Here 500mW gives a range of 6km @ 868MHz as opposed to 11km on 459MHz (discounting terrain). 25mW will give one a whopping 1.4km. What also starts to play a huge part is natural obstructions. 868MHz bounces nicely and is also absorbed extremely well, the one characteristic it does not have is penetration! The main application that I have worked on has been water projects where sites are spread over a number of kilometres and through fairly dense foliage. I don't see 868MHz working here too well. Ireland, offering a lot of potential at the moment, I believe would not be a place I would like to try 868MHz at all. One other consideration would be that I can see people testing the equipment on bright (ok grey, but at least dry!) days. Tomorrow it's gonna rain and the equipment falls on it's....... (rain stops high frequencies, badly!).
This is where I see most of the problems. There is a limited understand of the concept of 'half range coax length' i.e. the length of coax to reduce the range by half (induce a 6dB loss). At 459MHz this is 14.2 metres RG58 or 30m RG213. At 868MHz this reduces to 9.8m RG58 or 21m RG213 (this does not take into account any connector loss or antenna gains/losses). One cannot guarantee the typical length of coax on installations as they do vary dramatically.
A contentious issue in itself. Many people do not understand antenna principles and could prove to be a headache if offered on 25mW systems. It's all very well offering 16dB (to get to 500mW ERP) but then one is bringing the required energy into a narrow beam. This is fine for modem links but not for multi-station systems. I also have a problem with collinears as the radiation pattern is so flat local coverage suffers especially when used on mountain tops. What 868MHz does now offer is helical antennas (they're a bit bulky at 459MHz) which offer a fantastic natural immunity to interference (and never mind phasing differences from being bounced off aircraft and the like!).
If it came to choice I would offer nothing but 5/8ths for omni-directional applications at normal ground level, omni-directional stacked arrays for mountain tops, Yagis on difficult sites, and helicals on links. The price of a Yagi on 868MHz is far more than at 459MHz.
Taking all the above into consideration it leaves two aspects to consider. Is one going to spend the extra on the components on the board or rely (and I used this word carefully) on the abilities of an installer to ensure the best output. At 868MHz even BNC connectors start coming apart.
To the somewhat uninitiated 25mW is immediately thought of as "20 times less power". To us of course "it's only 13dB". The problem here is it's not always feasible to use gain Yagis especially in a multi-station environment. In this situation we're talking of then a 4:1 range difference.
My conclusion; Certainly offer 25mW as standard but have the space on the board to add the extra 13dB. From a technical point of view I would like to know 500mW is available to upgrade a site should the need arise. I see this as a far cheaper option than trying to add an external amplifier to the system when required.
One advantage of low power is the lowering of interference to neighbouring systems. Radio congestion is a real problem.
As the installation of this frequency is now of paramount importance i.e. keeping antenna cables to a minimum and antenna high and out of the way leads one to have to possibly install the equipment in out-of-the-way places and difficult to work on.
Here I would like to twist arms to offer a unit that is the 'radio portion'. A RS485 network then feeding a number of I/O modules. This does pose one problem in that the radio would be run on long power wires (of reasonable gauge) but will suffer dips as it transmits.
Just picture it, a 3-4 metre pole holding a small well graded box housing a radio with a 5/8ths antenna plugged straight in the top. No coax losses, the 5/8ths offering the last 1.8dB to bring one to 500mW ERP, and a simple stock standard 4 core cable as used for USB (two power cores and a twisted data pair) linking this to its I/O mates. No fancy RF connectors to go wrong, electrician friendly wiring (extendable, cropable, changable, u-name-it).
| | Ask a Question |