RFM95 LoRa Shield low range problem


I recently carried out a range measurement with two Dragino LoRa shields with rfm95 modules (or rather PCBs derived from them with the same circuit). Already at 500m and direct line of sight the reception at SF11 and 64kHz was about -140dB (calculated by SNR and RSSI as stated in the datasheet) and hardly any more packets got through. i used 14db tx power at PA_BOOST and LnaBoost at the receiver. The maximum range should be significantly higher with these settings, does anyone have any ideas what the problem could be?

I will be happy to provide more information about hardware and software if necessary.

thanks in advance,

And just as importantly how did you implement the antenna? If ant good then if true line of sight the board implementation is poor. Moving from 125khz bw on 200khz channels to 64khz will extend range further (closer to narrow band implementation), but also takes it way from the LoRaWAN implementation used for TTN and hence out of scope here. As you be aware users the world over get much further with similar configurations, under LoRaWAN, so clearly you have something wrong. Distances of 10, 20, 30km or more not unusual…without resorting to HAB implementations! Over the summer I saw 65-75km achieved ground to ground down in Cornwall UK…ok from low hills and cliff tops :wink: Make sure your ant matching components, values and build implementation is correct and clean…

There’s Line of Sight and then there is Line of Sight - was this in a built up area, any suspicious looking masts in the area, what was the weather like (ie, raining hard) - many different factors. But 500m sort of implies that no antennas were involved so checking all the connections is a good idea. My HAB record from a hand-cut wire antenna on payload to a mag-mount antenna at home was ~117km, so I definitely know they can go a fair distance.

Well, if its two LoRa modules your using them in point to point mode, so not for TTN, so not normally supported on here.

One of the potential causes of very short range is reduced power output, possibly caused by operating them without antennas.

Unfortunatly there is not a lot of point persuing why you have short range until you have verified that the modules are indeed transmitting 14dBm and for that you really need an RF power meter.

Keep in mind that this is the TTN forum, and TTN is LoRaWAN.

So node-to-node experiments are only meaningfully on topic if they are modeling LoRaWAN usage which is between a node and a gateway. There are indeed reasons (portability and channel measurements) to consider using a node radio for survey, etc type tests, but using radios in ways that LoRaWAN doesn’t is a bit divergent.

This is not a channel bandwidth used by LoRaWAN

Nor is it a very practical one. Throughput may be comparable to SF12 BW 125 (which is already painfully slow and takes for ever just to move the LoRaWAN framing even before considering any useful data), but by going narrow band you may be making yourself even more susceptible to frequency error. There are particular radio settings appropriate to the slowest modes that LoRaWAN does use, which are documented in the recommended settings for LoRaWAN app note. Also you probably want a TCXO when using high spreading factors.

Thank you very much for all the suggestions. I will take a closer look at the antennas and the actual radiated power. I’d be surprised if it’s the circuit itself, as it’s derived directly from the Dragino LoRa shield for Arduino, but I’ll do some more comparative tests with the original modules. The nodes are to be used for both LoRaWAN and individual connections. I have only carried out the tests for the time being with arbitrary values for a fairly large range. In a few days I will be able to take further measurements and examine the hardware again. I’ll get back to you if I have further questions or findings, and maybe it will help others in the development of nodes with similar issues.