Feather LTE RF Circuitry Design - Band Preference?

galeap

Hi Jared,

Firstly, thanks for all your great work and generosity in sharing all of your hard work with the greater community.

I have a couple of questions regarding the LTE antenna and associated matching circuit on the nRF9160 Feather that I hope you can help me with.
Reading through your original Feather build blog, I noted a section where you highlighted that the RF circuitry design for the feather was “not straight forward” and you included a reference to the Nordic Thingy91 RF schematic. In that circuit Nordic used a switched multiplexing arrangement to cater for all the supported operating bands. You mentioned that you forwent this design arrangement due to space limitations on the Feather.
Given this, what specific LTE Band was the PY matching circuit components selected for on the Feather? As I can’t see how the capacitor and inductor values selected compare against those shown on the Thingy91 circuit (In particular for Band28 which I am specifically interested in).

My IoT application sees the Feather installed in an unfavorable LTE comms location (low to the ground or in cable pits), and with the supplied Unictron WC12H0 antenna I’m ending up with a weak signal with RSRP around -112dBm just testing indoors in my office. However, I happened to noticed that the nRF9160DK on its internal antenna (ethertronics PB22601) connects with a stronger signal that is on average around 20dB better connecting with an RSRP of around -92dBm.
I am now trying to work out what I may be able to do improve my connection strength. I’ve tried a few different antennas that had different gain values and radiation patterns. I’ve tried to select ones with higher gains shown on the datasheet for the 700Mhz band (B28) that I am interested in. But so far, I have been unable to close the signal strength gap in comparison to the DK. I’m by no means an antenna expert, and am now reaching for straws and looking for circuit differences that may be the cause of my observation, and that lead me to the matching circuit and the questions above.

Thanks for any time and guidance and help you can provide on the above.

jaredwolff

@galeap you could always remove all components in the matching circuit and replace C1 with a 0 ohm. Do this at your own risk though as modifying your board isn’t covered.

I’ll do this myself in the near future to compare. For an external antenna that is matched to 50 ohm a matching circuit is usually not needed. The new nRF91×1 design doesn’t have matching circuitry.

galeap

Thanks for the super quick response Jared and test idea.

I was actually wondering about the need to have the matching circuit components if the the antenna has a 50 ohm impedance. Yesterday I received a SWF Connector/SMA test lead that I ordered to see what would happen if I connect my current external antenna onto the 9160DK. Looking at the 9160DK board’s circuit, this should place my external antenna direct onto the 9160 antenna input pin with no matching circuit. I do these tests and make the mod you suggested to the Feather board and compare the results.

Out of curiosity though, was the Feather’s RF design more aimed at one LTE Band over others? How did you come to select the PY circuit component values? You design blog page is great, but the section on the RF feels like a murder mystery book missing the last page of who did it 🙂

Thanks again for your help and input.

jaredwolff

galeap it was an artifact of an older design. I’m also very used to have pi network for Bluetooth designs so I never thought much about it!

AchimKraus

The new nRF91×1 design doesn’t have matching circuitry

Just as note:
At least on my side, the new (beta) nRF9161 feather seems to require less power. Even after 10 days the battery level doesn’t diff for 1% . Using a nRF9160 feather the 1% was reached after 6 days.

Edited:
:-): OK, I just did a mistake, my nRF9161 reports (wrong) to run on external power source ;-). Therefore the device didn’t calculate a forecast, even if the 1% of the battery energy was used (-:.

jaredwolff

I do not see any crazy improvement in RSRQ on band 12. Seems like every time I run AT+CESQ the results are all over the map:

Before:

AT+CEREG=5
OK
AT+CFUN=1
OK
+CEREG: 2,"0608","00CC2911",7
+CEREG: 5,"0608","00CC2911",7,,,"11100000","11100000"
AT+CESQ
+CESQ: 99,99,255,255,24,44
OK
AT+CESQ
+CESQ: 99,99,255,255,22,42
OK
AT+CESQ
+CESQ: 99,99,255,255,20,42

After:

AT+CEREG=5
OK
AT+CFUN=1
OK
+CEREG: 2,"0608","00CC2911",7
+CEREG: 5,"0608","00CC2911",7,,,"11100000","11100000"
AT+CESQ
+CESQ: 99,99,255,255,25,43
OK
AT+CESQ
+CESQ: 99,99,255,255,24,42
OK
AT+CESQ
+CESQ: 99,99,255,255,24,42
OK
AT+CESQ
+CESQ: 99,99,255,255,24,42
OK
AT+CESQ
+CESQ: 99,99,255,255,24,43

Do you know what band your device is trying to use @galeap?

jaredwolff

The performance is also highly dependent on the properties of the antenna you’re using. For instance here is the return loss & vswr for H2B4MH1F2F0100

Depending on what band your device is using, your performance will be different. Closer to VSWR of 1 and anywhere near -20dB for return loss is going to yield the best performance.

Also antenna placement will affect the antenna tuning. For best performance it should be mounted away from anything metallic. At least A few inches.

Band	Cat-M1 (LTE-M)	Cat-NB1/NB2 (NB-IoT)	Frequencies (Uplink/Downlink)
1	✔	✔	1920-1980 MHz / 2110-2170 MHz
2	✔	✔	1850-1910 MHz / 1930-1990 MHz
3	✔	✔	1710-1785 MHz / 1805-1880 MHz
4	✔	✔	1710-1755 MHz / 2110-2155 MHz
5	✔	✔	824-849 MHz / 869-894 MHz
8	✔	✔	880-915 MHz / 925-960 MHz
12	✔	✔	698-716 MHz / 728-746 MHz
13	✔	✔	777-787 MHz / 746-756 MHz
14	✔	✔	758-768 MHz / 788-798 MHz
17		✔	704-716 MHz / 734-746 MHz
18	✔	✔	815-830 MHz / 860-875 MHz
19	✔	✔	830-845 MHz / 875-890 MHz
20	✔	✔	832-862 MHz / 791-821 MHz
25	✔	✔	1850-1915 MHz / 1930-1995 MHz
26	✔	✔	814-849 MHz / 859-894 MHz
28	✔	✔	703-748 MHz / 758-803 MHz
66	✔	✔	1710-1780 MHz / 2110-2200 MHz