You talk about 3km and a hill, this should already ring a bell, what other cutter do you have? Buildings, trees or?
Your evaluation map is 1.88km, that is just about level with a hill, again what clutter are there besides the hill? All obstacles - clutter do attenuate, block RF, some more some less.
Yes I’m aware you’re communicating with the network server and not the gateway itself, I was just trying to describe my issue.
I found the elevation map data after making my post so yes I agree, that is probably part of the problem.
I’m just after any advice and information I can get. I’ll research the Fresnel Zone, thanks.
There’s not much between the two points, some residential buildings spread out on 10 to 50 acre blocks of land and not too many trees, but there’s certainly some the signal would have to pass though.
I obviously can’t really show the area in depth without giving out my location which I won’t do but from what I’ve read up today I don’t believe there’s anything wrong with the gateway setup, it’s most likely going to be a combination of the land and obstacles.
I think I’ll try get the antenna a bit higher up and keep trying.
Elevation will defiantly help, with LOS and fresnel zone, normally the first fresnel zone free of obstacles, but LoRaWan are a bit forgiving.
What is the RSSI, SNR of the node say 50-100m from the gateway? Give you a good idea if the antennas are good and the noise in the area.
Thanks,
So when I first made the post I had an RSSI of -72dBm and an SNR of 13.2dB. This would have been approximately 50 to 80 meters from the gateway antenna LOS.
I have since moved it probably another 80 meters. Say approximately 150 meters LOS from the antenna. The RSSI is now -84dBm and the SNR is 10.5dB.
I’ll try and get some more accurate measurements and also some readings from a further distance and post them here.
Thanks. 
If you double the distance, the RSSI should decrease by 6dB (in theory). But still there are more than 20dB missing. Try to turn the node by 90 degrees and observe what happens to RSSI.
As @wolfp has mentioned, double the distance and the RSSI should drop by 6dBm.
So your RSSI numbers are either wrong or there are obstructions in the way or you dont have true LOS, since the difference appears to be 12dBm.
I’ll get some more accurate measurements and post them here.
Could a faulty coax link between the antenna and the gateway cause unexpected drops in RSSI vs distance? Just out of curiosity. Poor connection in one of the joins, moisture, etc.
Thanks.
I’m logging the data to my database and I took a screenshot of it to post here. It does seem to move around a lot, before and after I relocated it:
It has always been clear line of sight. The antenna is above the peak of the shed roof as seen in my previous picture. The node this data is coming from has always been about 1 meter off the ground, but clear LOS.
Nope.
I would not care about ±3dB of RSSI or SNR. My gateway receives a Node that is transmitting a 19 bytes message every 12 seconds. It seems to be in a fixed position but RSSI (-116 dBm) and SNR (-3dB) are changing by ±3dB every reception.
Sometimes the SNR is even worse - the message can not be received (to much “noise”).
I don’t know where and how far away the node is situated, if I knew the owner I would tell him about FUP. On the other hand this is an interesting beacon.
If you had 3 gateways and could log a few 1000’s of it’s messages, you can get fairly close on the location of it. (but you will use RSSI for location, you could be out a few 100m thought)
What is the MAC ADR settings for the node?
And are all the RSSI from the same gateway? As if two gateways RX the node the NS dont always show the gateway with the lowest RSSI first in the JSON
Yeah all from the one gateway.
I’ll check for the information tomorrow.
So I’m currently at work but noticed I’ve lost the node, no RX at all for the last 2 hours.
This actually happened a couple days ago briefly.
I still suspect something isn’t right… I have an MKR WAN 1310 I’ve also had intermittent issues with. The LHT65 is brand new but I can’t rule it out either.
The last RSSI/SNR was -85dBm/8.5dB.
I suppose it could also be interference from something… very odd. I wish I had the equipment and know how to work this out. 
I’m going to order new, quality coax and adapters I think.
Edit: Getting received data sporadically every couple of hours instead of every 20 minutes all day so far.
If problem with multiple nodes suspicion turns to the gw(*) and any associated internet backhaul. What do the gw logs show wrt status/traffic etc. what does the TTN console tell you wrt gw connectivity……do the node ‘outages’ match to gw issues? Where in the world are you, what backhaul are you using?
Appreciate that, but the more detail and surrounding info we get the better the prospect for the forum volunteers helping find a resolution or providing effective advice. We don’t need your burglar alarm code 
(*) if gw not your own can you contact gw owner to ask them to share appropriate info. Is there a local TTN community, are you/gw owners members? Perhaps others in the area/community see similar drop outs….may help track down cause….gw, backhaul, interferers….
I’m not sure how to check the gateway logs on the Raspberry Pi with the RAK2245. I’ll definitely investigate though.
I previously had a RAK831 running on AS923 and never had an issue like this, with the LHT65 in the same location. The gateway I’m using now is a new Raspberry Pi + RAK2245 (AU915) Pi Hat and new coax cables.
The Pi is physically connected via Ethernet back into my network, and I have an ADSL2+ back-haul running at about 22Mbps downstream and 1Mbps upstream. The line has 100% uptime and runs no CRC/FEC errors at all.
There’s no local TTN community, I’m in a rural location and it’s a small town, so I’m kind of on my own with this one.
But I’ll try find some logs tonight as soon as I get home. Thanks.
Last RX was 4 hours ago at the moment. Should be every 20 minutes.
Probably the the easiest way to test if the RSSI your seeing is about right is to use a couple of Arduinos with LoRa modules, one programmed as a transmitter the other as receiver and see what the RSSI is over the same distance when simple 1/4 wave wires are used as antennas.
Hey guys,
I thought I’d give an update, would love to hear your opinions.
Since my last post here nearly two weeks ago, I had the gateway continuously running and the node was in the same location. The RSSI and SNR was still all over the place, and sometimes I wouldn’t receive any data from the nodes for hours at a time even though they are sending data every 20 minutes.
I tried programming both nodes (Two LHT65 and one MKR WAN 1310) to use AS923 (instead of AU915) and configured the gateway to use AS923 as well and got the exact same results/problems keeping all other variables the same.
I tried different coax, antennas, u.fl → sma, etc. Nothing changed anything. Confirmed good continuity from RAK PCB (pad before u.fl connector) all the way to the N type connector at the antenna. I even changed the Pi 2 to a Pi 3B+ that was being used with the Pi Hat just in case the Pi 2 was having SPI issues.
So I went and bought a RAK WisGate Edge Lite 2 from here.
I’m using it in the exact same location, same coax etc, and the node is in the same spot and I’m consistently getting an RSSI of -88dB every time, and an SNR of either 13.5dB or 13.25dB every time. Everything seems very stable and so far not a single missed RX like I was getting on the RAK2245 for hours at a time.
So I’m going to continue to monitor it. This has been frustrating as I have limited resources and knowledge to test things to try and figure out the issues, but do you think this may indicate that the RAK2245 Pi Hat was somehow at fault?
Never had an issue with the RAK831 and so far the WisGate Edge is performing well. I did originally take extreme care setting up the RAK2245. I wore an anti-static strap, didn’t ever power it up without an antenna connected, so maybe something was wrong out of the box?
Thanks
Thats very suggestive of an RF breakdown of some type, faulty antenna or connector etc. Possibly the Gateway board also.
If the ‘node’ is at 80m then calculations suggest that the RSSI @ 14dBm @ 868 with 2dBi antennas and good line of sight between the Gateway antenna and node antenna should be -51dBm (since the free space loss is -65dBm), which is a long way from the -88dBm you are reporting. However measurements I have done at 434Mhz suggest the RSSI indicated by the SX127x is considerably in error, maybe Gateways are the same.
It was later moved to ~150m which may help account for some additional loss, also assuming both antenna are vertically aligned, if there is a vertical offset then that may account for some additional loss due to polar plots of both antennas, this would be worse if antennas close - original 50-80m - deminishing in impact as they increasingly move into the far field and relative angle deminishes I guess.