I have a routerboard 433 with a r52hn pci card and 2 omnidirectional anntenas. question. does 1 anntena pigtail go to chain 0 and the other anntenna got to chain 1 on the same r52hn? and if so do i need special settings in the software to optimize it as a AP? Any help would be great im still new at this and dont want to fry the board or pci card

Nope. You have to add another wifi interface on RB to attach second antenna.

Nope. You have to add another wifi interface on RB to attach second antenna.

you are wrong
He can put two omni antennas to one R52HN

So i can connect both annentas (which looks like boat annentas)? and if i can will that increase my throughput?

So i can connect both annentas (which looks like boat annentas)? and if i can will that increase my throughput?

That depends on many things.
First you need to enable the chain 1. It is disabled by default. Only enable any chain when antenna is attached, or you'll blow the radio amplifier...

If you are going to use 802.11n protocol it is best to have one antenna horizontal and the other one vertical. But due the nature of the omni's you'd be better of with a special dual polarized omni antenna.
Using the 'n' protocol on both ends (AP + Client) will increase throughput and also extend the reach of the signal a bit. But in a setup with two normal omni's and one horizontally placed the benefit might not be that much..

If you are going to use legacy a/b/g protocol both antenna's serve to create something that is called "antenna diversity". It might create better signal (thus possibly better throughput) in difficult spectral conditions.

Be aware of interferences of other AP's. Strong signals is only part of the story....

Thank you for your replys. currently one antenna is connected per pci card. when a heavy load is put over wireless, the data traffic becomes strained and we start losing packets. im going to try attaching both antennas to strengthen the throughput. now i guess its time for me to figure out how to enable chain 1. i'm guessing chain 0 is enabled by default then? There are lots of normal wireless routers and ap's in the area detected by the scan. they maybe to far away though to cause any problems (i hope)

so i believe i set ht-tx/rx-chains to 0 and 1 to enable both chain 0 and 1 . what if i set tx to 0 and rx to 1. would that be a better through put than putting both to 0,1?

Nope. You have to add another wifi interface on RB to attach second antenna.

you are wrong
He can put two omni antennas to one R52HN

Sure he can connect them physically, but that would be just a waste. that was meant to be used with dual polarization antennas. Two same polarized omnis would just cause interference.

Nope. You have to add another wifi interface on RB to attach second antenna.

you are wrong
He can put two omni antennas to one R52HN

Sure he can connect them physically, but that would be just a waste. that was meant to be used with dual polarization antennas. Two same polarized omnis would just cause interference.

so all the ruters with 2 and 3 antennas are just for fun?
and strange, have the same polarity
You should stop misleading people...

Two same polarized omnis would just cause interference.

Wrong, that is called antenna diversity, a technology used by many cellular phone operators. Two antenna's can just pick up better signal on one of them then only one can. But it can only be done by radio's with 2 or more signals outputs that are therefore in full synch. With two radios in the same frequencies the interference will kill the link. (Or they have to be synch'd too!)

I use two chains with 802.11a with two sectors attached. This way I can cover a wide sector with very good antenna and still use only one frequencies/radio.

Diversity is meant to be used in narrow closed space not in open field. In open, it is just call for trouble.

You cannot just attach two antennas to single device. You have to match them properly (use exact the same antenna, use exact the same cable lenghts measured properly and space antennas properly).

If you need to use more antennas to cover space, then use more wireless adapters too.

Diversity is meant to be used in narrow closed space not in open field. In open, it is just call for trouble.

You cannot just attach two antennas to single device. You have to match them properly (use exact the same antenna, use exact the same cable lenghts measured properly and space antennas properly).

If you need to use more antennas to cover space, then use more wireless adapters too.

I don't know where you get your knowledge from. You are wrong. Mobile operators use it in the open field like I do.
And with a simple splitter you basically already do what you say is not possible?

I have one system setup with on one chain an omni antenna and on the other a sector. Cable lengths are different.
It works to my full satisfaction. Omni is some 4 meter higher and picks some client that the sector won't see and the sector reaches distant client in a heavy populated area that the omni delivers/receives too little signal too/off. (The difference is 6dB!). All with one card and both chains in 802.11a connected.
Before I needed two radios to do the same job. Now I saved one frequency channel and some power consumption (tower runs on solar energy) and less heat build-up in the box (this one card with both chains enabled produces less heat than two cards doing the same! Here in Spain heat production of cards in remote solar/battery powered boxes is something you need to take in account when designing the system.)

I have two other towers where I use same system but with 2 sector antennas roughly 1 meter separated from each other in such angle that both sectors slightly overlap each other. Cables are same length.
One tower has the rb5Hn card and the other tower has an UBNT SR-71-15 card that delivers 24dBm at each chain! (Before I had 2 R52H cards on each sector. By replacing them with the SR-71 I still have the same signal at the client but only one frequency in use!)

In other words: Antenna diversity with the use of both chains in 802.11a mode is a success!

(Why not use these duo chain cards for mimo? Well, than I need to upgrade all existing clients to mimo capable cards as well plus I need to again work with 2 different cards+frequencies again..... Presently my .11a system outperforms my competitors .11n system!)

When both chains are used the ROS uses only that signal of station that is strongest.

I don't know where you get your knowledge from. You are wrong. Mobile operators use it in the open field like I do.
And with a simple splitter you basically already do what you say is not possible?

If you want to play that way, where did you learn to read? I never said it is not possible.

Diversity used in wifi cards is targeted to handle multipath interference. That means, receiver uses both antennas to be able to choose better received signal and then responds using one antena - one through signal is received. That works fine to handle interference caused by reflections seen in closed space or short range urban areas.

If you use that to make long distance connections with different client you are going to have problems, as while signal is transmitter through one antenna, clients using other antenna might not be able to hear it. If they do not hear it, they would try to transmit and cause interference.

It might work with few clients, where time share could be sufficient, bit if there are lots of them, it becames problem.

If you want to play that way, where did you learn to read? I never said it is not possible.

"You cannot just attach two antennas to single device." reads to me as it is not possible....

Diversity used in wifi cards is targeted to handle multipath interference. That means, receiver uses both antennas to be able to choose better received signal and then responds using one antena - one through signal is received. That works fine to handle interference caused by reflections seen in closed space or short range urban areas.

If you use that to make long distance connections with different client you are going to have problems, as while signal is transmitter through one antenna, clients using other antenna might not be able to hear it. If they do not hear it, they would try to transmit and cause interference.
It might work with few clients, where time share could be sufficient, bit if there are lots of them, it becames problem.

Well, 1 tower has 30+ CPE's connected and each is at least able to run 5M/1M down/up and it works to my satisfaction. Like you stated yourself, the receiver uses the antenna with the strongest signal to communicate with a particular client. And that is exactly what I want it to do and it works fine...
I must say that I use NV2 (TDMA) therefore your 'interference problem by not hearing one antenna by the client' is evaded this way. What you prescribe is sort of same as the 'hidden node' issue normal 802.11 networks are victim off.

I expected you to pull out NV2.

Well, that is proprietary protocol, developed and used by single manufacturer. I am talking about accepted standards.

If you use special case as an argument, then you should point out that is special case. What you did is irresponsible, creates confusion, and leads people to troublesome configurations.

The problem i was running into while connected with just one antenna hook up was the wireless ip cameras while being viewed are eating up all the bandwidth. i have tried 1 antenna hook up to each pci r52hn card (using 2 antennas in total). tried using just 1 card with chain 0 and 1 activated. tried chain 0 for tx and chain 1 for rx. im still having the issue. now i have no idea what the problem is.

I expected you to pull out NV2.

Well, that is proprietary protocol, developed and used by single manufacturer. I am talking about accepted standards.

If you use special case as an argument, then you should point out that is special case. What you did is irresponsible, creates confusion, and leads people to troublesome configurations.

I see in your status your are the master. I rest my case....

Right, when you have no arguments then go personal

My account at this forum has nothing to do with my experience and knowledge with Mikrotik and, especially, wireless. I am familiar with diversity issus since Mikrotik did not even exist.

leave us please
make your own topic about your ideas and maybe a pool.

Inox you could do better than giving me bad carma with comment that I am stupid.
It says more about you than about me

WirelessRudy, so, your expertise lays on personal and insulting comments and giving bad carma? I expected better from expert like you.

all of you better stay on topic. And it says - Connecting 2 omnidirectional anntenas to r52hn

if required you always can go out and test the solution. And see what are the benefits or drawbacks of the set-up.

Inox you could do better than giving me bad carma with comment that I am stupid.
It says more about you than about me

WirelessRudy, so, your expertise lays on personal and insulting comments and giving bad carma? I expected better from expert like you.

so if I'm an expert why you decrease my karma? Why you make this topic your own one? And now lets see one of the meanings:
stupid [stoo-pid, styoo‐] annoying or irritating, characterized by or proceeding from mental dullness.
Can't you see you are against everyone here?

so if I'm an expert why you decrease my karma? Why you make this topic your own one? And now lets see one of the meanings:
stupid [stoo-pid, styoo‐] annoying or irritating, characterized by or proceeding from mental dullness.
Can't you see you are against everyone here?

As I said in carma comment, I gave it because of your rude behaviour which I find totally unnaceptable. You revenged by giving me bad carma and calling me stupid. As I said it talks about you, not me.

I am not against noone. I spoke about general rule (and I gave explanations for my point of view, unlike you). You spoke about specific case. When that was cleared up, I saw no more disagreement. For some reason you still do.

It is much better to make things clear and explained in detail to be better understood by others who need to learn from that.

I have several AP's running multiple omni antenna. It works *very well* with vertical separation. It doesn't appear to work as well with horizontal separation, tho I only have one AP like that and it has some quirks of its own.

It isn't strictly diversity in this situation. Diversity would choose the antenna with the strongest signal. This is 802.11n working as designed and using both chains to reconstruct the signal. If it only chose one antenna then why would we see 6dB gains vs a single chain?

My best N AP has a 12dBi omni and a 9dBi omni, both mounted directly to the aluminum housing, 12 on top, 9 on the bottom. If I disable the 9dBi antenna I lose 4-6dB signal.

All clients have clear line of site to this AP, signals from -50 to -65, 65mbps, 90+ % CCQ.

The AP where the antennas are the same height, about 4 ft apart, performance is well above the signal or CCQ. I also have to disable and enable the cards after each change, so there is something wrong with this board.

My best N AP has a 12dBi omni and a 9dBi omni, both mounted directly to the aluminum housing, 12 on top, 9 on the bottom. If I disable the 9dBi antenna I lose 4-6dB signal.

Hmmm,

Every 3dB means doubling (or halving) the signal.

2 x 12dBi antenna would give double the signal of only one, so 15dBi
1 x 12dBi + 1 x 9dBi (which is thus half the power of the 12dBi) give only 50% more energy, or 12 + 1,5 = 13,5dBi
So, when both antennas work you would see 13,5dBi as energy coming from both antennas.

If you now switch that 9dBi antenna off again you basically only loose 50% of the total energy, which is 1,5dBi. So 13,5dBi minus 1,5 means 12dBi again. That makes sense, just the 12dBi of antenna 1.

The fact you loose actually much more is probably because the receive path, or any other issue in the radio-to-antenna-to-client unit chain has issues leading to that the 9dBi antenna is getting much less energy delivered to the client than what could be expected due pure calculations only.
I also believe that 802.11n is more depending on 2 same strenght antenna's with only some cm's spreading if both have the same polarisation. If both have different polarisations than distance is of lesser impact. The wavelenght of wifi is only some cm's so separation of antenna's only some cm's apart is already sufficiant to create enough multipath differentiation to work with 'n' protocol. (Look at all these home 'n' wifi routers. 2 or 3 antenna's only separated some cm's).
The reason you loose much more dB's is probably to do with some other factors. I would check the radio-connector-cable-connector-antenna-client chain to see where something can have gone wrong. Better, I woud use same gain antenna but in different polorisation.

This happens in my AP's with vertical and horizontal separation. If I disable chain 1, signal drops anywhere from 4dB to 6dB, even when scanning for other networks from the AP.

I think the baseline is 3dB gain from dual chain, sometimes more, sometimes less. My AP with horizontal separation is the one that is kind of flakey, and if I turn off chain 1 then I lose close to a third of the clients. I think the tower may be in the way in this situation.

The vertical separation AP, the clients literally can see the tower.

Looking at it tonight, the one with the most change is closest to the tower, so this may not be the best example. The reasoning behind a 12dBi on top and a 9dBi on bottom was to help with blind spots beneath the tower. Looks like I did what I had planned. The difference is more pronounced when scanning and the radio picking up my other towers with weak signals 20 miles away. The top three are clients, the bottom is the 5GHz feed. It is also a strange night as far as weather is concerned. Foggy, rain is coming in and unseasonably warm.

I also don't know if the 3dB gain from dual chain is completely dependent on identical antennas. I mean, that makes perfect sense, but in practice it hasn't seemed to be quite so simple. I think there will likely be 3dB gain on the receive side simply because there is another antenna, another source for the 802.11n signal reconstruction to work with.

It was an experiment that worked out. I am putting up another with vertical separation in the next week or so. I don't think I'll be trying the horizontal separation again until I get 3x3 with beamforming.

edit: the pic is the same AP, snapshots taken with 2 chains and with one disabled. The top 3 are clients, the bottom is an 802.11n card connected to an older tower with 802.11a.

Hi ppl,
I think everybody talks about something else.

- you could phase antennas on one single RF port. Than you need to match them so you get your desired directional footprint and your gain. It kind of works even if not done properly, or you just could get lucky.

- you put 2 antennas on 2 ports of a regular (b/g) system, then the 2 antennas are switched by a RF switch on the antenna line (that small 6 pin IC on the output) and the best antenna will be used by trial and error. NO interfecrence is possible because of CDMA access (AP doesn't send on busy RF channel) and because ONLY 1 ANTENNA is connected to the transceiver at any point in time.

In both situations, the throughbut for the same link speeds is not increased, just the error rate can be brought down.
(but in better condition, a higher bit rate can be negociated).

- in 2 chain n systems, both antennas work in the same time, depending on the conditions even with different data contents. On this i have no idea how the algo picks the best sollution. But here you have true multipath correction and, under ideal conditions (2 good separated H/V antenna) you can have effective double the data rate.

So don't mix these 3 situations. Each can be a sollution for a specific situation. And the "it won't work" issue is an incorrect statement.

Greetings, Marius.

- you put 2 antennas on 2 ports of a regular (b/g) system, then the 2 antennas are switched by a RF switch on the antenna line (that small 6 pin IC on the output) and the best antenna will be used by trial and error. NO interfecrence is possible because of CDMA access (AP doesn't send on busy RF channel) and because ONLY 1 ANTENNA is connected to the transceiver at any point in time.

This is regular diversity and is not supported by Mikrotik.

so how is it using both omnis? i suppose putting them in L shape will not make difference to interference?

On a n-capable device this makes sense and will improve your link.
But take care that a horizontal placed omni antenna will not be omnidirectional any more, but prefers a direction perpendicular on the whip.

But there are vertical dual linear polarized omni antennas on the market (e.g. by ubiquity) which allow you to connect both chains to one antenna (they have 2 antenna connectors on the bottom...) and enjoy 2 polarity multipath.

- you put 2 antennas on 2 ports of a regular (b/g) system, then the 2 antennas are switched by a RF switch on the antenna line (that small 6 pin IC on the output) and the best antenna will be used by trial and error. NO interfecrence is possible because of CDMA access (AP doesn't send on busy RF channel) and because ONLY 1 ANTENNA is connected to the transceiver at any point in time.

This is regular diversity and is not supported by Mikrotik.

Well, Mikrotik itself calls it "diversity". So what is "regular diversity" exactly?
On the MT cards 'antenna diversity' means the listening cycle on the MT-card is on both chains (same time or one by one I don't know) and the radio then talks to an CPE through that chain that gives best signal.
Since this all happens within the normal cycle of rx/tx I would call that "diversity".

But maybe you have a more technical explanation on the issue?

so how is it using both omnis? i suppose putting them in L shape will not make difference to interference?

okay =) i.ll play a bit
@moment do not have enough money to spend on that base, to buy omni mimo antenna, so i have bought used ones 12dbi, if there will be need to improove, i'll exchange them perharps i should use both in vertical position to have better signal etc