[awelkie]

Sure, but the probability would be low-ish of that happening, and the other system could either switch frequencies or beamform a null in the direction of the interferer if they were also a phased array.

Maybe the EIRP shouldn't be unlimited, but I still think it would be beneficial to encourage spatially selective systems.

[adrian_b]

In an unlicensed band, everyone may have receivers and transmitters, so the probability of other receivers in the same direction as yours is not low at all, but it is very high, unless you live in the middle of nowhere, so you do not have neighbors.

There is no justification for imposing additional costs for others in order to accommodate your desires that do not matter for them.

Nobody stops you to use a phased array antenna only to obtain a higher gain for reception, in order to increase the communication range.

Even without phased array antennas, using just classic directive antennas that are placed on high masts, it is possible to communicate through WiFi at tens of km (but only at low bit rates and not in all countries, as some have more severe EIRP limits).

The problem of directive antennas is that they are usable only for fixed positions of access points and wireless stations.

Phased array antennas are not enough to enable mobility, because initially a mobile wireless station must discover the direction of the AP and the AP must discover the direction of the station, by using omnidirectional transmission, which limits the range to what can be achieved without phased array antennas.

To use a mobile wireless network that works at distances greater than possible with omnidirectional antennas requires much more sophisticated equipment than just the phased array antennas. You also need means to determine the coordinates of each station (and of the access points, if they are also mobile) and maps with the locations of the access points so that a station that wants to associate with them will know in what direction to transmit. You also need a protocol different from standard WiFi, e.g. the access point may need to scan periodically all directions in order to allow new associations from distant stations.

[Dylan16807]

> so the probability of other receivers in the same direction as yours is not low at all, but it is very high

But on average, even with unlimited EIRP, I'll see 1/n as many interfering signals that are each n times as strong. That's not a bad tradeoff.

But having a moderate EIRP increase for focused signals would make things better for everyone. Let's say a signal that's 10x as focused can have 3x the EIRP, and everyone switches their equipment over. That drops the total power output by 3.3x, and interference drops significantly for almost everyone.

> initially a mobile wireless station must discover the direction of the AP and the AP must discover the direction of the station, by using omnidirectional transmission, which limits the range to what can be achieved without phased array antennas

You can do discovery at a lower bit rate to get a big range boost.

[adrian_b]

For a radio receiver it is irrelevant how many interfering signals exist.

The only things that matter are the radiant intensity (i.e. power per solid angle) of the interfering transmitter and the percentage of the time when that transmitter is active.

A single interfering transmitter with high radiant intensity (a.k.a. EIRP) will blind the radio receiver for all the time when it is active.

Doing discovery at a low bit rate is fine, but that means that your fancy phased array antenna cannot achieve any higher distance for communication than an omnidirectional antenna, but it can only increase the achieved bit rate at a given distance.

That would be OK, except that it is achieved by interfering with your neighbors, exactly like when using a transmitter with a higher total power than allowed.

Limiting EIRP is the right thing to do in order to limit the interference that you can cause to your neighbors.

The law does not stop you to use a phased array antenna or any other kind of directive antenna, with the purpose of lowering the power consumption of your transmitter, while maintaining the same quality for your communication and the same interference for your neighbors.

What you want to do, i.e. increase the interference for your neighbors, is the wrong thing to desire. If that were allowed, your neighbors would also increase the radiant intensity of their transmitters and then everybody would have worse reception conditions and you would gain nothing.

The hope that only you will increase your radiant intensity and your neighbors will not, is of course illusory.

[Dylan16807]

> For a radio receiver it is irrelevant how many interfering signals exist.

> The only things that matter are the radiant intensity (i.e. power per solid angle) of the interfering transmitter and the percentage of the time when that transmitter is active.

And if a transmitter isn't pointed at you, then it isn't an interfering transmitter. This is a crucial factor in the math.

Or for a more realistic analysis of directionality, the radiant intensity is only high for a small fraction of observers, and is very low for the rest of them.

In the first scenario I talked about, total interference is probably the same.

In the second, total interference is almost always much less.

> A single interfering transmitter with high radiant intensity (a.k.a. EIRP) will blind the radio receiver for all the time when it is active.

If a moderate boost blinds the receiver, then the alternative is being almost blind for much longer (because there are more interfering transmitters), so I'm not convinced that's a problem.

> Doing discovery at a low bit rate is fine, but that means that your fancy phased array antenna cannot achieve any higher distance for communication than an omnidirectional antenna, but it can only increase the achieved bit rate at a given distance.

I don't understand what you mean.

If you don't care about speed, the maximum distance is the same for both antennas, and is defined by obstacles alone.

You can always slow down to compensate for a lack of gain. And it's a proportional slowdown, not very expensive. Especially when you only need to send a beacon that's a few bytes long to initiate contact.

Just some example numbers: Your functional requirements are 1Mbps of bandwidth with pretty tight focus. You send the few bytes of initial omnidirectional contact at 1Kbps. Your slow omnidirectional signal actually reaches further than your fast focused signal.

> Limiting EIRP is the right thing to do in order to limit the interference that you can cause to your neighbors.

If your only concern is the worst case of everyone being pointed at the same spot, yes. In normal situations the average level of interference matters more.

> The law does not stop you to use a phased array antenna or any other kind of directive antenna, with the purpose of lowering the power consumption of your transmitter, while maintaining the same quality for your communication and the same interference for your neighbors.

The law says that I can maintain quality and decrease interference, but I don't gain any real benefit because I'm only saving half a watt. So I'm not very motivated to do so. I'd prefer if it was legal to split the difference between increased signal quality and somewhat decreased interference.

> What you want to do, i.e. increase the interference for your neighbors, is the wrong thing to desire.

Where do you think I said that?

Edit:

> your neighbors would also

Here, I'll elaborate on a scenario.

Originally, me and my 6 neighbors are all transmitting omnidirectionally and causing 1 unit of interference to each other person. Everyone gets 6 units of interference.

I want to lower my transmit power but double my EIRP. I am entirely selfishly motivated, and just want a better signal to my devices. As a consequence I will now cause 2 units of interference to a single neighbor, and 0.1 units of interference to all other neighbors.

What happens when everyone thinks this way and does the same thing? Well now the neighborhood receives 2.5 units of interference on average instead of 6. Even with a bit of variance as devices move around, everyone is better off now. I love that my neighbors did the same thing I did.

[adrian_b]

> And if a transmitter isn't pointed at you, then it isn't an interfering transmitter. This is a crucial factor in the math.

This does not really matter, because in the dense wireless networks that are typical for the unlicensed bands, due to the ubiquitous WiFi and Bluetooth devices, there is always a transmitter pointed at you. Typically there are many transmitters pointed at you.

So any argument based on this idea that there are no transmitters pointed at you would fail badly in practice.

Moreover, the interference in digital communications is not something that grows linearly. It grows in jumps, when some thresholds are exceeded and the error correction used by transmitters fails to work. So at certain thresholds the interference would force your device to reduce the bit rate, until a certain threshold where communication would become impossible until the interfering transmitter stops transmitting.

Also, 6 neighbors do not provide 6 units of interference. The amount of interference depends on many factors. The neighbors that use the same communication channels will attempt to not transmit simultaneously, to avoid collisions.

When a neighbor transmits on the same channel, then it is guaranteed that the interference is so high as to prevent other simultaneous communications. So the interference that we discuss is from transmitters that use other channels than yours.

Besides the fact that the neighbors are partitioned in groups within which only one transmits (but almost all the time there is an active transmitter), the interference depends greatly on the distance to the transmitters.

So to estimate the change in interference when all replace their omnidirectional antennas with directive antennas, increasing the radiant intensity, is far more complex than your simple arithmetic.

The worst case, which can never be excluded, is that there will be at least one transmitter pointed at you and its higher radiant intensity will be enough to cross the threshold at which communication becomes impossible for yourself. In this case it is completely irrelevant if you no longer have interference from other transmitters that are not pointed at you.

Planning wireless networks cannot be done based on hopes that you will be the luckiest in the universe and Murphy's law will not apply to you.

[Dylan16807]

> there is always a transmitter pointed at you

A transmitter, sure. If you go from having 40 transmitters pointing at you, to now having 6 transmitters pointing at you, that makes a big difference. Even if they're running at twice the EIRP now, that's a big improvement.

> So any argument based on this idea that there are no transmitters pointed at you would fail badly in practice.

My argument doesn't depend on that.

> The amount of interference depends on many factors. [...] So to estimate the change in interference when all replace their omnidirectional antennas with directive antennas, increasing the radiant intensity, is far more complex than your simple arithmetic.

Yes I simplified. But does that completely upend the result? If so, show me the math that makes it happen.

> groups within which only one transmits (but almost all the time there is an active transmitter)

> cross the threshold at which communication becomes impossible for yourself

And guess what? If everyone doubles their EIRP but transmits in a much narrower beam, the area in which that happens becomes smaller. The number of transmitter pairs that need to time-share decreases.

> Planning wireless networks cannot be done based on hopes that you will be the luckiest in the universe and Murphy's law will not apply to you.

I think your argument depends on me being lucky in the omnidirectional case but unlucky in the directed transmit case. That's not a reasonable way to assess alternatives.

For every percent chance that higher-EIRP directional transmit causes me problems, there's a bigger chance that higher-total-power omnidirectional transmit causes me problems.