Co-linear antenna modeling2-1.pdf (700.0 KB)
Thanks @kylix, if the black device was not a balun but a piece of plastic I wonder how the IRNAS gateway performed as the VSWR without a groundplane is high. 4:1 @LoRaTracker tested these antennas some time ago and reported the results to be something like 4:1
@dicktonyboy, thanks for the pdf. Will test and see how close the predicted VSWR curve is to my build. Wasn’t able to reply earlier, was attending our local TTN/IOT Meetup.
Build it and check the sensitivity of each element to change. Try to only change one parameter at a time to isolate performance changes and then see if you can scale to move the VSWR to the frequency of your determination but use a ground plane everywhere and observe the 10 lambda rule. - easy eh? see my time sump loom!
thanks @dicktonyboy, will do. Before the 915 version was published I scaled the 868 version to 915MHz. I’ll go back and build the 868MHz version and compare to the modelled performance. I’ll then start moving the frequency. interesting the published 915 version has significantly smaller diameter loops, never understood how they came about. Far more than simple adjusting for different frequency, almost designed to a different concept. Anyway, thanks again for the document, i’ll give it a try.
The ‘touchy-feely’ argument is to consider the wave travelling straight up the plain wire aerial and then look at the delay incurred as it has to travel around the loop - which is ‘orthogonal’ to the main wire. What happens if the loop is one quarter, one half and one wavelength long? c = f λ (remembering that c is slower in cable by the velocity factor).
Today I tested this https://www.aliexpress.com/store/product/868MHz-high-gain15dBi-omni-base-fiberglass-antenna-GSM-900M-outdoor-roof-monitor-antenna-N-female/413214_32772798538.html antenna for a friend who bought it via Aliexpress.
The result is terrible, gain performance is not 15dBi but 1 to 2 dB less than a simple 3 dollar dipole . Noone should ever buy a “glide base” antenna, it has a totally fake gain figure.
I must say that the VSWR was good, better as specified. Almost as good as a dummy load
@dicktonyboy, understand about the phasing loops between the antenna sections so I rescaled the 28mm OD loop in the 868MHz design to 26.5mm in my 915MHz adaption. Then when Fabien’s 915MHz design was published it had 13mm OD loops!!! Could not reconcile such small loops. Would mean the velocity factor > 1 and that can’t happen.
So i got myself a new gateway, a Laird Sentrius RG186, but the range is a bit disappointing. Could be my lora test device of course, it has a 1/4 wire antenna. But about 500 meters is max (@ SF12).
I want to get an external antenna and think i have to decide between these 2:
Any other recommendations? Will these antennas increase range significantly?
that depends on where and how you place them.
in general … don’t use a very long cable and use a good quality cable… with low loss @ 900 mhz.
the antenna should be place on a spot with no other object within1 mtr or so
ideetron is a lora specialist shop so I think they can advise you about the cable type to
Also depends on what antennas yoyu are currently using and where they are located.
I’m using the default antenna that came with the gateway, it’s inside about 4 meters high under the rooftop.
Are you getting good coverage from this? I have several roof-space mounted GW’s around the country (including a bunch of the Lairds) and in each case I have had to be careful with placement due to various roof constructions, especially in modern or recently refurbished buildings.
Often the roof itself will be heavily lined with thermal insulation that these days is increasingly made with an additional layer of metal foil - a great RF blocker!..sometimes several layers. (Foil backed plaster-board/dry-lining increasingly common also). The foil doesn’t just ‘block’ but if too close to ant seems to help ‘detune’ also. In other instances have discovered a layer of mineral loaded water resistant ‘felt’ to ensure good weather proofing - again some of these materials appear to badly affect rf propagation. In these cases have found that rather than going to the highest roof point assuming height = better range, have found dropping to attic floor level has actually increased range as signal ingress/egress under insulation better.
Best yet a small hole under the eves, later closed to stop insect ingress, allowing use of external antenna on short run of ultra low loss feed cable usually works wonders (typically 10-30db signal improvement )
There’s actually no insulation besides the asbestos roofing… Don’t know if that blocks the signal.
I’m going to put an antenna on the roof anyway, but if i’m going to spend money on that i 'd rather buy a good one.
With these 2 options:
I think the GP901C is the better one? There’s also a GP868C, wouldn’t that outperform the GP901C?
The GP901 C has a Frequency Band of 870-960 MHz where the GP868 C has 835-900 MHz.
I think the GP868 is better suited for Lora in Europe.
Yes, if you have managed to increase c - you’d better prepare the universe What link of Fabiens are you referring to exactly? I’m trying to gain access to a radio lab here over the student holiday period. Annoyingly we did just build an array which was incredible at about 550M and then we scaled it to 868MHz - I’ve lost the paperwork but the unit is up on the roof. @loratracker has achieved some remarkable results too (but is currently side-tacked, deep in a local, (unused), railway tunnel proving that the Semtech chip has a noise issue).
Here is the link to the two different designs. its on the IRNAS GitHub site https://github.com/IRNAS/ttn-irnas-gw
@dicktonyboy, hope you can find the paperwork for the converted 550M antenna. Guess you’ll be shimmying up a drain pipe to measure. Over the weekend I will also try a Coaxial Colinear, bit fiddly to build but hope the VSWR match is better than the 2:1 in Fabien’s modelling work. If it looks promising, I’ll test the gain and pattern.
I hope to deploy some more gateways using @Charles’s pcb to integrate Rak831’s with Pi Zeros. I ordered an antenna with some Rak831’s so would be interested in data on tuning or performance?
I just received one last week but I haven’t done any testing yet…
My gateways have SMA pigtails and the RAK antenna is delivered with an RP-SMA connector, so I need some additional bits to swap easily…