nRF52840 and nRF9160 sensitivity to out of band blockers

[Stephen Snyder]

Hello,

We have a design that includes both the nRF52840 and the nRF9160.  We are doing custom antenna designs in a miniaturized form factor and would like to specify required isolation between antennas.  In order to do this, I need to understand the impact of out of band blockers for each receiver.  Note that this is not the same as RX selectivity or ACI, because I would expect the receiver can tolerate more RF power hundreds of MHz away from it's band than only one or a few channels over. 

I'd like to know the following:

1) For the nRF52840, please provide information on how much RF input power in LTE bands 2 (1850 - 1910MHz), 4 (1710 - 1755MHz), and 12 (699 - 716MHz) cause a specific degradation in RX sensitivity.  For instance, the answer may look like: "A CW blocker of -30dBm will degrade RX sensitivity by 3dB)".  I'd like to see this for a small degradation, something in the range of 1-3dB.

2) Similarly, for the nRF9160, please provide information on how much RF input power in the 2400 - 2500MHz range will cause a specific degradation in RX sensitivity.  Answer should look like the one above.  There may be a different answer for each LTE band.

The image below shows how this is specified for a GNSS receiver.  This is a great way to document impact of out of band blockers (when coupled with sensitivity degradation caused by these blocker levels).

I'd also like to know if you expect modulation on the blocker to be a factor.  Please note we also have 802.11n WiFi radio, so the 2.4GHz will use those modulation schemes as well.

I realize this is a detailed question, but I think the answer here would be highly valuable to many customers, so thanks in advance.

Steve

[ketiljo]

This is not easy to answer. Blocking or selectivity tests are normally done at a specific level where the packet error rate is required to be below a certain threshold. For the nRF52840, the test outlines in the Bluetooth core spec. is used, there the wanted level is at -67 dBm and the blocker is increased until BER = 0.1% https://infocenter.nordicsemi.com/topic/ps_nrf52840/radio.html#unique_595666871

The wide band blocking, at > 12 MHz from the wanted frequency is - 52 dB. Meaning that the blocker can be 52 dB higher than the wanted frequency before PER increases to 0.1%. 

For the LTE modem, this is not stated in the product specification. There are SAW filters in the RX chain on the module that will handle WLAN without any impact on the LTE receiver. The receiver requirements for a LTE modem is defined in ETSI EN 301 908-13, chapter 4.2.7. https://www.etsi.org/deliver/etsi_en/301900_301999/30190813/13.02.00_20/en_30190813v130200a.pdf

For co-location of BLE and LTE, it's the power from the LTE transmitter that you have to worry about, not the other way around. A band pass filter before both the BLE and LTE radio will be needed to keep the 2.4 GHz receivers from being blocked by the LTE transmitter. LIke it's done on the nRF9160-DK https://infocenter.nordicsemi.com/topic/ug_nrf91_dk/UG/nrf91_DK/bluetooth_ieee_network_processor.html

[Stephen Snyder]

Thanks for the response.

For the nRF52840, for the > 12MHz, how far away from the wanted signal is the blocker positioned?  If this test is only run with blockers in the 2400-2500 MHz range, this spec may be overly conservative.  Meaning, a blocker in the LTE bands (699-716MHz, 1710-1755MHz, or 1850 - 1910MHz) may be able to be tolerated at higher power levels.  If we knew that it would reduce the requirements on the SAW filter.  Are you able to find out if you have data with blockers in these bands, or if not, could you do some testing?  This seems applicable to many customers.

The info on the LTE radio embedded SAW filters is very helpful, thank you!

[ketiljo]

The response outside +/- 12 MHz is relatively flat. Remember that the LTE transmitter is up to 23 dBm. And if the 2.4 GHz BLE antenna is close to the LTE antenna, the power transfer between the two can be high. To avoid reduced sensitivity of the BLE receiver, you need a good band pass filter. Let's say the coupling between the LTE and BLE antenna is 30 dB and the LTE transmitter is at 23 dBm, you have -7 dBm on the BLE antenna. The selectivity is 52 dB so the sensitivity limit will be at -59 dBm. To get the full sensitivity of the BLE receiver of -95 dBm, you need a filter that can can attenuate >36 dB at the frequency of the LTE transmitter.

The closest to a blocking test that covers the LTE bands can be found in the ETSI EN 300 328 receiver blocking https://www.etsi.org/deliver/etsi_en/300300_300399/300328/02.02.02_60/en_300328v020202p.pdf, chapter 4.3.2.11. The nRF5240 will pass these tests without a filter, but the power levels will be higher for the blocker when the LTE module is close. 

[Stephen Snyder]

Knowing that the response is flat is very helpful.  Thanks for the example and details.

[Stephen Snyder]

Hi Ketiljo,

As we get closer to releasing hardware, I have a follow up question on this topic.

Reviewing ETSI EN 301 908-13 chapter 4.2.7 and looking at our antenna isolation measurements, it's possible we may have power in the 2.4GHz band on the LTE front end in excess of Pinterferer (-15dBm).  I'd like to have an additional filter to block 2.4GHz power to avoid desensing the LTE RX.  Can you recommend any filter part for this purpose?

Thanks!