nRF52810 does not work with Radio Test, does not modulate with Direct Test

[dbur]

We are finalizing a product that has a BL652 (nRF52832) in the main unit and a BL651 (nRF52810) in a remote control unit.  We are trying to get our regulatory radio testing done where we need to set up 100% duty cyle continuous modulated RF transmission.

Using Radio Test we are successful doing that with the nRF52832, but the nRF52810 is not responding on it's serial port.  We have re-compiled the code with the legacy serial switch on, but no change.

As an alternative we tried with Direct Test.  Direct test connects to both devices over the serial port and CW transmission works with both.   But for some reason when we try to turn on modulation (PRBS9 with packet length 20) then no RF transmission occurs in either device.

We are at a dead end with Laird tech support on the Direct Test.  It works on the eval board (DVK_BL652) for them, but not for us on the same eval board.  No solution presents itself.  Same problem on our target device PCBs as with the eval board.

So now we are focused on making radio test work again.  I think we need some tech support on how to get Radio test to talk with the nRF52810 serial port.  We know the port works since it works with Direct Test (direct_test_mode_pca10040.hex).  It is just unresponsive with radio Test.  (radio_test_pca10040.hex)

Can we get some support getting this serial port to function?  Thank.

[Håkon Alseth]

Hi,

 

The nRF52810 (and nRF52811) does not have a floating point, so it must be compiled with soft ABI.

For DTM, you should be able to use the emulated "pca10040e" (precompiled: direct_test_mode_pca10040e.hex) project for the nRF52810.

 

The radio_test example does not have a emulated project, so you will have to follow this algorithm to make it into a nRF52810 project:

https://infocenter.nordicsemi.com/topic/sdk_nrf5_v17.0.2/nrf52810_user_guide.html?cp=7_1_5_0_2#ug_52810_software

 

Note that in most IDEs, the second step "In the project settings, change to device "NordicSemiconductor->nRF52810_xxaa"" will select soft floating point for you, but in gcc, you need to change the Makefile, by changing out "-mfloat-abi=hard -mfpu=fpv4-sp-d16" to "-mfloat-abi=soft"

 

Kind regards,

Håkon

[dbur]

Thank you this helped.  We had no luck getting Direct Test to enable the correct modulation required for the testing.  Radio test worked fine,  though we had the problems with the serial port for the BL651.  We ended up needing to buy the full compiler version since the BL651 code base exceeded the size limit for the free version.

[Håkon Alseth]

Hi,

 

There is a bug in dtm for larger packets (>63 bytes), more specifically in ble_dtm.c::on_test_setup_cmd(), line 1305.

It is currently:

m_packet_length = (parameter && LE_UPPER_BITS_MASK) << LE_UPPER_BITS_POS;

It should be:

m_packet_length = (parameter & LE_UPPER_BITS_MASK) << LE_UPPER_BITS_POS;

 

However, with your setting (PBRS9, 20 bytes), this should work even without the above mentioned fix. I tested both on a nRF52810 just now, without issues.

 

Kind regards,

Håkon

[dbur]

Tech support at Laird had the same result.  But we are supposedly using the same FW and same eval board as them but it does not work.  The issue remains unresolved, though we are good to go with Radio Test instead now.

[Håkon Alseth]

I'm glad to hear that you're happy with using the radio_test example at this point, but also sorry to hear that DTM still does not work. If you try to get DTM running in the future, I'd recommend that you try different GPIOs for RXD and TXD, just to see if there's something there.

 

Kind regards,

Håkon