Can you share more about the hardware you're using for this mesh? I'm still struggling a bit with my design.
The point to point nanobridges are nice and cheap...$100 a site. But as soon as I want start to think about having multiple links in more of a hub fashion, I look at a Rocket ($90) with a sector or omni ($150). Links between omni sites look to need p2p dishes unless you put the omni sites close. Ugh. design decisions.
I won't step on Andre too much as he is doing a lot of the central planning here.
I do want to say that my major node is a Rocket+Sector and is aimed strategically to cover a very large population percentage (I'm lucky in this one case that we have a mountain to the north that means I don't need to cover the north, and I'm against the Ocean so KG6JEI-West only needs to cover a 90 degree window to cover a major portion of the population -- this is one of the few times I am thankful for the terrain -- of course the same range of hills blocks me from linking to Orange County network so I curse them at the same time)
I'm aware of only ONE omni in this network and to my knowledge it is a DUAL POLARITY omni, not a lower cost single polarity omni.
You can probably tell from the the node names that a fair number are NanoStationM2's (NSM2's)
What is actually more interesting on the topology is I would say 3/4 of these links didn't show up until moving to channel -2 @ 5MHz and that many of these links were never planned, it just happened a node pointing at another node picked up other nodes as we made the switch.
I'm aware of one link that showed up at ~10 miles with one antenna pointed 180 degrees way from the node.
Its started to restore my faith in being able to do more 'mesh' vs 'hub and spoke (though I think the network I'm working with Andre on will give us better performance by being more designed infastructure)
It shows on the screen that my KG6JEI-NSM2-1 is aimed towards Andre Sleeping Indian node, its however working my KG6JEI-Oceanside-West node practically off the back side of it as well.
I'm tempted to try either move smaller beamwidth antennas and try and shield and do more directed between both (being a relay point) or switching to a wider antenna (true sector) and see if the added gain and wider beam path works for an advantage (though the NS is much easier to keep hidden on property and less noise as more directed)
I see myself wanting to get those nanobeams working for some strategic PtP links... Just hope its soon (one of those devices that we keep dedicating time to getting to work)
Yes, it's remarkable how many additional links popped up with the move to -2 and 5 MHz. What I'm curious about is which change had the greater effect...
I'm seeing the same thing in Orange County. Before, my high mountain node station was seeing 4 to 5 nodes, now I see a dozen--on 2.4Ghz. At the same time, we moved a ch 149 (stayed in part 15) to 10Mhz BW. This definitely improved the LQ/NLQ as well--around ~20% LQ. Both are a factor. Based on the example I see here. the out-of-unlicensed space was a bigger improvement than 20Mhz->10Mhz. The 3Ghz band wins top price. I've yet to find the distance out before the LQ drops below 100%. Current testing was 15.3 miles out using a NANOSTATION M3 (120 deg Sector on other end)! I gave a demo this week at the City of Irvine to Saddleback mountain.
Now I'm wondering how many 2.4G client nodes are going to pop up in Orange County before it becomes a problem. The high level station is seeing nearly everything out ~20 miles. I'm seeing at my house a 90% LQ (one direction only) signal that is 90 deg perpendicular to my grid dish. The station is ~14 miles away on ch -2.
Yep the new problem, now were going to be 'jamming' ourselves so LQ will likely drop in that regard, but thats a problem I rather have as there is something I can do with that (shield antennas, change beam paths, etc) vs fighting non linked hardware.
Plus in theory they get dense enough we can start hopping off them (Actually doing that with one of hte paths here, it was inadvertent link, and its getting me a more reliable path from "North County" to "South County" than the first "unintended' path was.
I had one node with a massive noise floor (try -85 to -75) I had to get get away from the noise to make it to that one, channel -2 had an effect where as I can tell just from spectrum viewing that the hardware was getting clobbered by local home RF 5MHz would help, but it sure wouldn't of solved the problem of RF right on top of my node.
5MHz has been field tested in San Diego as having a major positive effect on some paths, I'm not sure however they were full line of sight (local brush at one end) but were virtually line of sight where LQ was strong one way but not the other until smaller bandwidth. Unfortunately relationships broke down at that site and the node isn't there for more testing so we never got to test changing the new "clean" (5GHz) channels on it. I am aware of one report where another station on these forums (not in SoCal) reported it was a negative to go smaller in their case.
I did Ethernet captures in the past an at times (on chan 1 @ 20) you would get maybe 1 packet every 4 minutes from some devices (or burst of like 3 packets at once) all at 'near noise floor' level which means smaller channels more compact power spectrum would in theory help by a few DB.
Getting away from other signals will always help as they are no longer there to 'jam' the channel on the receiver.
If one is out in the country channel -2 will likely have no effect, in heavy metropolitian like SoCal and Las Vegas its going to be much more significant.
There could be many more stations out there generating interference on the unlicensed channels than you can actually see individually in a spectrum survey. WiFi base stations emit beacons every 100 ms. Lots of them, individually too weak to decode, could certainly generate something that looks a lot like pure noise. One of these days I ought to put up my satellite S-band dish and downconverter so I can actually listen to the band.
I believe we see some because we see LQ in one direction will be 20% better on the node that doesn't see as many as compared to the nodes nodes that are are 'high point' central location that see everyone, though I can't really give a good quantification on this yet.
This is however predicted in the design plan of our structure.
It should be noted some of these are LONG HAUL paths which mean that their nodes are lower signal than everything local which (in theory) reduces the affect they will have as the more local nodes will have stronger power to get in, these links become somewhat 'emergency backup' links in my mind but will have to see how this evolves.
Our BIGGEST hidden terminal affect to my knowledge was actually with one group of 3 nodes where all of them had the RF Distance parameter set TOO SHORT (set to 0) that was the biggest issue, once those were changed to further than real world distance that issue cleared up (Line was very choppy on a SIP phone and LQ went through the floor and latency went through the roof) the nodes ended up flooding the RF channel from both directions and thrashed the RF locally because they thought the node was closer than it really was.
Is there a network topology displayed on a map? Would anyone have any objections to me creating one based off the topology above. In my opinion this makes it easy to identify the locations of each node and what the is the closest node to link into.
Are there smaller mesh topologies running the BBHM WRT54G Hardware/Firmware. I would think that having the backbone running on the -2 for long haul traverses and then smaller & cheaper networks running on the BBHM systems would be feasible and connected via Ethernet for local around neighbor hood setups. My current plan, based on the information I have at the moment is to tie into the Otay Mesa Node from Imperial Beach and tie that node into BBHM to connect Chula Vista and Imperial Beach for some of the Hams in the my local area. Any thoughts or suggestions?
The Linksys devices offer no advantages over Ubiquiti gear and, for example, the NanoStation M2 for the $75 street price can't be beat when you consider it's already weather-proof, tower mountable (for the cost of 2 tie wraps), includes an internal 11 dBi antenna, and is powered over the Ethernet cable. You will find, as the rest of us in San Diego have, that the Linksys gear was fun to play with in the lab, but not at all suitable for outdoor deployment without considerable expense.
If you will reply to my email from yesterday, I'll add you to the local mesh distribution list and invite you to join us for the next San Diego Mesh Working Group meeting. There's lots going on throughout the county.
I would like to say that today 5 March 2016 I rode the 2.4Ghz electromagnetic waves 25 kilometers from Imperial Beach and joined the node on Mt Otay. My Node has the callsign KK6YKC. I have a weak SNR at the moment but I hope to rectify this in the near future.
I have tried to adjust antenna no luck getting the SNR any better. I may need to upgrade to a higher gain/directional antenna. Here is the antenna I am currently using:
Most all of us are using Ubiquiti antennas with Rockets. Perhaps we should get together and test each of these antennas. You bring the Lcom and I'll bring both a NanoBridge M2 and a RocketDish M2. Maybe if you're coming to the Del Mar meeting next Monday we could set it up in advance. I wouldn't buy anything until this test convinces you of a combination that will work.
The 2G24 "rocket dish" is being discounted these days - for a point-to-point link it would be good.
It would be interesting to see a "shoot out" between the various vendors (Andrew, Laird, L-com, Ubiquiti, Ventev, Radiowaves, Til-Tek) to see if the specs they publish are real.
Hams sometimes have an antenna gain measuring activity at hamfests. Its kind of a tricky business to measure gain and directivity in the real world.
But, if you have antennas all tested on the same range, you can get a pretty good comparison between them. I am not sure they are used to MIMO antennas though ...
Of course, this does not work for antennas with integrated radios (or radios with integrated antennas) like the nano station.
I will not be able to make the meeting as I have work. But I am willing to do this another time. I have some engagements up until May then I am free. Looks like the hail just knocked my link to Mt Otay. So that is good know....lol. I am personally a fan of Laird and L-Com antennas. I am sure Ubiquiti are great also I just have never used them.
I am up and running 100%. Just installed vs 3.16 and meshchat. I have a good SNR and can access other nodes.
I am wondering if there are tools that can be installed on the nodes such as ping, tracert/traceroute etc? next project may to build a topolgy map in either google earth pro or arcgis.
I will be traveling up to Paso Robles this weekend and taking with me my mobile node. Does anyone know if there is a mesh net there?
You can simply Telnet into the node using PuTTY. You'll connect to its Linux computer where these basic network tools are native. You can also use Windows versions of the tools from your PC CMD window.
It's not particularly complete yet, but here's the map of AREDN nodes Darryl has created. It's based on the Optional Settings info you provide and upload from the Basic Setup screen: http://usercontent.arednmesh.org/K/5/K5DLQ/livemap2.html
Topology Image:
Can you share more about the hardware you're using for this mesh? I'm still struggling a bit with my design.
The point to point nanobridges are nice and cheap...$100 a site. But as soon as I want start to think about having multiple links in more of a hub fashion, I look at a Rocket ($90) with a sector or omni ($150). Links between omni sites look to need p2p dishes unless you put the omni sites close. Ugh. design decisions.
Mike
Mike, I'll forward you the San Diego design. I'm not quite ready to post it on the forum.
Andre
I won't step on Andre too much as he is doing a lot of the central planning here.
I do want to say that my major node is a Rocket+Sector and is aimed strategically to cover a very large population percentage (I'm lucky in this one case that we have a mountain to the north that means I don't need to cover the north, and I'm against the Ocean so KG6JEI-West only needs to cover a 90 degree window to cover a major portion of the population -- this is one of the few times I am thankful for the terrain -- of course the same range of hills blocks me from linking to Orange County network so I curse them at the same time)
I'm aware of only ONE omni in this network and to my knowledge it is a DUAL POLARITY omni, not a lower cost single polarity omni.
You can probably tell from the the node names that a fair number are NanoStationM2's (NSM2's)
What is actually more interesting on the topology is I would say 3/4 of these links didn't show up until moving to channel -2 @ 5MHz and that many of these links were never planned, it just happened a node pointing at another node picked up other nodes as we made the switch.
I'm aware of one link that showed up at ~10 miles with one antenna pointed 180 degrees way from the node.
Its started to restore my faith in being able to do more 'mesh' vs 'hub and spoke (though I think the network I'm working with Andre on will give us better performance by being more designed infastructure)
It shows on the screen that my KG6JEI-NSM2-1 is aimed towards Andre Sleeping Indian node, its however working my KG6JEI-Oceanside-West node practically off the back side of it as well.
I'm tempted to try either move smaller beamwidth antennas and try and shield and do more directed between both (being a relay point) or switching to a wider antenna (true sector) and see if the added gain and wider beam path works for an advantage (though the NS is much easier to keep hidden on property and less noise as more directed)
I see myself wanting to get those nanobeams working for some strategic PtP links... Just hope its soon (one of those devices that we keep dedicating time to getting to work)
Yes, it's remarkable how many additional links popped up with the move to -2 and 5 MHz. What I'm curious about is which change had the greater effect...
I'm seeing the same thing in Orange County. Before, my high mountain node station was seeing 4 to 5 nodes, now I see a dozen--on 2.4Ghz. At the same time, we moved a ch 149 (stayed in part 15) to 10Mhz BW. This definitely improved the LQ/NLQ as well--around ~20% LQ. Both are a factor. Based on the example I see here. the out-of-unlicensed space was a bigger improvement than 20Mhz->10Mhz. The 3Ghz band wins top price. I've yet to find the distance out before the LQ drops below 100%. Current testing was 15.3 miles out using a NANOSTATION M3 (120 deg Sector on other end)! I gave a demo this week at the City of Irvine to Saddleback mountain.
Now I'm wondering how many 2.4G client nodes are going to pop up in Orange County before it becomes a problem. The high level station is seeing nearly everything out ~20 miles. I'm seeing at my house a 90% LQ (one direction only) signal that is 90 deg perpendicular to my grid dish. The station is ~14 miles away on ch -2.
Joe AE6XE
Yep the new problem, now were going to be 'jamming' ourselves so LQ will likely drop in that regard, but thats a problem I rather have as there is something I can do with that (shield antennas, change beam paths, etc) vs fighting non linked hardware.
Plus in theory they get dense enough we can start hopping off them (Actually doing that with one of hte paths here, it was inadvertent link, and its getting me a more reliable path from "North County" to "South County" than the first "unintended' path was.
Depends on where one was.
I had one node with a massive noise floor (try -85 to -75) I had to get get away from the noise to make it to that one, channel -2 had an effect where as I can tell just from spectrum viewing that the hardware was getting clobbered by local home RF 5MHz would help, but it sure wouldn't of solved the problem of RF right on top of my node.
5MHz has been field tested in San Diego as having a major positive effect on some paths, I'm not sure however they were full line of sight (local brush at one end) but were virtually line of sight where LQ was strong one way but not the other until smaller bandwidth. Unfortunately relationships broke down at that site and the node isn't there for more testing so we never got to test changing the new "clean" (5GHz) channels on it. I am aware of one report where another station on these forums (not in SoCal) reported it was a negative to go smaller in their case.
I did Ethernet captures in the past an at times (on chan 1 @ 20) you would get maybe 1 packet every 4 minutes from some devices (or burst of like 3 packets at once) all at 'near noise floor' level which means smaller channels more compact power spectrum would in theory help by a few DB.
Getting away from other signals will always help as they are no longer there to 'jam' the channel on the receiver.
If one is out in the country channel -2 will likely have no effect, in heavy metropolitian like SoCal and Las Vegas its going to be much more significant.
There could be many more stations out there generating interference on the unlicensed channels than you can actually see individually in a spectrum survey. WiFi base stations emit beacons every 100 ms. Lots of them, individually too weak to decode, could certainly generate something that looks a lot like pure noise. One of these days I ought to put up my satellite S-band dish and downconverter so I can actually listen to the band.
With so many nodes on one frequency, do you folks see any hidden terminal effects?
73, Mark, N2MH
I believe we see some because we see LQ in one direction will be 20% better on the node that doesn't see as many as compared to the nodes nodes that are are 'high point' central location that see everyone, though I can't really give a good quantification on this yet.
This is however predicted in the design plan of our structure.
It should be noted some of these are LONG HAUL paths which mean that their nodes are lower signal than everything local which (in theory) reduces the affect they will have as the more local nodes will have stronger power to get in, these links become somewhat 'emergency backup' links in my mind but will have to see how this evolves.
Our BIGGEST hidden terminal affect to my knowledge was actually with one group of 3 nodes where all of them had the RF Distance parameter set TOO SHORT (set to 0) that was the biggest issue, once those were changed to further than real world distance that issue cleared up (Line was very choppy on a SIP phone and LQ went through the floor and latency went through the roof) the nodes ended up flooding the RF channel from both directions and thrashed the RF locally because they thought the node was closer than it really was.
Hello to the group first. My name is Gary KK6YKC.
Is there a network topology displayed on a map? Would anyone have any objections to me creating one based off the topology above. In my opinion this makes it easy to identify the locations of each node and what the is the closest node to link into.
Are there smaller mesh topologies running the BBHM WRT54G Hardware/Firmware. I would think that having the backbone running on the -2 for long haul traverses and then smaller & cheaper networks running on the BBHM systems would be feasible and connected via Ethernet for local around neighbor hood setups. My current plan, based on the information I have at the moment is to tie into the Otay Mesa Node from Imperial Beach and tie that node into BBHM to connect Chula Vista and Imperial Beach for some of the Hams in the my local area. Any thoughts or suggestions?
73's
Gary
KK6YKC
Hi Gary,
The Linksys devices offer no advantages over Ubiquiti gear and, for example, the NanoStation M2 for the $75 street price can't be beat when you consider it's already weather-proof, tower mountable (for the cost of 2 tie wraps), includes an internal 11 dBi antenna, and is powered over the Ethernet cable. You will find, as the rest of us in San Diego have, that the Linksys gear was fun to play with in the lab, but not at all suitable for outdoor deployment without considerable expense.
If you will reply to my email from yesterday, I'll add you to the local mesh distribution list and invite you to join us for the next San Diego Mesh Working Group meeting. There's lots going on throughout the county.
Andre, K6AH
Hello AREDN,
I would like to say that today 5 March 2016 I rode the 2.4Ghz electromagnetic waves 25 kilometers from Imperial Beach and joined the node on Mt Otay. My Node has the callsign KK6YKC. I have a weak SNR at the moment but I hope to rectify this in the near future.
73's
Gary
.KK6YKC
I have tried to adjust antenna no luck getting the SNR any better. I may need to upgrade to a higher gain/directional antenna. Here is the antenna I am currently using:
2.4 GHz 11 dBi Dual Polarization Diversity/MIMO/802.11n Antenna - 3ft RP-TNC Plug Connector
Tyring to find somethingh like this http://www.l-com.com/wireless-antenna-24-ghz-18-dbi-dual-polarity-mimo-d...
What is the community using with a rocket m2?
Also I have used LNA,s in the past and have access to some. Is this recommended?
I found a Nano Bridge M2 are those compatible with the firmware the chart does not have the Nano Bridge M2 but the 2G18.
Also I see some folks have services where can i find a list of them?
73's
Gary
KK6YKC
Most all of us are using Ubiquiti antennas with Rockets. Perhaps we should get together and test each of these antennas. You bring the Lcom and I'll bring both a NanoBridge M2 and a RocketDish M2. Maybe if you're coming to the Del Mar meeting next Monday we could set it up in advance. I wouldn't buy anything until this test convinces you of a combination that will work.
Andre
The 2G24 "rocket dish" is being discounted these days - for a point-to-point link it would be good.
It would be interesting to see a "shoot out" between the various vendors (Andrew, Laird, L-com, Ubiquiti, Ventev, Radiowaves, Til-Tek) to see if the specs they publish are real.
Hams sometimes have an antenna gain measuring activity at hamfests. Its kind of a tricky business to measure gain and directivity in the real world.
But, if you have antennas all tested on the same range, you can get a pretty good comparison between them. I am not sure they are used to MIMO antennas though ...
Of course, this does not work for antennas with integrated radios (or radios with integrated antennas) like the nano station.
I will not be able to make the meeting as I have work. But I am willing to do this another time. I have some engagements up until May then I am free. Looks like the hail just knocked my link to Mt Otay. So that is good know....lol. I am personally a fan of Laird and L-Com antennas. I am sure Ubiquiti are great also I just have never used them.
San Diego mesh,
I am up and running 100%. Just installed vs 3.16 and meshchat. I have a good SNR and can access other nodes.
I am wondering if there are tools that can be installed on the nodes such as ping, tracert/traceroute etc? next project may to build a topolgy map in either google earth pro or arcgis.
I will be traveling up to Paso Robles this weekend and taking with me my mobile node. Does anyone know if there is a mesh net there?
Gary
KK6YKC
You can simply Telnet into the node using PuTTY. You'll connect to its Linux computer where these basic network tools are native. You can also use Windows versions of the tools from your PC CMD window.
Andre
Thanks for the information. Is SSH supported also?
Gary
It's not particularly complete yet, but here's the map of AREDN nodes Darryl has created. It's based on the Optional Settings info you provide and upload from the Basic Setup screen: http://usercontent.arednmesh.org/K/5/K5DLQ/livemap2.html