Does anybody has the knowledge what actually happens when we set channel to work with one of the mentioned channel bandwiths?
I am referring to the reciever. The transmission part is clear, transmitter stansmits in the bandwith given.
But the reciever, is the filter still alowing bigger channel to enter the reciever than what the transmitter is set for.

Or, in day to day problem shooting: If AP is set to work with smaller channels to create more useable working channels in band, will same AP still pickup inteference from other AP's that still work in 20Mhz (or bigger) band of adjacent freq's?

Example: I set my AP to 5610Mhz-10Mhz. So the channel range is 5605-5615Mhz.
I also set other AP to 5600-10Mhz, the channel range is 5595-5605Mhz so it just stays clear of the first. As long as they both are not in one mast and clients don't pick up both signals at same strength we have no problems
But competition now is not aware of these channels because default is 20Mhz so when he scans for a free channel he is not seeing my AP's. He decide to use 5600-20Mhz. His channel range is now 5590-5610Mhz which overlaps both of my AP's!
Do both of my AP's now still allow the full energy of the 20Mhz radiowave to hit the reciever? Or are the recieving filters now also narrowed down enough so only very little of his 20Mhz energy hits my recievers?

Please some insight in this process?

Does anybody has the knowledge what actually happens when we set channel to work with one of the mentioned channel bandwiths?
I am referring to the reciever. The transmission part is clear, transmitter stansmits in the bandwith given.
But the reciever, is the filter still alowing bigger channel to enter the reciever than what the transmitter is set for.

I asked the same question http://forum.mikrotik.com/viewtopic.php?f=7&t=49835,but did not get any replies from radio card manufacturer's?

Example: I set my AP to 5610Mhz-10Mhz. So the channel range is 5605-5615Mhz.
I also set other AP to 5600-10Mhz, the channel range is 5595-5605Mhz so it just stays clear of the first. As long as they both are not in one mast and clients don't pick up both signals at same strength we have no problems
But competition now is not aware of these channels because default is 20Mhz so when he scans for a free channel he is not seeing my AP's. He decide to use 5600-20Mhz. His channel range is now 5590-5610Mhz which overlaps both of my AP's!
Do both of my AP's now still allow the full energy of the 20Mhz radiowave to hit the reciever? Or are the recieving filters now also narrowed down enough so only very little of his 20Mhz energy hits my recievers?

This is a classic example of the limitations of freq scan on MK, when you scan on 10mhz it will not show the AP’s that are using 20Mhz bandwidth in the scan, same applies to 5Mhz scan,
For us WISP’s having access or the purchase of a cost effective spectrum analyser which will perform the functions we require for our business without being over specified is a must have item?

MT, any reply/comment on this?

This is a classic example of the limitations of freq scan on MK, when you scan on 10mhz it will not show the AP’s that are using 20Mhz bandwidth in the scan, same applies to 5Mhz scan,
For us WISP’s having access or the purchase of a cost effective spectrum analyser which will perform the functions we require for our business without being over specified is a must have item?

Is it really that hard to set it to 5, 10 or 20 when you are scanning for other networks?

This isn't just Mikrotik, this is the way every Atheros based AP I've ever used works.

I use the frequency usage tool, and in The Dude you have a spectral scan if you have a new enough chipset, also works in terminal, but I've never tried it there.

maybe in future we will try to include a option to scan through all the bands 5/10/20Mhz, but currently no date yet.
I would recommend to use spectral-scan/-history feature to see what is going on in the wireless network.

maybe in future we will try to include a option to scan through all the bands 5/10/20Mhz, but currently no date yet.
I would recommend to use spectral-scan/-history feature to see what is going on in the wireless network.

Good idea but as always looking for extra functions is it possible to include in this scan a option to reveal propriety stuff or interference spike in the band?

Is it really that hard to set it to 5, 10 or 20 when you are scanning for other networks?

I would ask the question how many users are aware of this limation of the scan?

maybe in future we will try to include a option to scan through all the bands 5/10/20Mhz, but currently no date yet.
I would recommend to use spectral-scan/-history feature to see what is going on in the wireless network.

Nice, this would make life a bit easier again. As long as users are still aware that none 902.11 signals in the same frequencies are still not picked up. A good spectrum scanner is needed to see what is really happening, I agree on that.

But this still leaves my initial question un-anwered: By squeezing the transmitter to 10 or 5Mhz channel or open it up to 40Mhz. What happens on the reciever end?

How is this working? If a radio has to listen to a chosen freq. in a band, how is the recieving filter set to have the reciever only listening to that specific frequency. This must be something that is arranged by the software. Most card we use can cover 4.9 to 5.8Ghz radio band (and 2,4Ghz) but I don't believe reciever is open for all signals in that band. So it is set according to what the radio transmitter is set for.

But does the same story counts for 5,10,20 and 40Mhz bandwith?
If radio A is set to work with 5Mhz, does it still pick up a lot of signal from another radio B only 10Mhz away. When radio A would be set to 29 or even 40Mhz but in the same working (centre) he WOULD be sharing that frequency with that radio B!

Or is the software capable of also limiting/filtering to/on the narrow bandwith only?

This might not seem to be so important if you have plenty of spectrum available; just go to anoter freq. channel. But in 5Ghz crowded areas, and we will see this more and more in urbanised areas, this interference issue, is going to be a problem one day or another. Just like it already is a big problem on 2,4Ghz band...

There must be info available or some radio technician must be able to tell us how, and how good, it is done....

Perhaps a referemce to this document would help.

http://src.acm.org/raghavendra/raghavendra.html

Very interesting reading, but then you know my views on huge bandwidths!

I could start a huge fight on the forum by asking the questions!

Do you have to provide 100Mbps UDP/TCP to a client or is it because YOU WANT TO!

PS, Rudy, I am not having any problems with MF in 4.17 with NV2, but I am still having the original issues with drop outs with assorted reasons, basically I have fixed the scan to the AP channel and the disconnect/reconnect is within the same second.. So I am not going to waste anymore time with it until NV2 has had ALL the bugs sorted out. (and I am not intending to convert anymore of the network to it, until resolved)

I still don't understand your views regarding NV2 and adjacent channel interference, the 2 are not related.. NV2 is a handshake protocol and has no impact on bandwidth, however close proximity to another operator who is using "N" is a different matter completely!

Maybe one of the competitors is using some of the higher frequencies that we discussed, some equipment manufacturers are using the 5ghz band like an IF stage. One of those in proximity to you will also cause a problem.

As a foot note, I don't allow my customers to use N protocol on home based routers!!!!

maybe in future we will try to include a option to scan through all the bands 5/10/20Mhz, but currently no date yet.
I would recommend to use spectral-scan/-history feature to see what is going on in the wireless network.

Nice, this would make life a bit easier again. As long as users are still aware that none 902.11 signals in the same frequencies are still not picked up. A good spectrum scanner is needed to see what is really happening, I agree on that.

But this still leaves my initial question un-anwered: By squeezing the transmitter to 10 or 5Mhz channel or open it up to 40Mhz. What happens on the reciever end?

How is this working? If a radio has to listen to a chosen freq. in a band, how is the recieving filter set to have the reciever only listening to that specific frequency. This must be something that is arranged by the software. Most card we use can cover 4.9 to 5.8Ghz radio band (and 2,4Ghz) but I don't believe reciever is open for all signals in that band. So it is set according to what the radio transmitter is set for.

But does the same story counts for 5,10,20 and 40Mhz bandwith?
If radio A is set to work with 5Mhz, does it still pick up a lot of signal from another radio B only 10Mhz away. When radio A would be set to 29 or even 40Mhz but in the same working (centre) he WOULD be sharing that frequency with that radio B!

Or is the software capable of also limiting/filtering to/on the narrow bandwith only?

This might not seem to be so important if you have plenty of spectrum available; just go to anoter freq. channel. But in 5Ghz crowded areas, and we will see this more and more in urbanised areas, this interference issue, is going to be a problem one day or another. Just like it already is a big problem on 2,4Ghz band...

There must be info available or some radio technician must be able to tell us how, and how good, it is done....

Very interesting topic. What I see is that "filtering" is dependent on signal strength and frequency spacing.
So a strong signal even 100MHz away floods the receiver. Using 10MHz Channel with the same power you
add some frequency spacing but the energy put to this 10MHz Channel is the same you put to the 20MHz
Channel before.
I see this behavior depends on the wireless/card equipment used. So the maker of the card has influence on this.
There should be a parameter for the wireless-card which tells the buyer how good the "filtering" of nearby
frequency is. Like Front to Back ratio for an antenna.

There is a TX frequency mask which tells how steep the signal curve falls to the sides of a signal
(we've to deliver this parameter to get licenses for licensed frequency-gear). Better designed
card/gear has a steeper curve. Such a mask should be known for the receiver to see how
far away and how much power a nearby signal can have to generate noise below a given level.

The quality of this nearby signal filtering might even influence the efficiency of 802.11/nstreme/nv2.
As nstreme/nv2 do more traffic on quiet links, it might help to go back to 802.11 on some links to
reduce interference (Or sync so you dont listen when your neighbors shout )).

But if you put the radio in plastics and use low gain antennas all of this does not really matter

Perhaps a referemce to this document would help.

http://src.acm.org/raghavendra/raghavendra.html

Simon, indeed a very interesting document. Although I do not see the implications when same dynamic channel bandwith is to be used in 802.11n devices. I think here a bit more work is needed to get it done. But anyway, future as I see it.

Very interesting reading, but then you know my views on huge bandwidths!

As you know I started using 10Mhz channel for some of my AP networks and so far the results are in line with some of the conclusions we read in mentioned document. Clients at the edge of AP's range did indeed see an increase of 2-3dB. Links to these and closer clients ALL became more stable.
Since clients can consume up to 4Mb sustained (subscribed client speeds) only chosing a connection rate of 18Mb which give 9Mb actual data throughput has proven to be sufficiant and stabilized links even more.
5-6Mb tcp bandwith test are possible.

PS, Rudy, I am not having any problems with MF in 4.17 with NV2, but I am still having the original issues with drop outs with assorted reasons, basically I have fixed the scan to the AP channel and the disconnect/reconnect is within the same second.. So I am not going to waste anymore time with it until NV2 has had ALL the bugs sorted out. (and I am not intending to convert anymore of the network to it, until resolved)

I use nv2 now on several AP and backhaul links. But I had an upgrade to 5.0 fail due an disabled wireless package afterwards that needed a roll out. This issue seems to be under the attention of MT so I am waiting until 5.1 has fixed this to avoid the risk on more roll outs for units that still are up for the upgrade.
But all my links running in sole MT network with a mix of 4.16, 4.17 and 5.0rc12 are relative stable after I re-arranged several channels and narrowed several AP's to work in 10Mhz range.

I still don't understand your views regarding NV2 and adjacent channel interference, the 2 are not related.. NV2 is a handshake protocol and has no impact on bandwidth,

NV2 is TDMA. that is a different way how AP communicate with station with time slots etc. to achieve a better use of spectrum. NV2 also comes with some other options:
Taken fm MT manual: "# Reduced propagation delay overhead - There are no per-frame ACKs in Nv2 - this significantly improves throughput, especially on long distance links where data frame and following ACK frame propagation delay significantly reduces the effectiveness of media usage.
# Reduced per frame overhead - Nv2 implements frame aggregation and fragmentation to maximize assigned media usage and reduce per-frame overhead (interframe spaces, preambles). "
So yes, it has impact on bandwith.

however close proximity to another operator who is using "N" is a different matter completely!

Why shoud this be a ´diferent matter completely!"? nv2 imho opinion has influence on any other frequency band as long as the reciever of unit in same or adjacent freq. channel can ´hear´ the nv2 radio.
I don´t see why ´n´ protocol is different in this respect than legacy ´a´?

Maybe one of the competitors is using some of the higher frequencies that we discussed, some equipment manufacturers are using the 5ghz band like an IF stage. One of those in proximity to you will also cause a problem.

Hmm, could be. Hence again my original question comes up: how narrow is radio wave recieved at the reciever compared to its set sending bandwith. Or; how good is the mask in front of the reciever to protect it from not wanted radio wave energy! I want to find out if a 40Mhz channel from a ´n´ unit, or a 20Mhz, that is only 15Mhz away from my centre 10Mhz channel ´hurts´ my reciever?
The interesting document fm your post is not giving any clue to this. Nor can I find much usable info on the internet about this.

As a foot note, I don't allow my customers to use N protocol on home based routers!!!!

You are telling your customer not to buy, recieve or use wifi routers they buy in the shop?
These are becoming the standard and other WISP users are going to use them as bonus to get new clients in. I don't see your point in preventing your clients to use these.
If you are afraid of these disturbing your network, what about his neigbour that is not your client?
Most WISP's moved fm 2,4Ghz to 5Ghz to stay ahead of the ´domestic user interference´ issue. Or they flee to Wimax.
Now we have NV2 (TDMA) to stay away from this monster a little longer but in the end we have to keep running for it.
I already pick up domestic 5Ghz signals from private users on my links shooting over urbanized areas. 5Ghz domestic signal usually come from the latest generation domestic ´n´ routers.
Netgear and some others already have wifi routers that are multi channel all protocol (a/b/g/n) and if these become widespread in urbanized areas we can only hope 802.11 free license band doctors came up with better protocols to ´harden´ our links otherwise we are doomed...

The quality of this nearby signal filtering might even influence the efficiency of 802.11/nstreme/nv2.
As nstreme/nv2 do more traffic on quiet links, it might help to go back to 802.11 on some links to
reduce interference

I'm afraid I don't completely understand this. From reading and what I see happening is that it is more the other way around. My links running nv2 don't see any problems. It is my ´old´ 802.11 links and PtMP networks that seem to suffer.

(Or sync so you dont listen when your neighbors shout )).

Well, you must have seen my input in this in the related topic.

But if you put the radio in plastics and use low gain antennas all of this does not really matter

I can't agree more with you! I already sprayed my first plastic boxed internally (don't forget the top and use an earth wire to connect all part to earth and eachother!) with a double layers of metal spray paint. And from now on I buy only metal casing for all my AP's. And I always used higher gain antenna's than actually needed anyway. High gain directional gain there gain by convergence the signal and also the ´listening´ caracteristics are more directional. Thus avoiding signals you don't want.

But yes, a ´wizkidd´ experimental build 1kW transmitter will blow your cards to ashes anyway.... special in front of my 32dB dish!

We can only do what we can do and it has to stay reasonable and affordable.

The quality of this nearby signal filtering might even influence the efficiency of 802.11/nstreme/nv2.
As nstreme/nv2 do more traffic on quiet links, it might help to go back to 802.11 on some links to
reduce interference

I'm afraid I don't completely understand this. From reading and what I see happening is that it is more the other way around. My links running nv2 don't see any problems. It is my ´old´ 802.11 links and PtMP networks that seem to suffer.

Yes. Switching your nv2 links back to 802.11 will help the other nearby links.
So it might help to stay at 802.11 on links with low traffic which do not need
the power of nv2.

The quality of this nearby signal filtering might even influence the efficiency of 802.11/nstreme/nv2.
As nstreme/nv2 do more traffic on quiet links, it might help to go back to 802.11 on some links to
reduce interference

I'm afraid I don't completely understand this. From reading and what I see happening is that it is more the other way around. My links running nv2 don't see any problems. It is my ´old´ 802.11 links and PtMP networks that seem to suffer.

Yes. Switching your nv2 links back to 802.11 will help the other nearby links.
So it might help to stay at 802.11 on links with low traffic which do not need
the power of nv2.

Ah, that way, I understand. Not an option for me I'm afraid. Some provider needs to be the first in my region using tdma. It better be me.... nv2 is not only about higher throughputs. It has other more important features that are important for idle links too.

I can't agree more with you! I already sprayed my first plastic boxed internally (don't forget the top and use an earth wire to connect all part to earth and eachother!) with a double layers of metal spray paint. And from now on I buy only metal casing for all my AP's.

Just curious what antenna are you using, I would be concerned while the shielding is a good idea it could impact on the performance of the antenna as in the internal distance between the radiating element and the case now shielded (wavelength?), have you noted any signal difference on your cpe's,

I would not purchase metal case for AP's simply because metal is a very good temperature conductor which from experience with cpe’s with aluminium case bolted onto a steel antenna, at night low temps and each morning a build-up of condensation inside the box which resulted moisture droplets falling on to the radio card and eventually failure, so for me spraying the inside of a plastic AP unit overcomes this issue ?

I can't agree more with you! I already sprayed my first plastic boxed internally (don't forget the top and use an earth wire to connect all part to earth and eachother!) with a double layers of metal spray paint. And from now on I buy only metal casing for all my AP's.

Just curious what antenna are you using, I would be concerned while the shielding is a good idea it could impact on the performance of the antenna as in the internal distance between the radiating element and the case now shielded (wavelength?), have you noted any signal difference on your cpe's,

I would not purchase metal case for AP's simply because metal is a very good temperature conductor which from experience with cpe’s with aluminium case bolted onto a steel antenna, at night low temps and each morning a build-up of condensation inside the box which resulted moisture droplets falling on to the radio card and eventually failure, so for me spraying the inside of a plastic AP unit overcomes this issue ?

http://marsantennas.com/item/g_antennas_type_1-43.html
has a metal compartment. Available in 23db, dual 23db and up.
Never found condensation inside.

I can't agree more with you! I already sprayed my first plastic boxed internally (don't forget the top and use an earth wire to connect all part to earth and eachother!) with a double layers of metal spray paint. And from now on I buy only metal casing for all my AP's.

Just curious what antenna are you using

I started to use for CPE: http://www.interprojekt.com.pl/acconet- ... p-581.html
I have sprayed two and am going to test them to see if it makes any different.
Most AP's of mine have separated antenna's. Spraying the boxes is merely to keep foreign radio signals out of the routerboard containing plastic box. I have several of these and think it to be a waist to just dump them. But new boxes I buy are full metal anyway.

I would be concerned while the shielding is a good idea it could impact on the performance of the antenna as in the internal distance between the radiating element and the case now shielded (wavelength?)

hmm, don't think so. Most boxes with embedded antenna come with metal backplate on the antenna anyway. That should shield radio energy of to the back. Secondly, typical 5Ghz wavelength is somewhere in the region of 3cm's. The depth of the box is always more. I don't think you have to worry about self-interference from the radio by internal bouncing. The reason of shielding the boxes is more to keep unwanted radio signals out. They can come from other radio's nearby. Very strong and even from other channels than the one you use. They can also come from same freq. range but from distant (other user/provider) units that happen to be radiating in your antenna's direction.

I would not purchase metal case for AP's simply because metal is a very good temperature conductor which from experience with cpe’s with aluminium case bolted onto a steel antenna, at night low temps and each morning a build-up of condensation inside the box which resulted moisture droplets falling on to the radio card and eventually failure, so for me spraying the inside of a plastic AP unit overcomes this issue ?

I never had any condensation issues. But I do have heat problems at times. Special on the plastic boxes. I had to shield them with alu paper on the outside to not have them heating up too much by the sun. Metal boxes fit to steel mast/tower get rid of their temperature much easier since the tower is always cooler because it is been cooled by the wind. But ok, my advantage (cooling by the mast) can be your disadvantage (condensation because box becomes too cool in humid climate). I think a metal spray paint on your plastic boxes could give you a bit advantage of both. Lesser condensation and some radiation shielding.

hmm, don't think so. Most boxes with embedded antenna come with metal backplate on the antenna anyway. That should shield radio energy of to the back.....

Yes you are correct must check if this backplate is earthed

....I think a metal spray paint on your plastic boxes could give you a bit advantage of both. Lesser condensation and some radiation shielding.

My thoughts are now to simply mount say 433 metal case on the inside of the plastic AP box http://routerboard.com/index.php?showProduct=61, once heat build up by the radio card in this metal box is not a issue.

Hi ...

5, 10, 20, 40 MHz channel BW means 2.5, 5. 10 & 20MHz low pass filters at IQ demod outputs of an - example - zero IF front end, which I guess is used in 99% of the chipsets.

I have a pair of Maxim's develop kit around, this chipset has an option to setup (config) this filters - a programable baseband low pass filter, which can be adjusted to 7.5, 9.5, 14 or 18MHz (15, 19, 28 & 36 MHz channel bw).

But they're not "brickwall".

Depends on the chipset architeture and the hability of the IQ AD converters to handle out of band signals. So the "real" filter may be implemented after the AD converter (something like DSP) where the math can implement any filter shape at an FPGA, for example.

But this was last decade technology lol.