The BIG and SMALL ANTENNA topic part 2


(Rustie0125) #25

Hi LoraTracker

Been keeping a keen eye on your research. Thank you for posting, As we know semtech IC’s has a max sensitivity of -145 more or less. the best/ worse signal i have received from the a node was -121 RSSI. now im not sure if this results is due to incorrect configuration or urban noise but i just dont want to miss out on the extra odd 20dB sensitivity that could result in extra range. So i guess what im asking is without breaking the law, what else could be done/tested to increase sensitivity.

It just seems odd that with all the work that has gone into antenna design exc that a simple GPA kick all the other antenna’s specifications without really costing anything.

I geuss one other option is to try and use a extreme high gain antenna for its receive gain but turn down the RF as to not break the law. the issue is the only reputable place i know where to buy a colinear 12db antenna is toaglas for $265. this is the only manufacturer that i have always found to be true to their word on spec.


(LoRaTracker) #26

As we know semtech IC’s has a max sensitivity of -145 more or less

Not quite true, the max sensitivity is -20dB below noise level. So if the receiver sees a noise level of -105dB (typical real world) max sensitivity is circa -125dB. Its not that difficult to demonstrate that the data sheet sensitivity figures do not apply to the real world, although you might be able to reproduce the data sheet figures in a lab in a Faraday shield.

The RSSI reading is thus of little help at the failure limit, look at the SNR instead. If your picking up signals (just about) at -20dB SNR, your at the limit, so forget what the RSSI reading or data sheet says.

However that is not really the issue.

There is a legal limit to the effective radiated power (ERP) its based on the amount of power into a dipole or the equvalent such as a 1/4 wave vertical with radials.

If your gateway is already transmitting the legal limit, which is very likely, then there is no point, if you want to stay legal, in fitting a gain antenna as you have to reduce the transmit power to nulify the antenna gain.

You can add gain antennas, as much as you like, to a node or gateway that is receiver only.


(Rustie0125) #28

@LoRaTracker I get what you saying. Iv just been changing RSSI improvement as it improves SNR aswell. In terms of 1/4 wave vertical with radials. Could I expect a better result by making a 1/2 wave version with radials ? Have you done any experiments in this regard ?


(LoRaTracker) #30

If your suggesting an end fed 1/2 wave length vertical antenna, then that should be a disaster as the impedance is in the 2000R+ range, not good for a 50R output transmitter.

Try a 3/4 wave length vertical that should have the same impedance\matching as a 1/4 wave.


(Tony Smith) #31

@Rustie0125, I have come to the exact same conclusion about all the antennas you listed. Without a ground plane in the form of 4 radials, the performance is compromised. I have recently been working on a Stacked Dipole Colinear. Here is an example of the type I’m using http://e.pavlin.si/2018/04/03/colinear-antenna-for-868mhz-lorawan-gateway/. Easy to tune, just trim the length of the top section. I’ve found I take off sections about 1mm at a time and quickly get it tuned spot on the design frequency.


(Tony Smith) #32

Hi @Rustie0125, interesting photo of a 1/4 wave GP. What is the purpose of the inductors in the ground plane radials? I would assume they are useful on a dual band antenna but don’t understand their purpose on a LoraWan system where you have a set of adjacent channels over a reasonably narrow bandwidth.


(Rustie0125) #33

Hi Tony. I have built 2 colinears of the same type you built and did not achieve much over the 1/4 vertical, what is your findings ?


(Rustie0125) #34

I had to lengthen the ground radials to improve the impedance matching range but wanted the device to fit in 75mm pvc pipe. So adding loaded ground radials got me better tunning range as I could cut them shorter or pull on them to extend them. I found that if I take straight radials of 86mm and cut them shorter to fit in pipe I could not get impedance correct( as you would expect , as the are to short.) the coiled radials are 55mm in length over all but made from 110mm of wire before coiling. I still want to coil the actual vertical to bring the height down but I am/was more concerned about diameter rather then height. Once I found the best compromise between size and performance I will probably build a few to sell to get feedback from others
The next step is to replace N type with something smaller maybe custom PCB and move the connector to outside of enclosure without adding cost. Lol the most expensive part about this antenna is the N type connector.


(Tony Smith) #35

Thanks, that’s really interesting. I too am working on antennas designed to fit a small diameter waterproof enclosure. I started at 75mm PVC tube, reduced that to 65mm and now aiming at 50mm as I can use 2" BSP or NPT threaded fittings to connect to a 7 metre length of 1" or 1 1/4" steel support pipe. I have a full Raspi based gateway running in a 65mm tube but working on a design for a gateway to fit a 50mm PVC tube. I’m aiming to sit the antenna directly above the radio module and not be too heavy.


(Tony Smith) #36

I’ve yet to perform a proper range test. which I hope to do in the next few weeks.


(Rustie0125) #37

@TonySmith this sound awesome ! I have 10 gateways currently running on raspberry pi zero using the RAK gateway module but am designing my own raspberry pi zero hat with the new sx1280 ic aswell for 2.4ghz nodes. Please let us know your test results and photos of your gateways that will be really interesting.


(Rustie0125) #38

@TonySmith one thing I must note! All my antena’s Are always tested/ Tuned with the extended cable that I will be using for installation(all my antennas are tuned with 1 meter LMR400). People underestimate the effect of the cable on the antenna.


(Tony Smith) #39

31932673_10214557533950715_2716543445802942464_o Photo of self contained outdoor raspi gateway with a commercial 6dBi antenna inside the radome. This one requires lots of machining of plastic to connect the housing to support pipe. Gateway is about 2 metres above the peak of the house. Get good range in cluttered urban environment


(Rustie0125) #40

That looks awesome ! I take it your devices are PoE powered/communicated ?


(Tony Smith) #41

Yes, built my own POE power supply and its connected via ethernet cable. Can also power from Solar Panel/Battery and/or plug in 3G/4G dongle.


(Rustie0125) #44

Talking about the antenna’s however. Who has put effort into Node antenna’s? The best SMD chip antenna I found was w3013 from pulse. Got 3.3km out of this antenna. Other then that can’t beat 1/4 wave piece of wire for cost/performance


(Tony Smith) #45

Waiting for one of these Molex antennas to arrive. Hear they are cost effective and have good performance. ideal for a node. https://www.molex.com/molex/products/datasheet.jsp?part=active/1052620001_ANTENNAS.xml


(Rustie0125) #46

If they are anything like these https://www.taoglas.com/product/pc81-868mhz-ism-mini-pcb-antenna/ you will be happy! They work amazingly just expensive.


(Tony Smith) #47

Ooooo that’s nice!! Do like the cable exits at the end of the antenna not the centre as per Molex. This allows much slimmer packaging and can get the antenna correctly polarised above the Node’s electronics.


(Rustie0125) #48

Ya and they do match there specifications. I have ordered a few of these they claim 2.63 db gain which seems a bit high but will test and give results. Hope they work