RFM69 antennas comparison

[zboblamont]

@gohan Fair comment, as always it is a trade off...

[sundberg84]

@zboblamont @gohan (or anyone!) - could i ask your opinion about an PCB trace antenna (like a normal Nrf24l01+) would benifit in different orientations - and which one is the best?

For example, most of my MySensors nodes are laying down flat (and therefore the antenna is pretty much going through the horizontal plane). Would it benifit to have the radio standing so the PCB trace is in the vertical plane instead?

[scalz]

@sundberg84
for "best" pcb antennas, I would say : monopole (meandered or not), and inverted F. they also need tuning of course, regarding board geometry, antenna design etc.
but a well tuned whip antenna (with a wire) will always radiate in a better way, sort of 3d antenna compared to a 2d pcb antenna.

[sundberg84]

@scalz - ok, but if you take the Nrf24l01+ radio with its PCB antenna - is the antenna performing worse if the radio is lying down compared to if I have the radio standing up?

[zboblamont]

@sundberg84 Not a great fan of PCB trace antennae, although I fully understand the advantage of small footprint. A whip or dipole antenna is problematic in terms of incorporating neatly, but will outperform the PCB a large factor.
Horizontal polarisation when received by a vertical cuts the gain considerably, at vhf/uhf etc up to 20dB, not sure the level at 2.4GHz as the wavelength is fairly short, but there will certainly be some attenuation on the outer fringes.
The other problem with horizontal polarisation is that the radiation lobe tends to be directional, so if nodes are at range you may have to compromise direction to 'catch' the edge of the radiation lobe.

[sundberg84]

@zboblamont - must admit I have a little hard time to fully follow your explanation and the terminology - but if I understand you correctly the answer is, yes I might have some loss if the radio is lying down.

[zboblamont]

@sundberg84 Sorry, yes there will be a loss.
This site may better demonstrate it graphically down the bottom...
link text

[sundberg84]

@zboblamont - perfect!! Thanks!
Edit: from your link I could google more and found this: http://www.edaboard.com/thread294139.html

Because meander type is a monopole, the polarization is almost the same as a standard vertical monopole antenna (or the duck antenna in your setup).
So, the best coupling between antennas (giving the best communications between modules) is to place both antennas in vertical position (keeping the module using the printed antenna in the position shown in the picture).

Edit2: https://forum.mysensors.org/topic/3459/nrf24l01-align-direction-position
https://forum.mysensors.org/topic/212/antenna-101

[gohan]

I usually try to put the antennae in vertical alignment since the communication direction is mainly horizontal

[zboblamont]

@sundberg84 The way to imagine field strength on an antenna is that it is circular looking on the end of a dipole, varying in strength from essentially zero at one end to max at the centre back to zero at the other end.
Under ideal conditions, the field is like the balloon explanation in the other link. Unless your receiver is hugely different in height to the sending antenna, the reception will be identical throughout 360 degrees at the same distance.
Rotate the antenna 90 degrees however and the the field forms principal lobes to front and rear, max facing centre on both sides, reducing to minimum facing the ends, viz this which makes the antenna directional compared to it's vertical orientation.

[Yveaux]

@gohan I ran into this HopeRF antenna paper. Addresses a lot of different antenna types.

[zboblamont]

@Yveaux Excellent summary, yet well worth reading through.

What I found fascinating elsewhere was where 2.4GHz compressed PCB antennae such as the zig-zag pattern underperformed compared to their U or J type countertparts, despite identical claims as to gain, which is bidirectional.

[gohan]

It is more or less related, but just this evening I had the garden sensor going out of range for 2 hours (from 6 to 8 pm). It happened a few time in the past but for shorter periods. Looking at the TX power it seems there is something affecting the communication from time to time. It would be nice to know the source since I live in the countryside and I have only another house near mine... I have to make other outdoor sensors to see if I get the same pattern

[gohan]

Added in OP the first test of the SMA antenna

[Tmaster]

@gohan
Hi. i know that its 3 months + old discussion .but i only read now :)
i have the same problem,probably same time that hapens to you. my nrf24 modules between front gate and gateway stops comunicating at end of day.
My bet it that signal its already on limit ,(at 50m between gate and home), and at the end of day ,air humidity increase and signal just don't arrive. I already change antennas but signal strengh it's on limit and some days are worst than others.

Now i will change all my nrf24 for rfm69's (they arrive today) and try again at 868mhz and simple wire antenna(dipole).
Just for confirm; are the grownd plane from rfm69hw good enought? or i need something more than a piece of wire on ANA pin ?

[zboblamont]

@tmaster Pending @gohan's response, I suggest you try a simple 1/4 plain wire whip initially.
You are using not simply different devices with differing characteristics but different frequencies, penetration of 2.4GHz and 868MHz wavelengths are not the same, lower frequencies have better penetration.
Should you require additional gain either to reduce transmission power input, or to increase effective radiated power it is easy enough to add a mirror 1/4 later to make it a dipole should you desire...

[gohan]

You could also use a PA LNA radio on the gateway and maybe on node. Also trying different antennae on the radio, as previously suggested, is an option. RFM69 have a good advantage of reportig signal strength (I published a small signal scanner for RFM69 in the My Project section) that gets handy in detecting blind spots.

[rozpruwacz]

@tmaster I had similar problems with nrf24 and also decided to switch to rfm69. I used nrf2rfm69 board to replace nrf24 modules. As an antenna I use a simple 1/4 wavelength flexible wire with isolation. And it works far more better than nrf24 even at 0dBm tx power. The modules are only twice the cost of nrf24 and perform much better. I guess the frequency (I use 868MHz) is the key factor here.

[Tmaster]

:+1: :ok_hand:
thanks

[zboblamont]

@gohan I meant to ask about that scanner project... Is this now fully realisable under v 2.2 ?