nRF52833 + nRF21540 Custom Board

RE: nRF52833 + nRF21540 Custom Board

mathiaso — Tue, 02 Sep 2025 11:01:45 GMT

Hi, Ashwini.

Asbjørn is currently out of office, so I'll try to help you out while he's away.

Have you had a look at this section in the documentation, describing how to enable GPIO+SPI mode support for nRF21540?

Are you sure you're not able to share a PDF containing your board's schematic and a zip folder containing the Gerber and drill files for the layout? Even not if the case is made private? This could possibly be very helpful for figuring out what may be the issue.

Best regards,
Mathias

RE: nRF52833 + nRF21540 Custom Board

Embel_Tech — Thu, 28 Aug 2025 08:00:23 GMT

Hello Asbjorn,

Due to certain limitations, the schematic cannot be shared.

I would like to request confirmation on the following points regarding the control and configuration of the nRF21540 FEM for our testing setup:

TX Gain Enable Sequence (Using GPIO and SPI):
After enabling the nRF21540, the initial state of the control pins — MODE, TX_EN, and RX_EN — is LOW (logic 0).
My current procedure is as follows:
- First, I write to CONFREG0 to set the TX gain (e.g., writing 0x1C for +10 dBm gain or 0x6C for +20 dBm gain).
- Once the write is successful, I set the TX_EN GPIO to HIGH (logic 1) to enable TX gain.
Could you please confirm if this is the correct sequence for enabling TX gain using a combination of SPI and GPIO control and values written to CONFREG0 for the corresponding gain?
TX Gain Control Using Only SPI (No GPIO Control):
In the case where we want to set and enable TX gain using only CONFREG0 (Only SPI control without using FEM control GPIOs),
- What is the correct sequence of SPI commands to set and activate the TX gain?
- What should be the expected or required states of the TX_EN, RX_EN, and MODE pins in this SPI-only control scenario?

Your guidance on these points will be greatly appreciated to ensure correct operation during our testing phase.

RE: nRF52833 + nRF21540 Custom Board

Asbjørn — Tue, 26 Aug 2025 13:35:40 GMT

Hello Ashwini,

[quote user="Embel_Tech"]as per schematic no external crystal is mounted on it. [/quote]

What schematic? I've asked about the schematic and layout for this board and the implementation, but haven't seen any. Would it be possible to share as this might also be related to the issues you're observing.

Best regards

Asbjørn

RE: nRF52833 + nRF21540 Custom Board

Embel_Tech — Thu, 21 Aug 2025 12:08:00 GMT

Yes, we are using LFRC. as per schematic no external crystal is mounted on it. Also documentation 500 ppm value reflects the expected accuracy after calibration for the internal RC oscillator. we cannot set it externally.

Also, tried with setting up configuration as:

CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=y
CONFIG_CLOCK_CONTROL_NRF_K32SRC_500PPM=y

but not getting drastic change in observations.

Is there need to set accuracy at different value for internal RC?

RE: nRF52833 + nRF21540 Custom Board

Asbjørn — Wed, 20 Aug 2025 14:19:23 GMT

Hello Ashwini,

based on your numbers alone it looks like the gain from the FEM is there and the numbers don't look too bad. Suspect that there might be two things worth checking here:

Q1: What does the layout and schematic look like on your board? Are you able to share this information with us if you set this ticket to be private?

Q2: RX timing, what clock source do you have available on your board. LFXO (32k crystal) or are you using LFRC (internal 32k)? It looks like you are using LFRC based on this: CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=y, but what are you using for the clock source accuracy? https://docs.nordicsemi.com/bundle/ncs-1.7.0/page/kconfig/choice_483.html#choice-483 Can you set it to https://docs.nordicsemi.com/bundle/ncs-1.7.0/page/kconfig/CONFIG_CLOCK_CONTROL_NRF_K32SRC_500PPM.html#std-kconfig-CONFIG_CLOCK_CONTROL_NRF_K32SRC_500PPM and give it a try? I fear that if you haven't set this correct, your RX windows are out of sync with the phones.

Best regards

Asbjørn

RE: nRF52833 + nRF21540 Custom Board

Embel_Tech — Tue, 19 Aug 2025 07:29:39 GMT

Additionally, we have conducted RSSI testing for various TX gain levels and distances.

Could you please review the results and let us know which values appear to be expected and if any seem unusual or inconsistent?

devzone.nordicsemi.com/.../nRF21540_5F00_TX_5F00_gain_5F00_Testing_5F00_18_5F00_Aug_5F00_25.xlsxYour feedback will help us validate our observations and proceed accordingly.

RE: nRF52833 + nRF21540 Custom Board

Embel_Tech — Tue, 19 Aug 2025 06:59:52 GMT

Hello Asbjorn,

Yes, we have the FEM integrated with a custom nRF52833 board, and there is no provision to separate them for testing. As a result, we are unable to trace the GPIO state.
Both antennas have been connected for testing purposes.

RE: nRF52833 + nRF21540 Custom Board

Asbjørn — Mon, 18 Aug 2025 12:37:07 GMT

Hi Ashwini,

This sounds like far from what I would expect. I asked about the custom board performance with the FEM connected, is that a possibility or is the FEM integrated into your custom board? If it is possible to separate, what is the performance of just the custom board?

Timing between states, from the product specification:

Could you capture the state of the control GPIOs across a couple of advertising and connection attempts? PDN, TX_EN, RX_EN, ANT_SEL?

Do you have the antenna connected to the selected antenna pin on the nRF21540?

Best regards

Asbjørn

RE: nRF52833 + nRF21540 Custom Board

Embel_Tech — Mon, 18 Aug 2025 06:02:03 GMT

Yes, at 40 meters I can only scan my device, and I'm only able to establish a connection at a range of 5–6 meters.

RE: nRF52833 + nRF21540 Custom Board

Asbjørn — Wed, 13 Aug 2025 11:14:25 GMT

Hello Ashwini,

Just to be sure how your custom board performance with the FEM. If you bypass the FEM and only use the nRF52833 on your custom board, what's the range you see then?

You say BLE scan range is 40m, are you able to establish a connection at this range or is that just the range you can see the advertising at?

Best regards

Asbjørn