Author Topic: LoRa 868Mhz Antenna Troubles  (Read 1092 times)

Offline Joules

  • Hero Member
  • *****
  • Posts: 1271
  • Country: gb
LoRa 868Mhz Antenna Troubles
« on: December 10, 2023, 06:32:30 PM »
I got bitten by a new bug, looking at a comms system called Meshtastic.  Think of it like an open source pager using a mesh network of LoRa devices.   Ordered my ESP32's expecting a rough spiral antenna with heatshrink on it.   What arrived was a stubby antenna of much nicer quality with an SMA connector, meaning I can try different antenna.  Soooo, of course I am going to test this little stubby antenna, I rebuilt my electronics lab over Covid thinking about keeping my hand in and also being able to service kit we already have.  Yeah, also playing about ranks pretty highly too.

The stubby was listed as being 868Mhz suitable, well my VNA test would disagree.  Anyone familiar with CB should know getting the SWR low is the aim of the game, around 1:1 being ideal, 1:3 not great but workable.  The VSWR here is bad, 1:15.21  .... WTF....
Yep, looks like they grabbed these off the shelf regardless of frequency.  So I scratch my head for a bit and think, been playing winding my own copper inductors to play with.  My thoughts drift to magnetic loops where a small inductive loop resonates a larger loop.  Could I transpose that ?   After several attempts with different coils I settled on double helical coil, first thought was that the inner would cancel the outer, but since I hadn't tried it, give it a go.

After juggling the coil position, on the plastic toothpick I found the sweet spot VSWR now 1:2.52   That's a hell of a lot better than where I started.  The stubby is now exciting my little coil that is now acting as the antenna, it also provides a better impedance match to the stubby making it appear more like 50-ish ohms.

Proof is in the pudding, a great feature of these NanoVNA's is that you can also use them as a signal generator, you just need a spectrum analyser to view the result....

I was lucky getting some work in during Covid that gave me the opportunity to buy a spectrum analyser, yeah I have a TinySA and I use that to check what I hook up as it's much cheaper to blow up a TinySA than the Rigol.  Set the telescopic antenna to the wavelength of 868Mhz, about 345mm and look at the RF signal.  I saw an increase of 5dbm over not using the coupling coil, the coil has paper insulation to stop the coils shorting against the inner and outer windings.

Conclusion, it was great to play with my toys get an actual result, but by far I would be better making my own antenna's from scrap wire, they are hardly huge at these frequencies.   Enough distraction, back to getting Meshtastic understood and working.

Honour your mentors, and pay it forward.

Offline BaronJ

  • Full Member
  • ***
  • Posts: 243
  • Country: gb
  • Grumpy Old Git !
Re: LoRa 868Mhz Antenna Troubles
« Reply #1 on: December 11, 2023, 02:17:54 AM »
 Hi Joules,

I've attached a picture of an antenna that I've used very successfully with a change of element length for TV reception and WISP WiFi. I simply recalculated the element lengths to suit the frequency that I wanted.

 
   

2.4Ghz Colinear.pdf
« Last Edit: December 11, 2023, 02:47:01 AM by BaronJ »
Best Regards:
                     Baron

Offline Joules

  • Hero Member
  • *****
  • Posts: 1271
  • Country: gb
Re: LoRa 868Mhz Antenna Troubles
« Reply #2 on: December 11, 2023, 03:04:08 AM »
 :lol:  Cheers Baron, so I discovered a colinear antenna of sorts.  As maybe you can tell RF electronics is not an area I have expertise in.  Thanks for the info, I shall play further.
Honour your mentors, and pay it forward.

Offline Joules

  • Hero Member
  • *****
  • Posts: 1271
  • Country: gb
Re: LoRa 868Mhz Antenna Troubles
« Reply #3 on: December 15, 2023, 02:40:00 PM »
Here is my Frankenpager, LoRa 868 based communicator.  Chucked in a couple of boxes as I couldn't get things working to begin with.  These Lora units can be used for text messages line of sight, or telemetry data.  They have a surprisingly long range if clear view and much better antenna choice.
Honour your mentors, and pay it forward.

Offline AdeV

  • Madmodder Committee
  • Hero Member
  • *****
  • Posts: 2434
  • Country: gb
Re: LoRa 868Mhz Antenna Troubles
« Reply #4 on: December 18, 2023, 04:45:10 AM »
Depending on what you want to achieve, maybe take a look at PainlessMesh? It's an Arduino library which works on ESP8266s or ESP32s (the latter preferred), it uses wifi frequencies & therefore the antenna on most ESP32 dev boards works just fine.

It's not the same thing as Meshtastic (that seems to be a dedicated hw/sw unit?), as you have to write your own s/w around the PainlessMesh lib... but worth a look depending on your use case.
Cheers!
Ade.
--
Location: Wallasey, Merseyside. A long way from anywhere.
Occasionally: Zhengzhou, China. An even longer way from anywhere...

Offline Joules

  • Hero Member
  • *****
  • Posts: 1271
  • Country: gb
Re: LoRa 868Mhz Antenna Troubles
« Reply #5 on: December 22, 2023, 11:18:40 AM »
Cheers Ade, but it is the text communication I am after, Meshtastic being the initiator of this and then me going off on tangents playing with antenna's.

Example being, further development from the crap stubby to coupled coil stubby and now surgery......

To the stubby antenna that is...   I tested a PCB ultra wideband antenna and got better results than my modified stubby so figured I can do better. Cut up the stubby, unsolder the tiny helical they use and make a half wave dipole.  A small top hat drilled for the copper wire and larger hole inside to cover the pin and wire soldered together.  Turned from Acetal this is slide over the feed wire. Once in place I could fold that copper wire 90 degrees.  First test results on the NanoVNA proved disappointing with high SWR, it was then I rotated the wire to 90 degrees reference the ground plane.  This made a huge difference, again the NanoVNA allowed me to tune the angle just past 90 degrees.  Hot glue fixed everything in position and I could setup the antenna's I have for comparison testing.

I opted for the TinySA Ultra to compare the original stubby, then the UWB and finally my DIY  V dipole.  Screen shots of the results.  Not exactly rigorous scientific testing, but a comparison using the ESP32 board with a 1m patch lead on all antenna's.

The V dipole is a big advance over the stubby and my coil loaded version.  It was nice to have a "tiny" bit of machining for this.
« Last Edit: December 23, 2023, 06:02:55 PM by Joules »
Honour your mentors, and pay it forward.