Do I need 2 Yagi xohm antennas

[SKD]

Hi: Both my samsung 4G card and zytel modem has two SMA connectors for two antennas. These yagis need to be oriented at 90deg with each other and 45 deg from vertical. So my q is do I need to buy two yagi cylindrical assemblies for my xohm usb card/ zytel modem or two with correct orientation are already provided in each cylinder. If not so, are there markings on the cylinder showing how the yagi elements are oriented? Appreciate your response

[thuor]

Mimo antennas for wimax

Where did you get the info on orthogonal 90degree orientation and 45deg elevation?

WiMax utilizes MIMO (Multi Input Multi Output) radio transmission technology. So while it does help to have two antennas for maximum MIMO gain you really only need to get one antenna to achieve antenna diversity.

If you do get two directional Yagi antennas the most important parameter will need to be the antenna separation. Theoretically two antennas operating in the same frequency band only need to be separated by a ½ wavelength to avoid coupling (recall wavelength lambda = c/f (where c is the speed of light 3x10^8 and f is the frequency ~2.5x10^9 for Sprint 4G Wimax so lambda ~ 12cm). Practically need several wavelengths separation distance.

On the other hand however to maximize MIMO channel gain it's important to ensure that the paths between the transmit antennas (cellular towers) and receive antennas (radio device) or vice versa are as uncorrelated as possible. The non-intuitive result of this is that MIMO channels require that you have a non-line-of-sight path (multipath propagation actually improves MIMO gain!!!). Practically this means that it would be beneficial to actually orient the antennas in different directions to pick up signals with different angles of arrival or arriving from different directions.

Finally you will need to carefully plan your link budget to ensure that you don't inadvertently create an extra strong signal at the input of your radios. If the input signal is too strong it will saturate the receiver front end.

[SKD]

That s very good information. I did some drive testing for Sprint using 4G card with two roof top stub antennas connected to the two antenna connectors on the card. This made the thruput really good even at 8 miles from the tower in Chicago. Here the on-the-air tower close to my home is less than a mile away and without any antenna connected the 4G card and modem give around half thruput expected. Hence looking at different antenna to boost the thruput. As mentioned earlier the modem got two sma cable male connectors. When I looked at the cell / BTS radios in the lab they have an array of antenna placed in a vertical line, each point having two transmitters/receivers oriented at 90deg to each other and 45 deg from vertical. If the Yagi directional lobe is in the same plane as the elements, why should a lobe 90 deg from this plane should interfere even if both lobes operate at the same frequency? I could be wrong but just curious.

Appreciate your input. I may be needing about 20 ft of cabling with SMA female endings on the modem side. What else I need to do for impedance matching for the length of the cabling?

[thuor]

Orthogonality on the source antennas or antenna polarization does help to ensure that the transmit signals are as uncorrelated to each other as much as possible but due to multipath propagation this doesn’t really have much of an effect in the far field. Typically only significant for Line of Sight (LOS) Point-to-Point (PtP) systems. In this case we really don’t worry too much about far field interference since the two antennas are really working together (no co-channel interference in this case). Only thing to concern ourselves with is MIMO channel gain which is dependent on the rank of the channel matrix which in term is dependent on how uncorrelated the multiple spatial channels are. If your signal is strong anyway then this doesn’t matter much to begin with – i.e don’t depend too much on MIMO channel gain since your SNR is already decent enough for higher modulation types and hence higher throughput.

Don’t need anything for impedance matching and in fact you really want to reduce the amount of components to reduce insertion losses and minimize the length of the cable to reduce attenuation. The cables are 50ohms and the radios require a 50ohm input impedance so they are already matched. The VSWR or Return Loss is typically good enough for such application that don’t require very stringent impedance matching.

So just to clarify everything you will need:

1) you need 2 Yagi antennas in the Xohm/clearwire wimax frequency band (2.5 – 2.7GHz) more >>

2) 20ft low loss cable with SMA-female connector and a connector to mate to the Yagi antennas more >>

3) Do you need lightning surge protectors as well? more >>

[burnin240sx]

I see in your 2nd post that you said you can use just 1 antenna. how do I hook it up. do I connect it to 1 port or do I use a splitter and connect it both ports? I don't want to blow my radio.

[thuor]

Hooking up antenna to Sprint 4G, Clear, WiMax Modem

Thats a good question. First two antennas will always outperform one. If you do use one antenna you connect it to the primary antenna port and NOT to a 1:2 splitter/combiner. This would beat the purpose of MIMO. It wouldn't really hurt though but these is just adding unnecessary components and also adding unecessary losses (and splitters can add an insertion loss of up to 3dB - half power in addition to other losses such as non-perfect impedance matching etc).

Also these are passive antennas (don't need to connect them to a power source) so less risk of blowing anything in the analog front end of your radio. Only concern though is lightning so if mounting this outdoors in a lightning prone area and where no other lightning arrestors exists consider adding a lightning surge device such as: Connecting Zyxel Wimax modem with lightning surge device. Remember though that the lightning surge device adds an insertion loss so keep the cables as short as practical.

Under certain propagation conditions, specifically under low multipath propagation environments where a line of sight or near line of sight exists one antenna works just fine. This is because MIMO doesn't have much benefit if the paths between the basestation and the antennas are correlated and they would be more correlated if you have a line of sight and less correlated if you have multiple reflections (obstacles) in the transmission path.

[burnin240sx]

OK I'm going to start with the RooTenna 2 panel antenna

http://www.rfwel.com/shop/?target=product&product_id=1664&cat=179&page=1

I'm only .72 miles (straight line) but I'm only getting 2 bars -85 - -89db gain on my modem. the other reading is CRRI (iirc) it getting 3-7db max.

so my plan is to get up on the roof and hook up the antenna above and see if this improves my signal. If it does then I may order another one to take full effect of the external antenna option as I want the best connection possible.
all impedance is 50 ohms and I'm planning on putting the modem in the antenna enclosure if there is space to reduce the loss over the cable. I will definitely add a lighting protection.

Do you see any issue in trying this setup?

[thuor]

Looks good. Check with sales before you order that antenna i know they had it on backorder. I know the 16dBi panels are in stock though: http://www.rfwel.com/shop/?target=pr...roduct_id=1721

I know MD doesn't get that hot unlike Arizona where it's hit 115 degF once this week but keep in mind the Zyxel modem is spec'd for an operating temperature of 45 degC (~113 degF).