[trust]

After having found what appeared to be temperature issues with some of the OPLINK air rcvr modules, I ordered several types of antennas and tested them - in this case on the base station module (on the theory that they are both the same xmtr/rcvr modules). The rcvr antenna is oriented at 90 degrees to the xmtr - this is normally the highest signal strength for a dipole like antenna. Antenna parallel to the ground. All tests with 10' spacing rcvr to base station.
The antennas tested are from left to right:
1) Antenna that came with OPLINK - when antenna parallel to rcvr antenna -40db on OPLINK sig strength

2) Version to try to match 1) above ordered on ebay link below - identical results to 1)
https://www.ebay.com/itm/Folding-433MHz-1-4-Wave-Antenna-2-dBi-SMA-Connector-105-mm/184610598934

3) Longer antenna on ebay - this seemed to be directional - when parallel, about -40db, but when end was pointed at rcvr, -32db - significant improvement
https://www.ebay.com/itm/433MHz-Rubber-Antenna-19cm-SMA-Connector/382614690368

4) Longest version - this too was directional as 3), but not as high gain - -36db
https://www.ebay.com/itm/SMA-M-VHF-UHF-Antenna-for-Baofeng-UV3R-Retevis-RT3S-YAESU-VX-3R-Two-Way-Radio-X2/362947038362
This antenna also seemed to shift around - sometimes signal strength would drop very low --60db or lower. All the others stayed pretty consistent in the -40db range or above.

[TheOtherCliff]

The rcvr antenna is oriented at 90 degrees to the xmtr

For linear antennas outside of near field effects (us at say 10 meters or more and away from clutter, normal flying) they should be oriented to be parallel for best reception. 

Antenna that came with OPLINK

In the beginning, OpLinks and Revos came with a T wire dipole.  They now come with "Small, cheap wire coil style" which is what I think these are.

Your medium length antenna might be a loaded antenna instead of the standard cheap simple coil antenna, but every 433 that looked like that that I have cut open (only 2) has been a coil antenna in the bottom of the bottom and the rest of the antenna empty.    The many different coil antenna case styles I have tried have varied from really bad to usable for local flying.  All my quads use the cheap coil type antenna for telemetry, but they aren't long range aircraft.  My fixed wing all use Retevis' whether they need it or not. 

For a test that is perhaps more like a range test set both ground and air to lowest power.  Try to find a place in the house for the air where it does not connect.  You may have to go outside.  Find a max range place for that set of antennas.  Now change either antenna and try again.  I bet your best range is with 2 Retevis', by far.    By my recollection I got 200+ meters outside at 1.6mW through trees and houses with a pair of Retevis' vertical / parallel at 1m (base) and 2m (hand carried airplane) off the ground.

[trust]

Just to be clear, I'm normally orienting antennas just like you see in the pix - parallel to each other. When I say 90 degrees, I mean 0 degrees would have the end pointing at the transmitter, 90 degrees is parallel.
True omnidirectional antennas shouldn't care about orientation, but with dipoles or antennas like these usually the poorest reception occurs when ends point at each other - so the antennas with gain are a bit of a surprise.
When the wx gets better I'll try some more outdoor tests. The Retevis are a bit on the heavy side - I was thinking more of using them for base stations, where weight doesn't matter. The pulsing signal levels are concerning though. Maybe I should move it further from the computer with a longer USB cable.

[TheOtherCliff]

There actually is no true 3D omnidirectional antenna for the same reason that a hairy ball must have a cowlick.

When standard literature talks about omnidirectional antennas, they mean in 2D, e.g. horizontally it radiates the same in all directions.  By this definition, linearly polarized antennas are omnidirectional.

Two factors affect antenna performance: radiation pattern and polarization.

Linearly polarized antennas (the ones we have here) have a radiation pattern that is like a doughnut with no radiation coming out of the top or bottom and very little when close to top or bottom.  All the radiation comes out the side of the antenna.

Polarization:  You can have two antennas that are parallel and pointing their sides at each other for strongest signal, but if you rotate them so that they still have their sides pointing at each other but are 90 degrees and no longer parallel, you have reached an orientation where the signal is strong but the polarization is weak and the antennas don't talk for polarization reasons.

Circularly polarized antennas still have the radiation pattern signal loss, but don't loose signal strength for polarization reasons like linears do.

When the antennas are close to each other or close to objects (within a couple wavelengths) (433MHz wavelength being about 70cm) there is a near field effect that causes signal strength to show unexpected variation.  Of course any reflective surface can cause destructive or constructive interference at even farther distances.

[trust]

Crossed dipoles come pretty close to omnidirectional, but circular polarized antennas tend to work better. The first VR systems I built years ago broadcast on 440 Mhz using crossed dipole antennas - used a yagi for rcvr antenna. I had a backpack I wore with the antenna fixed on it sitting above my head - by facing the direction of the aircraft, it would point the yagi at it.

Back in the 80s I built medium to high power full color laser projectors with all digital synthesizers I designed and built. Mostly xy scanned type.
But I also built a 3D full color laser video projector using self-resonant (8khz) scanners, which could rotate the beam polarization by 90 degrees every field of the video. Almost perfectly polarized. Wear polarized glasses to view in Technicolor red.

[TheOtherCliff]

I recently bought some more "Retevis RHD-771" 433MHz OpLink antennas from eBay and they test just as good (or better) than before.

This is the best 433 antenna I have found.  They are great for base stations or airplanes, but are a bit large (37.5cm/14.8inch) and heavy (36g/1.27oz) for small quads.

Code: [Select]

Today's 433 antenna test at 1.25mW
  quad in room perhaps 18m away indoors through 4 walls
  base station was always Retevis RHD-771

  Antenna type:            dB(m?) at GCS:
  ------------             -------------
  cheap coil type #1       -77 to -80
  cheap coil type #2       -76 to -79
  Nagoya NA-771            -58 to -59
  oldstock Retevis RHD-771 -56
  newstock Retevis RHD-771 -55 to -56
Note that this (-55dBm(good) - -78dBm(bad) = 23dBm) says that for each of two cheap coil antennas you replace with Retevis RHD-771 you will get about 15x better range.  (Double the range for each 6dBm.)  Replace both air and ground coil antennas with these and you get 225x better range.  I can't guarantee these numbers without testing more coil antennas and using better test equipment, but even if you only replace one cheap coil antenna and only get 10x better range... 

2 for USD $13 shipped.
https://www.ebay.com/itm/SMA-M-VHF-UHF-Antenna-for-Baofeng-UV3R-Retevis-RT3S-YAESU-VX-3R-Two-Way-Radio-X2/362947038362

How do I know what kind of antenna I have?
For our uses, a good antenna on 433MHz needs to be a half wave (37.5cm/14.8inch).  Quarter wave antennas (about half as long), such as Sanders style need a ground plain which we only have when using e.g. a magnetic mount quarter wave on a car roof, so no.  All plastic case 433 antennas I have taken apart have always been a cheap coil and a lot of empty space inside.

[jdl]

I vote for Retevis RHD-771, too. Excellent antennas for their price!

I have made some measurements, they prove to be very well tuned 430-440MHz antennas.





On the air side I prefer using DIY half-wave dipoles for planes (15+ km with stable control (1W Tx) and minor telemetry packet loss (0.1W Tx) @ 57600bps):

https://forum.librepilot.org/index.php?topic=4705.msg31806#msg31806





DIY coil-loaded dipoles on quads:

https://forum.librepilot.org/index.php?topic=4889.msg32763#msg32763









Even "Sander-style" quarter wave monopole works very well on my 7" quad, I'm using there the whole carbon-fibre frame as counter-poise (the frame is grounded) and I've carefully tuned the monopole length. The final result is almost the same if using a coil-loaded dipoles. Have flown to 5+ km without any control or telemetry issues.

[TheOtherCliff]

At about USD $6 each, the Retevis RHD-771 is very inexpensive.  It also is the best antenna I have found at even 5 times the price. 

I love that half loaded dipole!!!  When I first saw it (and couldn't see the load) I wondered if it was a J-pole.

For quarter wave antennas, I wonder about oddly shaped lobes below the plane (where we usually are) and a possible slight null at the plane, especially for our quads where the radials are oddly shaped and sized.  Probably work best with a 330mm class quad?

I'd be very interested to see the dB differences between that quarter wave and a Retevis RHD-771.  Just for a test. 

[jdl]

I'd be very interested to see the dB differences between that quarter wave and a Retevis RHD-771.  Just for a test. 

I'll test this

Quad with monopole, 1.25mW on Revo (RFM22B) vs Retevis RHD-771 on 1.25mW Oplink receiver (RFM22B). Measured against another Oplink @ 1.25mW (coordinator). Preferably in open space, at reasonable distance.