols, but it does not mean that it is impossible for some antenna diversity to be used with BLE. 

If there is not a specific use case or situation in mind for your application then it is likely best to go for one antenna.  Adding more antennas increases the cost and complexity in the physical design but also on the development side. 

akif said:

Okay either we use POUTA or POUTB. I understood. I thought we can select antenna gain of individual antennas. So when we set POUTA=20dB, does it set both ANT1 and ANT2 TX gains to 20dB? 

If you use POUTA  then the output power will be 20dBm with both antenna ports so ANT1 and ANT2. Only one antenna is used at the time so only one ANTx port is set at the time. 

akif said:

Sorry I don't have expertise on radio signals, I couldn't see if antenna have any effects on power model compensation. If we use a custom antenna on our custom board, do we need to create our custom power model or can we use built-in power model directly?

The antenna will effect the gain, since all antennas has gain, but exactly what type of gain in how much is not something the FEM can adjust for. The way to do this is to tune the matching circuits and antenna with a spectrums analyzer and a VNA to find the optimal result. Then the built-in model will handle the other variables to adjust the gain to correct levels. 


Regards,
Jonathan

[Ders]

JONATHAN LL said:

If there is not a specific use case or situation in mind for your application then it is likely best to go for one antenna.  Adding more antennas increases the cost and complexity in the physical design but also on the development side. 

JONATHAN LL I do have a specific use case for the same question.  
I have a BLE system where we have battery operated BLE devices and we had to back off on TX power to meet battery life.  The battery operated device talks to a wall powered gateway set to max TX power.  This creates a asymmetric communication where the gateway has difficulty hearing the battery operated devices.    To improve the gateways capability to hear the battery operated device I was planning to use your nRF21540 to utilize the LNA as well as gaining receive signal strength from antenna diversity.  

My configuration 

The battery operated nRF52840 is capable of +8dBm but backed off to +0dBm 

Knowing this use case what do you think of our approach and trying to rely on antenna diversity to pick the polarization with the strongest signal. 

We do not have control over the installation so I can't relay of directional antennas and the fact that the gateway will communicate with multiple battery operated devices.  

Thanks 

[JONATHAN LL]

Ders said:

Knowing this use case what do you think of our approach and trying to rely on antenna diversity to pick the polarization with the strongest signal. 

There is not a relay good solution for choosing what antenna would be best i think. You can use RSSI and look at the "strength" then switch antenna and compare, keeping the antenna that gets the best results. But I am not sure that there will be a huge difference in signal quality here so this might be a lot of work for little improvement. 

Is the battery powered unit relatively mobile and will change possession\location often then selecting antenna based on RSSI value over time might not net great results. 

The LNA will work, but antenna diversity might not be as effective as you want it to be, but this is difficult to say and you would have to test this and evaluate if it is enough of a bonus that it is worth it. 

Regards,
Jonathan

[Ders]

It's kind of a IoT system where the HUB/gateway is providing internet to the the battery powered BLE devices.  So the hub sees 5-20 devices.

• All installations are different and I'm trying to maximize the number of devices the HUB can connect to.
• The hub/gateway is fixed attached to a wall outlet
• The battery devices are also fixed and installed on a "walls". 
• Due to multiple devices I can't use directional antennas and also size will not allow for such antenna.
• The wireless enviroment at 2.4GHz is typically quite noisy due to residential scenario

I will attempt to ensure the antennas on the HUB and devices are in the polarized plane but due to nature of multipath/fading what have you, one never know what the real scenario is. Seeing the Nordic nRF52 + FEM devkit with 2 antennas made me think I can take advantage of the to improve the case when a wireless link is compromised due to polarization mismatch. 

Here is a quote from an article on the subject

"Polarization mismatch between antennas is characterized by polarization loss factor (PLF). The parameter is expressed in decibels (dB) and is a function of the angular difference in polarization between the transmit and receive antennas. In theory, PLF can range from 0 dB (no loss) for perfectly aligned antennas to infinite dB (infinite loss) for perfectly orthogonal antennas."

If you don't have any real nature test data on BLE and antenna diversity than that's fine and I'll just use the LNA and not the second antenna port.  In theory it seams like "free power" in a scenario where loss is caused by polarization mismatch.

[Ders]

Got some information from Nordic through email which explains the complication with BLE and antenna diversity.   

We don't have a similar example for other protocols. The reason is really because the 802.15.4 have a very long pre-amble. This allows you to sample the pre-amble with both antennas and decide during the pre-amble which has the best signal strength and used this. This can be done for every packet received. Bluetooth has a very short pre-amble so you can't change antenna during the packet reception so you can only switch antenna in between packets -> moving the antenna diversity out of the controller and onto the application space. You would then have to run something to a QoS service on the antennas and then switch antennas based on the QoS of each. This is not something we have.

Another complicating factor for Bluetooth (and other short pre-amble protocols): if you receive from multiple transmitters you will need to run a separate QoS algorithm for each transmitter. This is not the case if you have long pre-amble as you can do it per packet.

Also have in mind, for Bluetooth you are also frequency hopping, so the deep "nulls" will be different on each frequency. Antenna diversity is primarily useful if you have two stationary devices communicating on a single frequency; if you are moving or frequency-hopping, you might as well retransmit as change antennas.