N chasis female vs SMA chasis
@lex_ph2lb Thanks for your explanation in detail. 
yes, I plan testing the antenna for the tth_zh heartbeat, which should be received by many gateways in town.
The I prefer the male SMA because I’ve u.fl to female RP SMA which I use to connect to the (male) antenna
My problem at the monet is, that I couldn’t find cable ends with holes of 2.5/3mm; so I’ve probably to solder the ground plate wires directly on the SMA.
@BoRRoZ Indeed size does matter 
@ursm Soldering direct to the SMA is not a problem. But try not to melt the dielectric.
a while back I bought this set, thinking off converting it to 868 gp
but I thought the connector is pl259… which it isn’t 
but the quality is ok… maybe you have an idea ?
center is 4 mm
1 Like
First try on a sim of a 868Mhz GP antenne with 4NEC2.
http://www.ph2lb.nl/blog/index.php?page=simulation
Think I have to read the manual first
2 Likes
I have tried to simulated a 868Mhz groundplane in freespace, and 0.25m, 0.5m, 1m and 5m above average ground.
The results and the NEC files can be found on my 868 Mhz groundplane simulation page : http://www.ph2lb.nl/blog/index.php?page=simulation
GP in Free space
GP 0.25m above ground
GP 0.5m above ground
GP 1m above ground
GP 5m above ground
Conclusion : GP preform better above ground and height does matter.
Fancy 3D pictures on my page.
2 Likes
This is my Antenne.
1/4 Lambda 900Mhz GSM Cellular Car Antenna.
2 Likes
Thank you for the design! Here’s my version, let’s see how it does.
6 Likes
Looks good. Keep us posted on you’re experience with it.
1 Like
Results are very encouraging. I stuck it to the drain pipe with three ground planes instead of four and two meters of coax cable to a single channel (Jaap Braam repo) wemos. Before the antenna was a vertical 868MHz whip behind the glass of the window to the right.
The range seems to have doubled but needs more testing 
https://ttnmapper.org/?gateway=A020A6F42F21FAC2&type=radar
3 Likes
A materials question for Lex or any of the other clever RF guys here…
I have lots of old coaxial cable I was planning to scrap after I rebuilt my TV antenna. If I take three pieces of the solid copper core, anchor at one end in a vice and twist these tightly from the other end using a drill they form a very stable twisted wire, even more stable if the ends are tinned with solder. OK so it was a quiet night for TV programming yesterday, maybe I need to get out a bit more!
Anyway, in terms of RF and an antenna like this ground plane example, is there a difference in the way this 3 x twisted copper wire will behave compared to solid brazing rod for example?
I have a few N connectors on order so should I try this idea or is it doomed to failure by design?
Thanks in advance
Garry
1 Like
Well Garry, you can always try.
RF (or any AC current) has the strange effect that it loves to move around the outer skin of a wire so when twisting firmly them tightly i will behave like a solid wire. soldering the ends with will strengthen the wire (preventing any untwisting due o wind vibrations). The only thing you will get when using plain copper wire is a light layer of oxidation. This could on a very small scale change the behavior of the antenna.
One of my main HF antenna’s uses a non solid wire (3 strings iron and 3 strings silver copper wire) as antenna wire. I can’t find any difference in TX or RX between the original solid wire and this 6 string twisted wire.
But like I always say : hamradio so also TTN is all about experimenting. And remember FAIL means First Attempt In Learning.
Thanks!
Without your exotic equipment to test my antenna I fear my feedback will be binary at best but I will post something here.
Garry
2 Likes
The theory is far over my hat, but the building instructions are very clear.
Can I connect the Gateway via a coax cable (50 ohm?) with this antenna and what are the implications regarding to the length of the cable?
1 Like
Yes you can connect the antenna to the gateway with a coax cable and the length of the cable does has his influence.
The longer the cable the more attenuation to be expected. But also the type of cable has his influence.
For example a consumer cable like the RG-58 has a attenuation of 70 dB for a length of 100m at the EU LoRa frequencies. So a cable with a length of 10m would result in a attenuation of almost 7 dB (a -7dB amplification). Meaning that only 20% of the power put into the cable will be fed into the antenna (and vice versa).
When using a better cable like the AirCell 7 coax cable the attenuation for a 10m length will be just 2 dB (a -2dB amplification). Meaning that 63% of the power put into the cable will be fed into the antenna.
Formulas:
dB (power) = 10log(Pout/Pin)
for example :
Pin = 5W Pout = 1W (very heavy cable loses).
db (power) = 10log(1/5) = -6.989 aprx -7 dB
Pout/Pin = 10^( dB (power) / 10 )
for example : Pout/Pin = 10^(-7 dB/10) = 0.1995 aprx 20%.transfer meaning 80% loss.
It sounds like much even with LoRaWAN who has a good signal to noise ratio a RG-58 could work out.
For more info check out this page on my website : http://www.bolkesteijn.nl/blog/index.php?page=good-cables
3 Likes
… and that trial here.
https://www.thethingsnetwork.org/community/dublin/post/testing-learning-and-lora-wardriving
I will report back if I get something consistent from this, pro or con.
Thanks again
Garry
2 Likes
… I am no RF engineer but for 80mm of copper twisted two-up this works very well, great design.
I have no testing equipment so I can’t say how well it is tuned to 868mhz or how accurately I built this but the difference is clear. I was finding over 400m was hard to achieve, even with the node set to SF12 and the portable gateway placed high inside the roof void like this antenna. RSSI was never lower than -120db but it either worked or it didn’t but the range was disappointing.
The result with the same home made ESP8266 gateway in the same less than ideal position (geographically) but using this ground plane antenna is 4-5km already and I haven’t sough out further suitable places to try. The RSSI is agian in the -114 or -117db range so some room left yet I guess?
I may actually put this antenna 5m above my roof on the old TV antenna pole and try it out on the RAK831 for more testing. I suspect there is plenty more to learn about the node transmit antenna too… wonders if Ryanair will be flying to somewhere near the Belgium/Nederland antenna hangout?
Hope this helps? G
2 Likes
hello,
what the lengths of wires should be for 915Mhz? I suppose - ground plane wires should be 82mm and central wire - 76mm (82-6)?
You could give that a try.
If your building something like this;
Then with 0.25 wavelength radials, the antenna peaks when the vertical radiator is 0.228 wavelength (average over 3 modules). This is shorter than 0.25, but the diameter of the wire etc does reduce the dimensions a bit.
With the radiator at 0.25 wavelength you might loose 2dB or so.
Antenna tests done at 868Mhz, with a LoRa RFM96 as transmitter, antenna directly attached to module, results at 915Mhz should be similar.
1 Like