Hello, folks. I just joined! This is a cross-post from my Reddit post ( https://www.reddit.com/r/rfelectronics/comments/16eeloa ). I am working on a design using LoRa for communication, but I have no real RF knowledge or expertise. So, I have a few questions that I hope you all can help me answer.

1. The device will use LoRa operating at 915 MHz. I can calculate that the wavelength is (1s / 915e6) * (2.98e8m / 1 s) = 0.3257 m. So, a quarter-wavelength monopole antenna should be 8.14 cm long. My question is, in practice, how precise does this really need to be? What would the effect be of the antenna instead having a length of 8.10 or 8.20 cm? 7 or 9 cm? I could never have my antennas be exactly one quarter-wavelength, so is it better to be a little longer than shorter? Do I need them to reliably be 8.14 cm or longer? I will have the antenna wire essentially soldered into a PCB via. Is the 8.14 cm the length of the wire above the PCB, or can it include the little bit sticking through the bottom of the via?

2. I believe that the monopole antenna should have omnidirectional operation. Meaning that it will be able to transmit and receive in any orientation(?). But, what if it is not straight but curved so that it can fit in the device's enclosure? What if it were bent, or a spiral?

3. In my early prototypes, the solid-core 22 AWG wire I have used has often broken off at the solder point when the antenna has been moved around. So, I want to try switching to stranded-core wire. Is one or the other better for antennas? How much does the gauge matter? Should it be higher or lower gauge?

I would greatly appreciate any answers to these questions. Thank you!

 

It is an element of RF engineering you will discover in time, but the single piece of wire soldered to a board will not perform very well. The monopole or vertical antenna needs another potential, or ground-plane, to work against. The actual radiation pattern in three dimensions depends on the geometric relationship between the driven element (monopole) and the ground plane. There is an open-source simulation program called EZNEC that will allow you to answer your questions by inputting a simple description of your antenna geometry. It will solve the field equations to provide a set of radiation patterns in both azimuth and elevation.

You may encounter "antenna gain" in units of dBi. This is nothing more than the gain in dB of a particular radiation pattern, in a particular direction, with respect to an isotropic radiator. Such an isotropic radiator would be a point source in the center of an imaginary sphere that radiates the same in all directions. This is an idealization because in practice we need to contend with the existence of the literal ground, aka the surface of the earth.

The ground reflects RF energy and the ionosphere is transparent to UHF and microwave frequencies, but it reflects HF and low VHF frequencies. 915 MHz. is in the UHF range which means the ionosphere is transparent to the propagation of such signals.

 

For a straight mono-pole antenna with a ground plane the resonance is fairly wide with solid wire (don't use stranded wire) so it's not super critical to fine-tune the length.
https://medium.com/home-wireless/testing-and-reviewing-lora-antennas-5b37dfa594a3
Helical antennas are OK for testing but if you want full range you will need something better.

Several types of DIY designs.

 

@nsaspook illustrates an important point here. The ideal ground plan would be a solid metal plate cut in a circular shape. A practical approximation of a ground plane consists of some number of radial wires, usually an even number, perpendicular to the base of the monopole. Allowing them to bend away from the plane of the base of the monopole has a number of interesting effects. See the above reference to EZNEC and similar software to investigate these effects,

 

Thank you for the replies so far! I will read and look into them more. But, I should probably mention: The antenna will be soldered into the the corner of my PCB next to the LoRa module and be in the same net as the ANT pin of the module. The board does have a ground plane in the same net as the LoRa module's GND pins. Maybe that helps?

 

(don't use stranded wire)

Could you elaborate on this a bit? The answer here disagrees.

 

Could you elaborate on this a bit? The answer here disagrees.

Skin effect.

https://en.wikipedia.org/wiki/Skin_effect

Quoted from the article:

Skin depth depends on the frequency of the alternating current; as frequency increases, current flow becomes more concentrated near the surface, resulting in less skin depth. Skin effect reduces the effective cross-section of the conductor and thus increases its effective resistance.​

​

Everything gets worse the higher you go in frequency. Hardline coaxial cable used in UHF and microwave applications as described by rfparts.com:

Hard line coaxial cables have metal tubing (aluminum, copper, silver, gold or combination) as a shield. The center conductor consists of solid copper, or copper-plated aluminum. Connections must be air and water tight.​

​

For any "short" length it would be hard to argue that there would be a measurable difference between solid and stranded. If however, your antenna is on top of a 1200' tower now we are talking some serious attenuation at 915 MHz.

 

Could you elaborate on this a bit? The answer here disagrees.

Stranded wire has a complex surface geometry at much higher frequencies (UHF/SHF regions) when those discontinuities (energy reflections that can alter the properties of a antenna like length) on the surface (due to skin effect) start to reach a fair fraction of the wavelength. At much lower frequencies it usually does't make much difference for a antenna.

 

About accuracy:
First, the antenna wire shortening factor nevver is very accurate and it depends on the wire diameter, wire insulation if any, frequency etc factors. Yet in the most cases it may stand 0.95 plus minus 0.01 but hermetizing antenna into plastic box, pipe, or other insulator, in worst case may opress it down to 0.66 as an extreme. So, this is the basic reason why just cutting the proper length of wire NEVER works like a good antenna. To play the antenna role this wire MUST be measured by VNA (options beyound the Rhode&Scwartz available for hobby budget are P-100, N-1201, Nano-VNA). Then at those frequency range the half of mm accuracy already plays the measurable role, but mistake of 1 mm is catastrophic.

 

RE:""it would be hard to argue that there would be a measurable difference between solid and stranded""

The Focault effect of course gives a ultimate impact at those frequencies. Just at the 1 GHz the tiniest 0.6 microns layer is conductor and all the other wire cross section stays unused. Bit ago I wrote my experience of 160 MHz antenna making where I take 70 years layed in shelf very nice duraluminium pipe of 10 mm diameter. Antenna VSWR yet was catastrophic, if my memory not fails about 6 or larger. Then I just cleaned off the white aluminium oxide powder from surface by 400 number of sandpaper. VSWR became about 2. Then I polished it by sandpaper number 2000 and got 1.05. Next morning yet it was 1.3. I polished it by anew and got back 1.05. Next morning the tale repeated. Then I polished again and put over the good acrylic lack. So the VSWR was stable and good.

Therefore there are two advices - apply the antenna wire perimeter large enough according the power what will be expected. Or if perimeter means unacceptable large diameter, then go toward litz wire construction, what means the sum of all litz wires perimeters must be those demanded.... however I am far not sure do the inward looking litz surfaces are radiating anything, when shaded by equipotential field of neighbour wires. Thus, probably the insulator cylinder with few thin wires be guessed having better efficiency. But practice is to use a proper wire diameter, e.g. 2 mm, or 3 mm or 4 mm.

As the winding wires normally comes in high quality thin PTFE lack insulation, just avoid to injure that insulation, and wire will not oxidize thus loosing 100% of conductivity and the consequence of oxidizing deeper than 0.6 microns.

Last but no least is the litzwire often is slight twisted. And that for the sure changes the velocity factor dramatically, however I am not sure do exist any formulas to calc the value of this effect - it must be measured by VNA.

 

RE:""The ideal ground plan would be a solid metal plate cut in a circular shape.""
Yepp, BUT the angle of it have a giant role to the antenna impedance, thus one may decline from 75 Ohm to the 50 Ohm painlessly.
Other, probably the metal ring isnt bad idea, but practically, cheaper, more lightweight and not a yotta less effective are three, fout, six etc number of wires simulating that plane of "ground".

 

It is an element of RF engineering you will discover in time, but the single piece of wire soldered to a board will not perform very well. The monopole or vertical antenna needs another potential, or ground-plane, to work against.

With handheld ham radios, one trick is using a rat tail. Crude, but it helps.

The 2nd biggest issue is gonna be VSWR between whatever amp output there is, and the wire "antenna". They almost always need to be tuned using VSWR meter.

They do make antenna's for 915M, I would opt for that before trying to make a omni fit in an enclosure and try to get an acceptable VSWR.
Sadly, even these pre-made ones are only around 40% efficient.
https://www.mouser.com/pdfDocs/ant-915-cpa-ds.pdf