ESL / nRF5340DK / nRF21540EK

[Scott Grubb]

I am experimenting with the Nordic ESL sample application.  The central_esl app is running on a nRF5340DK.  I would like to add the nRF21540EK to this to improve the range.  I have been following the suggestions here: Using nrf5340dk with the nrf21540 EK - Nordic Q&A - Nordic DevZone - Nordic DevZone.  Specifically, I have added the following to the build configuration:

However, the ESL sample application uses an older version of the SDK (v2.5.0) than in the linked post and, although the application builds successfully and I can run it, the nRF21540EK does not seem to be getting enabled.

I can see that 'nrf21540ek_fwd.overlay' is being added to the parent image but there is no overlay added to the hci_rpmsg child image, and there are no references to the default MPSL FEM settings in the output .conf file for the hci_rpmsg child image as there were in the linked post.

I've tried using '-Dhci_rpmsg_SHIELD=nrf21540ek' instead because I think SDK v2.5.0 pre-dates the name change from hci_rpmsg to hci_ipc, but that didn't work either.

How do I go about enabling the nrf21540?

Thanks

Scott

[Scott Grubb]

A further update. I can now see activity on the GPIO lines, so I think the FEM is 'working'.  However, the RSSI values that I am seeing at either end of the connection are still much lower than I see when using the nRF5340DK's built-in antenna.

This is the antenna I'm using:
Molex 212860-0011 Whip WiFi Antenna with SMA RP Connector, Bluetooth (BLE), WiFi | RS

(I have one attached to both of the ANT connections on the FEM.)

Scott

[Kenneth]

Hello,

What have you set CONFIG_BT_CTLR_TX_PWR_ANTENNA to? Try 20.

If that doesn't help, maybe you can share the following files from your build folder:
\build\<name>\zephyr\.config
\build\<name>\zephyr\zephyr.dts

Maybe also upload a picture of your hardware setup.

Kenneth

[Scott Grubb]

I've tried plugging in both USB cables and also adding the following to the prj.conf file used by the hci_rpmsg child app.

CONFIG_MPSL_FEM=y

CONFIG_MPSL_FEM_NRF21540_TX_GAIN_DB=20

CONFIG_MPSL_FEM_NRF21540_RX_GAIN_DB=13

CONFIG_BT_CTLR_TX_PWR_ANTENNA=20

These are the sort of RSSIs I'm seeing:

The Rx value is the RSSI of the ESL tag as seen by the Access Point.  The Tx value is the RSSI of the Access Point as seen by the ESL tag (and communicated back to the Access Point).  The Tx value seems to have improved a bit with the latest changes, but both values are well below what I see when using the nRF5340's onboard antenna:

[Scott Grubb]

As a test, I tried running the peripheral_uart example on the nRF5340 with and without the FEM enabled (using the extra CMake arguments above).  I see the same issue: really low RSSI values with the FEM enabled; this time measured using the nRF Connect app on my phone.  With the FEM enabled, I'm seeing activity on the TX_EN and RX_EN GPIO lines, so it looks like the nRF5340 is at least trying to do something with the FEM.

[Scott Grubb]

Hi,

I'm still struggling to get acceptable RSSI levels when using the nRF21450EK.  I've had a closer look at the GPIO signals on a 'scope and they align with what I would expect to see (e.g. the TX_EN coincides with the ESL subevent advertising interval) so I'm beginning to suspect that the problem is on the RF side. 

Can I just confirm that the nRF21540EK is intended to be used with antennas and not just to be connected to lab equipment? (The Product Brief says it is, but the hardware description isn't clear: "The nRF21540 EK is a hardware platform used to monitor the performance of the nRF21540 RF front-end module (FEM) by connecting it to lab equipment or other radios with a SubMiniature Version A (SMA) connector."

[Scott Grubb]

I've just come across this:  RE: nRF21540 FEM TX Power with ESB 

I've tried forcing the PDN pin high as per the link but it had no effect in my setup.  Does the bug exist in SDK v2.5.0?

[Scott Grubb]

In an attempt to rule out the hardware as the problem, I have now switched to the following setup:

nRF52833DK + nRF21540EK

SDK: v3.0.0

I'm running the peripheral_uart sample app.

-DSHIELD=nrf21540ek

CONFIG_MPSL=y

CONFIG_MPSL_FEM=y

CONFIG_MPSL_FEM_NRF21540_GPIO=y

CONFIG_MPSL_FEM_NRF21540_TX_GAIN_DB=20

CONFIG_MPSL_FEM_NRF21540_RX_GAIN_DB=13

CONFIG_MPSL_FEM_NRF21540_GPIO_SPI=y

CONFIG_MPSL_FEM_NRF21540_TX_GAIN_DB_POUTA=20

CONFIG_MPSL_FEM_NRF21540_TX_GAIN_DB_POUTB=10

I'm still seeing significantly lower RSSIs when using the shield than without.  I've tried two different shields.

All that leaves is the antenna.  I'm using this one: https://uk.rs-online.com/web/p/wifi-antennas/2035105.

Is there an antenna that Nordic recommend?

Thanks

Scott

[Scott Grubb]

Using the setup described above:

nRF52833DK only build:

nRF52833DK+nRF21540EK build:

If I run the nRF52833DK+nRF21540EK build but don't plug in the shield (i.e. use the on-board antenna), the results are no different to when the shield is plugged in.  That might suggest that it's a configuration problem.

[Scott Grubb]

I have now tried running the SDK v3.0.0 peripheral_uart sample app across the following hardware combinations:

- nRF5340DK, nRF52840DK, nRF52833DK, nRF21540DK,
- nRF21540EK (x2)
- Murata cable (x2)

In ALL cases, I see the same behaviour: i.e. significantly lower RSSIs when using the shield than without.

When building for the shield, I'm using the following setup:

-DSHIELD=nrf21540ek

CONFIG_MPSL=y

CONFIG_MPSL_FEM=y

CONFIG_MPSL_FEM_NRF21540_GPIO=y

CONFIG_MPSL_FEM_NRF21540_TX_GAIN_DB=20

CONFIG_MPSL_FEM_NRF21540_RX_GAIN_DB=13

CONFIG_MPSL_FEM_NRF21540_GPIO_SPI=y

CONFIG_MPSL_FEM_NRF21540_TX_GAIN_DB_POUTA=20

CONFIG_MPSL_FEM_NRF21540_TX_GAIN_DB_POUTB=10

The shield has power, and I can see the TX_EN and RX_EN lines toggling.  I have Molex antennas connected to both antenna ports on the shield.

If I run the app on the nRF21540DK (built using the CONFIG_ items but not the -DSHIELD option) then the performance is no different to using one of the other DKs without the FEM.

Can someone from Nordic please get back to me?  I'm sure the solution has to be something simple like a missing configuration item, but it really shouldn't be this hard to get an evaluation kit working out of the box.

Thanks

Scott

[Kenneth]

Hi,

I am really sorry but I have not gotten around to test this myself.

I am wondering if you can send me a picture of each end of the cable you have between the DK and FEM?

Also, if you are able to connect ANT1 or ANT2 to a spectrum analyzer directly for comparison that would be helpful. That is what I plan to do when I get around to it.

Kenneth

[Kenneth]

Looks like the latest update from you is missing the CONFIG_BT_CTLR_TX_PWR_ANTENNA=20.

Kenneth

[Scott Grubb]

Adding CONFIG_BT_CTLR_TX_PWR_ANTENNA=20 seems to have improved things but the performance with the shield is still worse than without, just not by as much.

The cable is this one: MXHT83QE3000 Murata Electronics | Connectors, Interconnects | DigiKey

 

I don't have access to a spectrum analyser today but will do tomorrow. 

Thanks

Scott