Antenna advice


I recently started toying around with LORA. I send my first messages using a rn2483 hooked up to a arduino pro mini, as described here:

It worked pretty well, but I noticed that sometimes my messages got lost. I figured the problem would be the antenna, since I just used a 8.6cm piece of solid core wire.

To improved this, I designed a simple PCB that gave me the ability to hook up a proper antenna. I bought this antenna for that purpose:,searchweb201602_4_10065_10068_10000074_10000032_119_10000025_10000029_430_10000028_10060_10000067_10062_10056_10055_10000062_10054_301_10059_10099_10000022_10000013_10103_10102_10000016_10096_10000018_10000019_10000056_10000059_10052_10053_10107_10050_10106_10051_10000053_10000007_10000050_10084_10083_10000047_10080_10082_10081_10110_10111_10112_10113_10114_10115_10037_10033_10000041_10000044_10078_10079_10077_10000038_429_10073_10000035_10121-10033_10037_10077,searchweb201603_7,afswitch_1,single_sort_1_default&btsid=3ddb7acb-7f6a-4fcd-b87f-5de46314e2e0

unfortunately, with this antenna attached, almost no messages are getting through, only every now and then a message is actually received by ttn.

I made one error in my design, both the pcb connector and the antenna have a female connector, to solve that I used this:

Any idea if that could cause a problem, or are these chinese antennas just crap? and if so, any advice for a better one?


often you have to verify what you got is what you ordered or were expecting. In this case, pop the top of the antenna off and measure the wire length above the brass barrel. Pop it off by looking for the seam just beyond the hinge and then gently pull towards the top end while flexing it at that joint.

I found that the 915MHz antenna I purchased was in fact a 2.4GHz antenna. 915MHz is 78mm long so I had to very carefully cut the brass tube back to expose 78mm of wire. I used a dremel and a very thin and very sharp metal saw. Fingers and eyes need protection if attempting the same.

(Jcavera) #3

Use an SDR receiver to see if (a) anything is coming out of the tranceiver and (b) what the relative signal strength (with each antenna) is. As for antennas in general, cheap rubber ducks are usually worthless. If you're wanting something better than improvised wire, try adding a shielded feed-line and ground plane.


Please post a picture or the layout of your PCB. Improper HF design can lead to serious signal degradation.

Coincidentally, I just received the same antennas this week. I have nodes that function quite well with some China Export rubber ducky antennas. I'll give these antennas a try next week and let you know how they work out for me.


I did some testing with this and another antenna on two RFM95W nodes (Adafruit breakouts).
Antenna #1: 5cm RA rubber ducky antenna, €2,2/piece
Antenna #2: said 19cm RA whip antenna, €2,18/piece

I used it with the Radiohead library and, for sake of testing, with PA_BOOST enabled up to 23dbm. My results are in RSSI, with -100 being the worst and -22 being the best attainable results.

  • Rubber ducky on both nodes
    -22 in same room
    -73 with two concrete floors between them
    -100 with three concrete floors between them (node in basement)

  • Rubber ducky on one node
    -22 in same room
    -67 with two concrete floors between them
    -97 with three concrete floors between them (node in basement)

  • Whip antenna on both nodes
    -22 in same room
    -62 with two concrete floors between them
    -94 with three concrete floors between them (node in basement)

So despite the whip antenna being significantly larger (half-wave actually) the results are consistently around only 4% improvement. Still can be worth it sometimes. With the two rubber duckys the node in basement was not reliable, but with the whip antenna it was.


different antenna wave length designs will give you different radiation patterns. Open field testing is a better test of antenna(s). FYI, a 5/8 wave antenna will flatten the lobes of the radiation pattern so more energy goes out horizontally and therefore better range in the horizontal plane would be possible.

If you are testing with antenna's vertically located relative to each other then you're not going to get much. Did you try polarizing the antenna's to each other in the plane of the devices. ie lay each antenna horizontal in the same direction(N/S/E/W)?