improving 433 MHz range

[texane]

Hi,

I recently got this module:
http://blog.zakkemble.co.uk/wp-content/uploads/2012/12/modules.jpg
http://www.electrodragon.com/w/NRF905_Transceiver_433MHz-Wireless_Module

I want to use it to transmit data between my flat and my garage, which
is 7 floors below. Thus, this is a vertical configuration. For testing
purpose, I tried this:
. a module sends packets from the garage. The power gain is set to max
(10 db, adjustable in software). The packet size is limited to 6 bytes
CRC and automatic retransmission are disabled.
. a receiving module is located in the flat. 3 different messages are
printed if the carrier is detected, data are received, and all the data
bytes are correct.

In my flat, the carrier is not detected. 2 floors below, the carrier is
detected. 3 floors below, data are received and all correct. Thus,
communication between the flat and the garage seems possible to me, but
the range needs to be improved a bit. To do so, I plan to try different
antenna designs. For instance, I found this doc, where page 14 describes
the bulding of a ground plane antenna:
http://www.arrl.org/files/file/Get%20Licensed/Quickstart%20Guide%20EBOOK.pdf

Do you have any advice about antenna designs that are well suited for
this kind of vertical transmission configuration ? What are the other
points I could improve ?

Thank you for helping!

[Psi]

1. Make sure both antennas (RX and TX) are positioned in the same orientation. If one is at 90deg from the other the range will be bad.
Best range is when they're both in the same orientation.
2. Linear antennas have a null zone out the top and bottom of the antenna. So if your TX and RX are right above/below each other its best to have both antennas horizontal (That way you're not trying to communicate through the null zone).

If it's still not working try some better antennas, those short stubby 433mhz antennas aren't very good.
You can build good 433mhz dipole antennas really easy.

Get a SMA to SMA cable (assuming that's the connector on your tx/rx) and cut the coax in the middle (so you can make 2 antennas) then you can solder some up.

433mhz is used a lot in the FPV community so there's lots of info on making them.

Basically you want the coax coming out of the TX/RX and then solder two wires going off at 90deg and -90deg (to the coax). For 433mhz each wire should be 164.5mm long (http://www.csgnetwork.com/antennaedcalc.html)
One wire solders to the coax shield, the other to the main coax conductor

|
|
164.5mm
|
|
---Coax----<transmitter or receiver>
|
|
164.5mm
|
|

[texane]

Thanks for your reply. The connector is actually SMA. Is there advantage
of using dipole antenna in this kind of vertical transmission configuration ?

[Psi]

A dipole is better than a monopole,  i dunno what is inside your current small antenna, could be either but probably monopole.

I would try a horizontal dipole on the TX and RX and see what you get, it should give noticeably better range.

[hagster]

That ARRL antenna should work. You could add more 'ground' wires to make a bit of a cone. This would probably improve the pattern.

PSI's dipole would likely also work. It doesn't include a balun like the ARRL does.

The other good simple design is the sleeve dipole. This is what most WiFi antennas use. You can Google how to build it or just buy one ready made. http://uk.mouser.com/Search/m_ProductDetail.aspx?qs=K5ta8V%252bWhtZ00LNUmbKZ4g==

[hagster]

Of remember a dipole and monopole radiation patterns have a null end on. You might want to have the antenn horizontal to direct the pattern vertically. You could also place a reflector 1/4 wave from the antenna.

[Psi]

With those stubby 433mhz antennas the entire antenna is located really close to the RX electronics which likely increases the noise floor.
A small sma extension cable to get some separation may help.

Do you know what the RF output power is?

[texane]

Thanks for your answers, they are of great help.

@Psi:
I do not know the radiated power, but it is less than 90mW. I would
say 80mW. The NRF905 states that the total current consumed by
the module when in TX mode is related to the software adjustable gain
as follow:
-10 dBm: 27mW
-2 dBm: 42mW
6 dBm: 60mW
10 dBm: 90mW

My software sets the gain to 10 dBm.

[Psi]

With good antennas 80mW is plenty of power to go through a 7 story building at 433mhz

I think your issue is just with the antennas.

[NiHaoMike]

Just make a pair of directional antennas.

[hagster]

Thanks for your answers, they are of great help.

@Psi:
I do not know the radiated power, but it is less than 90mW. I would
say 80mW. The NRF905 states that the total current consumed by
the module when in TX mode is related to the software adjustable gain
as follow:
-10 dBm: 27mW
-2 dBm: 42mW
6 dBm: 60mW
10 dBm: 90mW

My software sets the gain to 10 dBm.

10dBm(10mW) is the output power. The 90mW is the power consumption of the module. You can get more sensitivity with a slower data rate if your chip supports this.

[texane]

Hi,

I purchased this antenna which corresponds to the model described by Psi if
I am not mistaking:
http://fr.mouser.com/ProductDetail/Linx-Technologies/ANT-433-MHW-SMA-L/?qs=%2fha2pyFaduizHDih74KwlXlZ4v2OdJFHHd%2f4HAnuCTScckk3zjpZ5KZM4%2frKQLyy

I now receive the signal in some precise positions in the flat. This is still to weak:
the carrier is detected, but data rarely arrive intact (but they do sometimes, confirming
that the carrier is not the one of some random device nearby).

I will try to build an antenna myself, as described by Psi ... but I do not believe it
will improve much.

Do you have any other idea ?

Thanks for your help !

[BravoV]

Too bad its way too late, last year TI had a great discount on its spectrum analyzer msp-sa430 that covers 389 – 464 MHz, official price is $250 and it was only $25 with free shipping, it was shopping frenzy -> here

[Psi]

I will try to build an antenna myself, as described by Psi ... but I do not believe it
will improve much.

Antennas are everything.

Since you only have 10mW you could investigate high gain directional antennas but i'd make up a quick full wave dipole at each end and try that out first. Should only take an hour to solder up two of them and test them out.

Full wave is 346.5mm of wire each side.
And remember to have it horizontal on both transceivers.

(Check if your tx/rx is actually working on 433.00mhz exactly, if its not use the calculator above to work out the exact length for your frequency)

[hagster]

Don't go for a Full Wave Dipole (unless it is a folded dipole). It will give you an awful impedance match and your antenna wont radiate.

I did some quick simulations to demonstrate this (50ohm source). The reflection coefficient comes out as -0.64dB. What this means is that if you send 10dBm of power down the cable to the antenna 9.36dBm of it will reflect right back at you.

The LINX antenna you bought looks pretty good with an SWR of about <2. The equates to a reflection coefficient of about -10dB. Meaning that for the 10dBm you transmit only 0dBm(1mW) will reflect back into the the transmitter.

If you need more gain than this I recommend going for a corner reflector (This should actually improve your SWR slightly).  Do a quick google and you will see lots of examples, but basically the reflector needs to be 1/4wave from the antenna(the same length as each element). For more gain than that a Yagi is a good option.