What are 2 antennas being used for?

snipped-for-privacy@ipal.net hath wroth:

Diversity reception.

Two antennas are diversity receive. Three antennas are MIMO which is quite different. No antennas are beam forming or beam steering.

Well, it would be nice if you would limit the question to a specific model or type of radio.

Wait a minute. You're asking a question and offering multiple choice answers? I recognize your deductive abilities but wouldn't it be better if you ask and not guess?

That's not the way it's usually done. Most dual band access points use common antennas for both bands. However, there are chipset that seperate the bands and it could be done this way. It's very common with the new 4.9Ghz/2.4Ghz MotoMesh access points, that have seperate antennas for each band. However, for commidity access points, the antenna is usually (not always) common for both bands.

That was done on some access points about 5 years ago, where one antenna was outside, and the other was inside the plastic box. I vaguely recall the DLink DWL-900AP+ was one of these. The inside antenna was almost totally useless for transmit, so all transmissions were from the outside antenna. I'm not sure why they even bothered with receive diversity.

This is the most common arrangement. What's not obvious is that there is a MAIN and AUX antennas. The radio sits on the MAIN antenna most of the time and only tries the AUX antenna if it detects a high error rate. The main advantage is to reduce frequency selective fadeing in a reflective environment (such as all indoor systems). See: |

| | a few articles on the common methods of wireless diversity.

Well, with two rubber ducky antennas, you'll never realize enough gain to make the effort worthwhile. Worse, with the antennas twisting in random directions, there's no way to aim it. That's not going to happen.

However, there is a class of access points that use beam streering, and another that uses beam forming, for directing the signal. Inside the box is an elaborate antenna array. For beam steering, the idea is to put the peak at toward the client radio. Beam steering is considerably more elaborate in that it also aims the peak, but also detects sources of interference and directs a null at the interference. |

That doesn't work. Read the Cisco article mentioned previous for details. Another major problem is that traffic through the access point, from one antenna to the other doesn't get switched fast enough. If you were to try to use it as a repeater between antenna ports, it would need to switch antennas on literally every packet. The diversity algorithm is nowhere near that agressive.

On 24 Jul 2006 19:30:05 GMT, snipped-for-privacy@ipal.net wrote in :

Spatial diversity.

John Navas hath wroth:

Not applicable. The above article is for common techniques used in telecom wireless. In all cases mentioned, the multiple signals received are combined in some manner. This is NOT done in any commodity Wi-Fi routers. It would be too expensive to have two identical receivers in one box. Instead, the methods of Wi-Fi diversity are called "switched diversity" or "selection diversity" as mentioned in this article: |

requires only one receiver and are admittedly not as effective as full time diversity (as described in the Wikipedia article).

On Mon, 24 Jul 2006 23:20:39 -0700, Jeff Liebermann wrote in :

Close enough for this context.

Update the article. ;)

On Mon, 24 Jul 2006 22:58:11 -0700 Jeff Liebermann wrote: | snipped-for-privacy@ipal.net hath wroth: | |>What are 2 antennas being used for? | | Diversity reception. | |>WRT54G and many others have 2 or even 3 antennas. | | Two antennas are diversity receive. | Three antennas are MIMO which is quite different. | No antennas are beam forming or beam steering.

So only one of the antennae is used for transmitting. Dare I ask which one :) ... ah, answer below.

|>There are a number of things I could envision what |>they might do with this. Anyone know what they actually are doing? | | Well, it would be nice if you would limit the question to a specific | model or type of radio.

It's a generic question because I see a number of radios with 2 or 3 antennas. The answer for one model would not necessarily give me an idea of the scope of what is being done overall. If two different models do different things, I'd like to get an idea of possibilities.

|>Here are some possibilities. | | Wait a minute. You're asking a question and offering multiple choice | answers? I recognize your deductive abilities but wouldn't it be | better if you ask and not guess?

The examples were intended to give the scope and level of answer I am looking for in that question. Were one of those guesses true, it would be an exact answer to my question. It's not really a multiple choice question. If the answer is something else, then all guesses are wrong.

|>1. Split band. For dual band models, one antenna could be attached |> to 2.4 GHz RF circuitry, while the other is attached to 5 GHz RF |> circuitry. This would avoid the need for an internal splitter and |> a dual band antenna. | | That's not the way it's usually done. Most dual band access points | use common antennas for both bands. However, there are chipset that | seperate the bands and it could be done this way. It's very common | with the new 4.9Ghz/2.4Ghz MotoMesh access points, that have seperate | antennas for each band. However, for commidity access points, the | antenna is usually (not always) common for both bands.

And so not likely concurrent operation on both bands at the same time other than simultaneous receive.

|>2. Split RX/TX. One antenna could be used for receive, while the other |> is used for transmit. I don't see any significant advantage to this |> unless the TX power is low enough to still allow RX on a different |> channel for models than can operate on 2 channels at the same time. |> For remotely wired antennae, this might be an advantage. | | That was done on some access points about 5 years ago, where one | antenna was outside, and the other was inside the plastic box. I | vaguely recall the DLink DWL-900AP+ was one of these. The inside | antenna was almost totally useless for transmit, so all transmissions | were from the outside antenna. I'm not sure why they even bothered | with receive diversity. | |>3. Diversity. Two receivers work together in case one antenna position |> would be in a null spot for another device, the 2nd antenna might |> get the signal. Transmit might also select which antenna provides |> the better path to the destination. With extra compnents used to |> realize the receive advantage, this might not be done. | | This is the most common arrangement. What's not obvious is that there | is a MAIN and AUX antennas. The radio sits on the MAIN antenna most | of the time and only tries the AUX antenna if it detects a high error | rate. The main advantage is to reduce frequency selective fadeing in | a reflective environment (such as all indoor systems). See: | |

| | | | | for a few articles on the common methods of wireless diversity.

You're right, MAIN vs. AUX was not obvious. I suppose if I have worked on trying to guess every possible combination of how to do things, that would have come up among them.

So for transmit, it would alayws be from the MAIN antenna, and leave it up to the other end to have receive diversity in the event that MAIN on one unit to MAIN on the other unit happened to be a bad RF path.

|>4. Directionality. Similar to diversity, but an RF stage taking signal |> on both antennae at the same time can get up to 3db gain in certain |> directions (and be entirely deaf in others). This could be used to |> enhance certain weaker devices and/or null out interference sources. |> TX could do the same. This would be difficult for the average user |> to manage, and require extra components, so I doubt it would be done. | | Well, with two rubber ducky antennas, you'll never realize enough gain | to make the effort worthwhile. Worse, with the antennas twisting in | random directions, there's no way to aim it. That's not going to | happen.

The distance between them would dictate a beam pattern, if merged at the RF level. But it would be more confusing with wireless networking than it would be trying to adjust rabbit ears on an analog TV (digital TV is going to mess that up, too).

| However, there is a class of access points that use beam streering, | and another that uses beam forming, for directing the signal. Inside | the box is an elaborate antenna array. For beam steering, the idea is | to put the peak at toward the client radio. Beam steering is | considerably more elaborate in that it also aims the peak, but also | detects sources of interference and directs a null at the | interference. | |

Much more expensive, no doubt.

| |>5. Placement. Wire a remote antenna on ONE of the connectors to get |> coverage in two distinct places. | | That doesn't work. Read the Cisco article mentioned previous for | details. Another major problem is that traffic through the access | point, from one antenna to the other doesn't get switched fast enough. | If you were to try to use it as a repeater between antenna ports, it | would need to switch antennas on literally every packet. The | diversity algorithm is nowhere near that agressive.

Or basically be 2 radios in one, just on the same band and programmed to not transmit at the same time (unless on different channels, but then, that's a different animal).

On 25 Jul 2006 15:56:40 GMT, snipped-for-privacy@ipal.net wrote in :

I think transmit will typically occur on the current antenna; i.e., if AUX is getting the best receive signal, then it will be used for transmit as well. Using only MAIN for transmit would be a poor design.

snipped-for-privacy@ipal.net hath wroth:

It's easy enough to check. Find a dual trace oscillosope and connect CH A to the diversity switch IC and CH B to some place that indicates if the radio is transmitting. See what it's doing. I did this about

3 years ago with a WRT54G in my rather sloppy attempt to reverse engineer how the diversity switch actually worked (and to figure out which is are the left/right and main/aux antennas). I also tried to connect a red/green LED to the diversity switch IC to indicate what it was doing, but the switching rate was just too fast to be useful.

If I have time, I'll do it again. It's fairly easy, but the local temperatures have been far to high to do anything useful in the shop.

It apparently varies with chipset. As I vaguely recall (not sure), the Broadcom chipset in the WRT54G switches TX along with RX. I believe (not sure) that some chipsets can be set to work either way. One thing I do recall is that the switch spent almost all its time on the MAIN antenna.

Not really. See:

an example. About $90. It's the mechanical parts that add cost. The smart antenna board eliminates a mess of connectors, cables, antennas, hardware, assembly labour, found in conventional diversity access points and is probably a cost savings.

What do you mean by 2 radios in one? I'll guess you mean full duplex, which is possible but has not commonly been implimented in consumer wireless. BelAir Networks and some others have dual radio access points that can act as repeaters. It's used in mesh networks to eliminate the bandwidth reduction caused by single channel systems. No diversity of any kind as each antenna is connected to a seperate radio. |

There was also Bermai, which had a dual receiver (TwinRX) diversity system for 802.11a that was targeting consumer wireless multimedia. With two receivers, they implimented full time diversity reception with no switching. | company closed down in 2004. No idea what happened to the assets or IP rights.

Or it could just transmit on both antenna at the same time; which would probably be the most sensible way, as well as the easiest.

Actually, this could be the worst possible thing to do, depending on phasing and antenna separation (to say nothing of antenna angle) you could get some really bizarre antenna patterns. [The thought of external antennas in this situation makes my brain hurt.]

On Tue, 25 Jul 2006 16:44:25 GMT John Navas wrote: | On 25 Jul 2006 15:56:40 GMT, snipped-for-privacy@ipal.net wrote in | : | |>... MAIN vs. AUX was not obvious. I suppose if I have worked |>on trying to guess every possible combination of how to do things, that |>would have come up among them. |>

|>So for transmit, it would alayws be from the MAIN antenna, and leave it |>up to the other end to have receive diversity in the event that MAIN on |>one unit to MAIN on the other unit happened to be a bad RF path. | | I think transmit will typically occur on the current antenna; i.e., if | AUX is getting the best receive signal, then it will be used for | transmit as well. Using only MAIN for transmit would be a poor design.

It could be a cheaper design (no TX switch), maybe by a few pennies. And maybe TX only on MAIN is adequate if the peer has 2 antennae. Of course the latter assumption cannot be depended on. So TX follow best RX would be a good idea. Better would be to test TX on each antennae and remember which works best per peer for the off chance one antenna works better for RX and the other worked better for TX, such as when the peer has split RX and TX without switching. But the adde software cost and extra test TX step might not be worth the few cases where this could help.

|>>... MAIN vs. AUX was not obvious. I suppose if I have worked |>>on trying to guess every possible combination of how to do things, that |>>would have come up among them. |>>

|>>So for transmit, it would alayws be from the MAIN antenna, and leave it |>>up to the other end to have receive diversity in the event that MAIN on |>>one unit to MAIN on the other unit happened to be a bad RF path. |>

|> I think transmit will typically occur on the current antenna; i.e., if |> AUX is getting the best receive signal, then it will be used for |> transmit as well. Using only MAIN for transmit would be a poor design. |>

| | Or it could just transmit on both antenna at the same time; which would | probably be the most sensible way, as well as the easiest.

There may be a case where transmit on both can reach the peer better than either alone. However, if the peer is in the null of the resultant pattern then doing this is counter productive. OTOH, a different phasing between the two antennae might be usable to target the peer better. But that is definitely more circuitry to independently rephase the antennae.

Ding, ding! Give that man a prize. At $50 typical retail for most wifi routers every penny counts. That and since the units are so cheap it's often better to use more than one. Taking care to properly configure the channels, of course. Something far too many installations fail to take into consideration.

If your access point is on channel 11 and there are other access points also on that same channel, then MOVE yours to a new channel.

snipped-for-privacy@ipal.net hath wroth:

There's a bit more going on with diversity than is obvious. The manufacturers would have eliminated the 2nd antenna long ago if they thought they could save a few pennies. Diversity significantly improves the reflection immunity of an indoor system, but such improvements are not what sells commodity routers. Price is what sells commodity routers.

One manufacturer did try to remove the 2nd antenna. The Dlink DI-614+ and DI-624 went from two antennas to one antenna in later mutations. I was told that retail sales immediately dropped. It wasn't a performance issue. Customers perceived that two antennas are somehow better than one and considered a single antenna router to be inferior. DLink still sells some older products with one antenna, but all the new DLink routers have two (or three).

OR do like I just did. Buy a "N" Router and force your next door guy to another channel. To piss him off put it on the same channel as he is. It does not affect yours, But it does his. :) My Motto in this case is what comes around goes around.

Hmmm, and with that sort of attitude I'd avoid standing anywhere near you should thunderclouds start forming. Sounds like you're asking for it with that sort of attitude.

"Look at the tail fins on that baby. It must be fast or else why would they put tail fins on it???"

I've always suspected that the whole two antenna hack started by someone wanting to differentiate their offerings from the crowd.

-wolfgang

On Thu, 27 Jul 2006 14:33:38 -0400, "Bill Kearney" wrote in :

Amen. With a neighbor like that I'd probably move.

On Thu, 27 Jul 2006 09:22:17 -0700, Jeff Liebermann wrote in :

Not that simple. Poorer performance tends to result in a higher return rate, which can be a killer on cost that far overwhelms the small savings. And people don't just buy on price alone -- reviews and recommendations are factors, for which performance is also important.

And they were right, whether they knew the exact reason or not.

John Navas hath wroth:

Chuckle. Return rate is a problem but not for the obvious reason that it affects the bottom line. The return rate on commodity wireless is sufficiently small that at least one manufacturer just throws them away and only retests them if a major distributor returns them en masse. They also unload them with large rebates, which requires destroying the packaging to collect, which therefore reduces the return rate. Return rate is a killer only because it trashes brand name recognition, which is the major reason consumers buy any particular device. If you look carefully at the advertising in the consumer magazines (not the trade journals), you'll probably notice that much of starts by congratulating the consumer on having made an intelligent choice and goes on to assure the consumer that the company is behind its products 150%. In other words, they're going for the repeat business based on name recognition.

Yep. Just ask Joe Sixpack which reviews he's read on his wireless router. Ummm...well...er...ah... Maybe it came with my computer? You, me, and Dilbert probably read reviews in detail. Only a few of my customers do that. Mostly they have a friend or consultant that has a wireless something, and they just buy the same thing and blame their friend or consultant when it doesn't quite work.

Sure. Perception is everything. Few can explain why two antennas are better, but intelligence and technology doesn't count with decisions based on perception.

There are other things that are amazingly important for retail sales. Color is one. Colors all have subliminal meanings. I walked in to a customer with a Watchguard SOHO router. It's in a bright red plastic box, apparently to capitalize on the firewall features. My customer asked if I had something in a different color? Huh? She said that it reminded her of blood, which made her feel awkward. I replaced it with an antiseptic white Netgear WG-614, which was deemed acceptable.

Weight is also an important feature. Given two almost identical products, the average consumer will usually pick the heaviest product. There's a perception that you get more for your money if it's heavier. I learned this the hard way when designing marine radios. We literally put a lead brick inside the box and sales immediately improved.

Criteria for commodity router selection (most important on top): 1. It's cheap. 2. A friend has one that works. 3. I've heard of the manufacturer from somewhere. 4. The box and color look cool. 5. It weighs like something that should work. 6. The literature is incomprehensible, so it must be powerful. 7. Larger numbers are always better. 108Mbits/sec instead of 54Mbits/sec. Is 802.11z later than 802.11b? 5.7GHz is bigger than 2.4Ghz. 8. The flashing lights sure look nice. What do they mean? 9. This box has more acronyms than the other box. 10. I read a review that said all I had to do is plug in the wireless router and it's ready to go. Optional: 11. Jeff L said it sucked so I guess I'll try it.