1km range at 868MHz? « Jee Labs Talk

Jee Labs Talk » Networking & wireless

1km range at 868MHz?

(10 posts) (6 voices)

Tags:

No tags yet.

divemasterp

I'm looking to build a one directional signalling system with a 1000m range and bi-directional reliability - i.e. either end can see if the link is lost. Only about three states need to be carried. The path is completely clear and unobstructed. The JeeNode would seem to be a sensible basis with a lightweight signalling protocol.

This is (I assume - I haven't bought the JeeNodes while I'm doing some other work) going to need more than a wire quarter-wave antenna. I don't want to spend £££ on a commercial antenna. Has anyone got relevant experience?

I'm wondering:

1. Does a television antenna for the UK's Group E have any gain left at 868MHz? It is supposed to cover channels 35-68 (586 - 850MHz) but I have always been a bit sceptical about wideband yagis.

2. Does anyone make a printed circuit antenna for this frequency? I've seen them for 2.4GHz but with a half wavelength around 170mm it ought to be possible.

3. Might a helical be the hacker's answer? The calculator at http://www.vk2zay.net/calculators/?body=helical.php suggests that a diameter of around 100mm should be OK and 3.5 turns would be 1.25m long and have a gain of 10dBi. That seems a plausible. I'll have to look for PVC pipe at the building store.

Happily I can mount the electronics at the antenna so there are no feeder losses to worry about.

I suppose 433MHz might have a more managable path loss but I worry about car keys etc.

Any thoughts?

Paul

Posted 04/03/2010 15:13:47
jcw

I'd be very interested to see how far we can push this with the RFM12B. The general info I have about range is: 100m line-of-sight, 20..30 m indoor. I haven't done any tests to find the actual range. Was thinking of hooking up my GPS to a JeeNode, and then use a central echoer. Then just walk away and let the GPS tell me how far I've gone.

I wish I knew more about antennas. Can't be of much help there, alas. My first (probably naïve) attempt would be a pair of home-made Yagi's.

To improve range, consider bringing down the baudrate of the RFM12B, and then reducing the bandwidth as well. Limiting bandwidth "focuses" more energy into the signal as far as I understand it - sorry for my laymen's way of saying this :)

Here's an example of that last point, taken to extremes: http://g4oep.atspace.com/qrss/qrss.htm

Posted 04/03/2010 16:48:18
andras

According to the RFM12B datasheet, the typical output power is 4 dBm and the typical receiver sensitivity is -105 dBm (1K2 bps, BW=134 KHz). The maximum allowable path loss is then 109 dBm. The Friis transmission equation states that the path loss over D=1000m is:

PL = 20log(4PiD/lambda) = 31+20log(D) = 91 dBm

That leaves 18dBm. As 10dBm should be taken as a safety margin, that leaves 8 dBm for other losses or uncertainty because your RFM12B has less output/sensitivity than the typical datasheet values. If you take the minimum instead of typical values, we have to substract another 7 dBm leaving only 1dBm for other losses.

It's still too close I think to do without a directional antenna. If you can get some antenna gain you should be able to make it (if we can believe the datasheet). Ofcourse you can also increase the receiver sensitivity with the methods JC sugests, but I don't know how much dBm's that would give.

Note also that just a line of sight isn't enough because of the Fresnel elipse; this is like a sigar-shaped space around your line of sight which should also be free from obstacles. At it's widest point in the middle of your link, the free radius around your line of sight must be 9,3m, so the antenna's must be at a high point. A penetration of this space outside 60% of that radius is not too dramatic, but within that 60%, losses will increase dramatically (so you should at least not get anything within a 6,6m radius at half the link distance). For 433MHz the sigar will be even wider (times sqrt(2)), so you may wanna stick with 868MHz because of this Fresnel elipse.

Posted 05/03/2010 21:47:02
jcw

Wow, I learned a lot from your post - thanks!

I'd like to run a simple test with Yagi's. Just to find out what simple means could accomplish. There must be tons of sites describing how to dimension the Yagi antenna, but I wonder how it should be hooked up - how does one go from the RFM12B antenna + ground pins to a Yagi without losing all that power gain again? I never understood all the impedance issues of RF ... :(

Posted 05/03/2010 23:30:04
divemasterp

Thanks Andras.

That's consistent with my gut feeling that you'd need about 15dB total antenna gain to make this work. I'd rather work a bit faster than 1200b/s but it would be a start.

Here http://www.tvaerials.com/product.aspx?productid=302 is a cheap-ish TV antenna that's available here in the UK. As I mentioned it is supposed to cover channels 35-68 (586 - 850MHz) so I'm wondering if cutting down the elements by a few mm would improve its performance at 868MHz. The element-element spacing would be a bit wrong but it would be much simpler than starting from scratch.

Here http://helix.remco.tk/ is a recipe for a 2.4GHz helix. It would be bigger at 868MHz but still manageable.

Also just found this design for 900MHz helicals http://home.earthlink.net/~w6rmk/antenna/helix900.htm

I'm thinking it will soon time to put a beacon at the end of the garden and try these ideas out.

Paul

Posted 08/03/2010 14:50:41
andras

I don't know much about antenna design; I usually just calculate with the specified gains from known antenna's.

HopeRF has a document on antenna design:

http://www.hoperf.com/upfile/ANTENNASMODULE.pdf

It gives a brief introduction on aspects of antenna design and gives some reference designs. These designs however trade form factor for performance (all perform less than the quarter-wave), so they are not very usefull for this use-case. Still an interesting read.

Posted 09/03/2010 23:02:46
tankslappa

I know this post is a few months old, but JC has just linked to it via his daily weblog, so I was wondering if you guys had had any success with antenna design.
I've found a nice DIY guide to making a yagi for 868Mhz (http://www.mictronics.de/2010/04/diy-yagi-for-868mhz/), but I really don't want directional if I can help it.

Posted 11/07/2010 07:53:18
jbeale

Antenna gain is directly proportional to directionality (forgive me if that is already obvious!). The "gain" of an antenna just comes from squeezing the power into a smaller viewing angle. You can get some useful gain with a multi-wavelength vertical; this antenna pattern looks like a pancake. More power goes out horizontally, less is radiated at high angles where it isn't useful anyway (unless you are talking to a balloon, or airplane :-). Of course, you need to be sure to keep the antenna vertical, no good for something that might be held or placed at any angle. Vertical antennas need either a ground plane, or an opposite element (dipole). Any other stray conductor near any part of the antenna will skew the pattern, probably in a way you don't want (although if the conductor is much smaller than 1/4 wavelength, it won't affect things too much.)
Note that a typical many element, log-periodic yagi TV antenna is designed to be very wideband, which causes design compromises. In particular, for any specific frequency channel, you're really using only about 3 elements out of however many there are in the whole antenna. The other elements resonate at other frequencies, which enables the antenna to work over the wide 2:1 ratio frequency range, but they are a useless waste of space if you use only a narrow frequency band.

Posted 25/10/2010 20:11:50
tankslappa

Hi Jbeale, Yes, sorry, my post didn't really make sense did it! What I intended to ask was for antenna tips to increase the range in the horizontal at the expense of the vertical.
So yes, the pancake or squashed doughnut (jam variety, not ring) distribution was what I was after.
Thanks for the information on the log period yagi TV antenna design, interesting stuff.

Posted 25/10/2010 21:28:39
ridezebigone

Hi,
some answers in this document : http://www.hoperf.com/upfile/ANTENNAS1212b.pdf
depending of your configuration you might not need a yagi antenna... what is sure is that high baudrate is not compatible with long range communication (high BER!). I'm electrical engineer, and used to work with hf transmission modules. by experience, I know that the lower baudrate, higher is reliability, higher is the range. I've reached 4km range with a yagi from a remote control and a dipole patch on the device (869Mhz, 500mW @ 2400bauds )
If results are 4km in a lab (page 22, with a high BER = 0.01 -> One bit over 100 is false, you might receive a lot of bad message!), you may hope to reach 1km in the real world ;-) But you will need to be very clean with the RF part (antenna, pcb routing ...)

Posted 31/10/2010 09:34:59