nRF9160 Bluetooth coexistence Insufficient Documentation

[optical]

Hello,

We are in process of FCC certification and we would like to use LTE and Bluetooth Coexistence to reduce number of measurement needed for the certification.

We are aware of SoftDevice support for the coexistence, we would like to use zephyr Opensource Bluetooth controller. We are willing to help with the Opensource implementation on the controller side.  However to do so we encountered an issue. There is a large gap in documentation of the COEX pin functionality. So there are several questions i would like you to clarify.

1. COEX0 functionality - Input to the LTE modem from external device.
   1. Does this effect only Transmission or also Receiving ?
   2. Is it currently implemented inside the Modem firmware ?
   3. Diagram shows on COEX0 State x State y State z, What exactly does that mean ? It is single GPIO which can only be 'one', 'zero' or 'high impedance' ?
2. COEX2 functionality - Shows the pin in '1' during whole operation of the LTE Modem.
   1. How does it cycle when the modem is NOT using PSM ? Doest it even cycle to allow Bluetooth to operate ?
   2. Is the COEX2 active during LTE receiving ? I found later that indeed the COEX can be setup using AT Command. This resolve this question
      
   3. Can the timing of the COEX2 pin some how combine with '4 – COEX2 pin toggling enabled, TX indicator. Pin ON when RF is transmitting.' ? What is the default timing for this option "as late as possible timing, approximately 50–100 μs anticipation."? 

Maybe you could give us an insight into operation of the coexistence in SoftDevice controller to help us.

[optical]

I have tested the COEX2=4 mode using logic analyzer. I found out that the the modem TX mode in poor signal conditions is taking a large amount of bandwidth shared of the device. 

In this state the Bluetooth will became unusable due to limiting the windows for its transmissions. (The image contains both cases for Poor and Good signal, enlarge for view)


The Modem firmware needs to have some option to give the Bluetooth time to transmit required data, or enable the Input functionality of the COEX0

How should we proceed with this ? I believe that the same issue must have been observe, when the Coexistence feature for the SoftDevices controller were developed. How did you solve this issue?

[helsing]

Hi Tomas, sorry for the late reply.

Have you had a look at the description of the 1-Wire coexistence protocol, as well as the  Bluetooth External Radio Coexistence examples? 

I hope to have a few more comments to you over the weekend.

[optical]

That is implementation for SoftDevice, which is not what i seek. Read the initial ticker description please.

The second this does not answer my question of behavior of nRF9160 COEX pin with poor signal.
How can bluetooth be operation when LTE modem takes more 90% of transmission time. There needs to be some settings inside the modem to limit this. Or there should be support for COEX0 inside the modem.

Could you pass this ticket to HW engineers from modem firmware team ? 

[helsing]

optical said:

I have tested the COEX2=4 mode using logic analyzer. I found out that the the modem TX mode in poor signal conditions is taking a large amount of bandwidth shared of the device. 

What is the modem doing in this plot? Is it an RRC connection, or just listening for pages?

As you point out, under poor conditions, you probably have multiple repetitions, which will increase the radio usage. Repetitions etc. are configured by the network, so it is limited what you can do. One suggestion is to have a look at %CONEVAL. This command could give the application the information it needs to decide if sending data would cause coexistence issues.

[optical]

The modem at this point was trying to transmit packets to the server. Is there an option how to configure the maximum number of repetition or some delay between them ? 

The Bluetooth operation would definitively be disrupted by this behavior.