I am using a Raspberry Pi 3B+ along with a Dragino PG1302 HAT module (868 mhz). I managed to add the gateway to the TTN network. As a newcomer to LoRa, I connected the gateway’s antenna to the PCIE card as shown in the picture below.
The gateway is being used with several devices: a Rak7201 4-button remote, a TTGO T3 module, and a Heltec LoRa module. Unfortunately, the range of my LoRa Gateway is severely limited. I can barely achieve a range of 50 meters. When the device is located next to the gateway, the RSSI is -49 and the SNR is 14.5. However, when I move it to the next room, the RSSI drops to -77, and the SNR is 13.75. When I go outside and walk approximately 50 meters away, the devices are no longer able to connect.
Is there something I am missing or any additional steps I should take to improve the range of my LoRa Gateway? Unfortunately other TTN Gateways are way to far so I need to rely on my own TTN Gateway. I have placed the gateway in the attic of the house with the default antenna. Without the antenna the RSSI quickly goes to -119db within a couple of meters so it appears the antenna is connected properly.
Is system configured to use the u.fl or to feed instead the SMA on the HAT? Signal will couple to either depending on config - most designs use a 0ohm resistor to redirect and a low power signal may get through either route - leading to a signal that is -10 - -30db lower than expected.
We normall recommend not running in close proximity as this typically leads to front end overload at gw or device or both, and channel bleed… in either case that often inhibits joining. We would suggest min 3m apart better yet 5-15m with an absorber such as a wall or thick window in between, the -49 is high but suggest they are not on top of each other or the ant’s are not aligned (say vertical vs horizontal), as typically we figures from -15–35 when very close and aligned. -77 is reasonable and is in the Goldilock’s band (~ -60 - -105dbm) that usually means your are debugging device issues vs fighting RF problems, however, -77 might even be a bit low for just one wall depending on construction…
That will in turn be determined by the construction of inside to outside partitioning - metal clad walls, foil back plasterboard, high metal/mineral content roofing or cladding/insulation felt etc.
This raises another question - how well connected is ant…make sure ‘tool’ tight not just finger tight, and double check you have correct polarity of SMA connections - male to female whether standard SMA or RP-SMA… a typical mistake is female to female (i.e. no direct center pin connection as signal will then capacitively couple at low level - similar to the 0ohm resistor problem on u.fl to SMA redirect and you will then experience a connection but loose anywhere from -10 to -35+dbm of signal. Effectively behaviour will be as if no ant connected at a distance/in far field…ironically LoRa sensitivity so good that systems will still ‘connect’ when in close proximity even with no ant, which brings us to
DONT do that! Running a device or GW without ant is a recipe for a dead or impaired device…had you run initially without an antenna? That could explain reduced sensitivity as when TXing with no load almost all the TX energy is reflected back into the front end causing burn out or inherent damage that can lead to reduced operation and potential long term reliability/failure mechanisms…
Yes, I checked female-male connections and have antennas on the u.fl and on the SMA connector. But the PG1302 seems to listen only to the u.fl. I wanted to know how I can set the PG1302 to only listen to SMA.
You can’t. The concentrator is only connected to the u.fl connector. If you want to use the SMA connector on the carrier you need to connect the u.fl connector on the concentrator board to the u.fl connector on the carrier. However, that will incur additional signal loss so isn’t recommended.
I had also a lot of Issues with this Board.
First, the Documentation (Link) said nothing about the U.FL Connector thing.
Therefor I ran the Board for a few hours without Antenna.
Now I connected the antenna, I get all my sensors, but in general I receive much less packages then before:
Maybe this is also software or configuration related? I do not use the draginofwd software, since it does not work on debian 11. I use this software instead.
Looks or is the same? If it only looks the same there might be significant differences underneath the shell. There should be an (very old by now) post on this forum with dissected antennas showing what is inside and how buyers are fooled…
What antenna were you using and what are you using now? Is the antenna in the same place and same orientation?
The antenna is connected by a pigtail with a SMA-connector. Is this a SMA-f connector? If it is, is the antenna equipped with a SMA-m?
It is possible to mount a R-SMA antenna to a SMA-f. In this case the inner conductor will not be connected and you receive no or only a very few nodes.
Mystery solved, the U.FL to SMA pigtail had a loose connection, I really tightened the antenna on the SMA socket. Now it looks like bevore or even better:
