nRF52832-CIAA-G-R7 Rx Sensitivity

[Star.Tong]

Hi There,

During the transmission and reception performance test, I found that the LE1M's maximum sensitivity was only -92 dBm, while the LE2M's maximum sensitivity was only -89 dBm. These values were measured using the IQ and added pathloss. The data do not match sensitivity of -96dBm in datasheet. Could you please confirm whether the configuration mode for SDK 2.6.1 is nRF mode?

[Asbjørn]

Hello Star Tong,

Regarding the sensitivity, it comes down to layout and implementation as well. So what sort of HW did you run the test on? Are you able to compare performance with a nRF52832 DK?

The results we have reported in the product specification come from performance on boards using the reference schematic and layout with an ideal transmitter. It's optimum conditions, but this is what we would expect to see customers getting close to as well with a reference schematic and layout.

---------------------------------------

I don't understand what you are asking here:

"Could you please confirm whether the configuration mode for SDK 2.6.1 is nRF mode?"

May I ask you to elaborate  what you mean by this?

Best regards

Asbjørn

[Star.Tong]

Hi Asbjørn,

Thank you for your reply.

I tested on my own custom-designed product board.

Regarding the layout issues you mentioned, I have the following additional comments from a hardware perspective:

1. I verified the Tx performance aligns with the link budget, so sensitivity degradation due to mismatch can be ruled out.

2. If the issue stemmed from desensitization, the 3dB degradation observed at LE2M compared to the datasheet would imply over 6dB degradation at LE1M—not just another 3dB. Currently, the 3dB sensitivity difference between LE1M and LE2M aligns with theoretical calculations.

3. I noticed the following statement in the datasheet:

'Typical sensitivity applies when ADDR0 is used for receiver address correlation. When ADDR[1...7] are used for receiver address correlation, the typical sensitivity for this mode is degraded by 3dB.'

Therefore, I request your assistance in verifying whether any errors exist in the register configuration at the software level.

Best Regard.

Star.Tong

[Asbjørn]

Hello Star Tong,

The -96dBm sensitivity 1M mode is for ADDR0 only. For the remaining pipes (1..7), it is 3dB worse, so -93dBm. Which address field are you using for your tests?

For 2M mode the sensitivity is -89dBm:

Q1: Aligned with TX performance, yes this is right. So as long as you have 0dBm out in TX, you should have good receiver performance as well given that the carrier frequency/crystal accuracy is good.

Q2 and Q3: I suspect you are using addr1-7 during these tests? Can you confirm or share how you configure the receiver for these tests?

Best regards

Asbjørn

[Star.Tong]

Hi Asbjørn,

Thank you for your kindly reply!

What information do you need me to provide? Could you tell me how to export the file?

THX & BR.

Star.Tong

[Asbjørn]

Hello Star.Tong,

I don't think you need to export anything to me as long as you check how you set up and configure the radio for your test. You must be loading some FW into the device to run your test, how does it configure the receiver, specifically if ADDR0 is used. You can check this yourself in your code as I don't know what sort of FW you have built for this purpose. Have you gotten the FW from somewhere else? What are you using as a transmitter during your tests?

-------------------------------------

Best regards

Asbjørn

[Star.Tong]

Hi Asbjørn,

Our software told me that he use the standard SDK and the version is 2.6.1.

Would you mind tell me how can I check the information related to the address?

I use the Litepoint MW-7G as generator during test, and the waveform file is attached.

Thanks!

Best Regard.

Star.Tong

[Asbjørn]

Hello Star.Tong,

you can check by either reading the registers or checking what you write to those registers in your FW while doing the test.

If you are using ADDR0, it would be what is configured as BASE0 and PREFIX0.AP0, these two specific registers have these fields:
BASE0:docs.nordicsemi.com/.../radio.html

PREFIX0.AP0: https://docs.nordicsemi.com/bundle/ps_nrf52832/page/radio.html#d939e14642

You would also have to know which address you are sending packages to, so that depends on what you have configured the LitePoint to transmit. I assume you have set it up to a format of this kind: https://docs.nordicsemi.com/bundle/ps_nrf52832/page/radio.html#d939e132 The "ADDRESS" there should match BASE0 and PREFIX0.AP0 to send to logic channel 0 which has the best sensitivity.

Best regards

Asbjørn

[Star.Tong]

Hi Asbjørn,

 

The software's response is as follows. Please help me check it:

BASE@0x4000151C = 0x89bed600

BASE@0x40001524 = 0x0000008e

BASE@0x40001530 = 0x00000001

Best Regards.

Star

[Asbjørn]

Hello Star,

''

It seems you have address 0x89bed68e configured for ADDR0. The remaining question is what HW are you measuring on and with? Is it a custom board and what do you mean with "1. I verified the Tx performance aligns with the link budget, so sensitivity degradation due to mismatch can be ruled out." What is aligned? What output power did you configure and what specifically did you measure? 

Best regards

Asbjørn

[Star.Tong]

Hi Asbjørn,

The custom circuit board I'm using

When Tx setting is 4, I measure 3.5dBm at the antenna end, corresponding to 4.1dBm at the chip output, with very flat in-band power.

I've measured PWR, Frequency error, ACP, ect... all of them are fine.

Best Regards.

Star

[Asbjørn]

Hello Star.Tong,

From what you are writing, the output power seems to be ok as you say.

------------------------------------------

There are still aspects related to your custom board that can affect this, like noise on the power supply, frequency accuracy and connections made, packet length, BLE mode or nRF mode. I don't know these aspects so it is difficult to point at what is causing the difference between the expected -96dBm and the measured -92dBm in 1M BLE mode. https://docs.nordicsemi.com/bundle/ps_nrf52832/page/radio.html#d939e34725

------------------------------------------

We do run all our chips through production testing and they need to show the same performance. In theory the output power between chips are the same, but majority will range with +/-2dBm from specification and some outlier could go out to +/-3dBm, but it would be unrealistically unlikely that you would have several of these outliers on all of your boards. 

------------------------------------------

I don't have all the information in this case, so at this point it is not possible to know or explain the findings. We can help you onwards, but it would require more information about the board and FW you are using. Have you tried comparing the performance with a nRF52832 DK?

Best regards

Asbjørn