Hi David
Well for 100M I make the free space path loss -72db. TX power is +20dbm so a RX signal strength of -52dbm or 0.56mV, in RF terms a pretty strong signal.
Of course this assumes perfect conditions. Zero cable loss from TX/RX to the antennas. Perfect unity gain antennas which would be a vertical dipole with a “proper” ground plane. Zero obstructions or anything like fresnel zones which could disrupt a clear signal path.
In the real world unfortunately this will not exist. Even in a dedicated propagation range the physical properties will be accurately known and can be allowed for or avoided.
I think the bogey man in this situation would be the PCB antennas. They would be far from ideal and I would not even hazard a guess exactly what their properties would be. Even how the DUT (device under test) was held and orientated would make a difference. All this would make any performance projections very difficult if not impossible.
Full marks to Core for even trying. They did all they could to put a practical figure on this performance. Just by walking away and finding a distance where the last seemingly reliable result was achieved. That appeared to be 100M. At least that could be regarded as a yardstick but with RF propagation there are so many variable factors even this would be open to individual experiment. Anything of this nature can be considered with a carefully surveyed path but in the home constructor/experimenter environment this is unlikely to happen.
In short, after all this preamble, I think your situation will be a “suck it and see” solution.
Set up the TX/RX in its final resting place with full power. Reduce TX power until signal us not useable then raise it at least 6db to allow for unseen degradation and go from there.
Cheers Bob
Not gonna lie - this was an intimidating project!
Our original version 0 design actually had an SMA connector on board. We did some soul-searching and considered the types of users that PiicoDev would attract and decided to go for the fully on-board route. While it would certainly sacrifice range, the user experience of having a device in a standard PiicoDev outline that requires no additional hardware (antenna) to get started with couldn’t be ignored.
As is so often the case with RF projects You can see why manufacturers quote the absolute best range they can - they want to sell product! We attempted here to get a pretty honest, practical range with an empirical demonstration in what seems to be a reasonable use case. Of course, your mileage may vary.
Yep. The RFM69HCW spec sheet you linked states sensitivity down to -120dBm @ 1.2kbps, but I have no way of accounting for all of those other various losses that you mention. So I couldn’t agree more.
Cheers, Dave
That is 0.2µV which is not a bad RX at all. Pretty good in fact.
Over the years I have installed and commissioned many point to point systems, from 500MHz to 7GHz. Provided the path profiles are properly prepared and accurate, The antenna gains and cable losses are known the expected RX signal levels can be predicted with a good degree of accuracy. There is the odd exception of course and I think in one case a fairly large frequency change had to be the solution.
A couple of 900MHz systems with the antennas being high gain Yagis. One system required a “passive” repeater to get over the escarpment on the NSW South Coast. This was 2 antennas each one being a stack of 4 of these Yagis. This later became “active” for other reasons too long to go into here.
The largest one was a parabolic grid type “dish” which was 7 metres wide and 6 metres high in a Tropposcatter system from Groote Eylandt to Gove NT. 193kM with a TX power of only 10W, not bad. 2 “dishes” at each end in a space / frequency diversity system. These dishes had a measured gain of 32db (design gain of 30db) which resulted in the 10W input being an Effective Radiated Power (ERP) of over 10kW. Worked well.
I think you did just that. Even outlined the simple method involved. Full marks as I think I agreed in my post.
You did not have a super good environment. The buildings etc along the street would have been providing lots of “freznell zones” with unwanted reflections cancelling and adding along the route. In fact you may have got to a point where the signal was almost completely cancelled and had you kept going it would have been highly likely (if this actually happened) that the wanted signal could return in all its glory. Strange stuff this RF.
Cheers Bob
You can see why manufacturers quote the absolute best range they can - they want to sell product! We attempted here to get a pretty honest, practical range with an empirical demonstration in what seems to be a reasonable use case. Of course, your mileage may vary.
