12-20-2018 12:02 PM - edited 12-20-2018 12:03 PM
I have a variety of questions about short link I have setup between two NSM2 Nanostations.
Before I ask them, here's some details on the configuration.
1. The link itself is about 500 feet through trees.
2. Both Nanostations are outside, and have RF armor.
3. The link is used to connect my Mom's house to mine, for internet.
For technical details on the link, see the screenshots below:
Below is the AP side (which is at my house). Please note for reference these screenshots were taking when it was raining outside and overcast.
Screen shot of the station (on my Mom's house)
Airview that was run from my AP (at my house). Note this was not taken the same day as the above screenshots from the Nanostations (the weather was clear when this was done)
Here's my questions.
1. Why the heck is my noise floor so high? I live in a rural area in the mountains of northeast georgia. I would expect my noise floor on 2.4ghz to be much lower living in such a remote area. I did recently realize that the reciever sensitivity for the nanostations is around -83dbm, is this why my noise floor is sitting there? It would make sense that the noise floor can't be any lower than the receiver sensitivity. If I move to a 20mhz channel, my noise floor goes into the high -70s.
2. I have a real problem with the horizontal polarization being severely attenuated (especially when it rains). Some of the trees I have to shoot through are pine (I can provide pictures of the actual trees/link, if requested), and I've heard that's worse for horizontal polarization, but it makes no sense to me? This is very confusing to me, because from everything I can find online about polarization orientation is the horziontal is usually better for shooting through trees. Vertical polarization signal is always pretty strong. At first I thought this issue was due to the metal roof on my house, and a few other structures that the link goes near. My link budget is pretty high on this link though, cause when it rains I loose almost 7-10dbm on both polarizations. I'm wondering if a -/+ 45 (aka slant polarization) antenna would really help with this.
Overall, I'm really happy with the nanostations. I've learned a ton using them about RF, etc. When the weather is clear, I have 99% CSQ typically, and I can get 45mbps of throughput on a 10mhz channel between the two nanostations. I'm thinking of switching the AP side out on my house for a Rocket 2AC prism with a sector antenna. I'm looking to add some more stations for IP cameras on my property (I have a big property) and want to provide better coverage than I think the nanostation can do. It seems like a higher gain antenna would help with shooting through all these trees, and providing a slighty more stable link? I'd like to run a 20mhz wide channel when I get all of these IP cameras online, so I don't chew up too much bandwith with them to interrupt other traffic on the link.
Any input and thoughts into this, would be greatly appreciated.
01-04-2019 09:23 AM
Though I'd post a follow up, as I've made some changes that made a big difference (and for others to see the results, and be able to learn from this).
So I turned off everything that is 2.4ghz in my house, and at my Mom's house, and did a airview and got this:
Around 2430mhz I noticed there was like barely any hits, so I decided to change the channel to that, and try lowering the bandwith from 10mhz to 5mhz. When I did that the chains suddenly are much closer, and the noise floor went much lower (which is expected).
Below is a screenshot of AirOS on the AP side:
The station side AirOS screenshot:
so a couple of thoughts.
1. I left it at 5mhz wide, cause I can get 20-22mbps with this, and its adequate for this link.
2. I have no idea why my horizontal polarization was so severely atenuated on the higher channels, but I suspect I was getting some interference (my AP is on my house and pretty high up compared to the surrounding area, and there's no leaves on the trees right now, so I suspect any one else transmitting on 2.4ghz is going pretty far). I don't know if Airview only shows one polarization, but maybe someone could clarify that? It also dawned on me, I'm shooting through the power lines that run to my house, and they are horizontal orientation, so I wonder if that is related? Maybe a harmonic or something? Everything I could find online indicated it shouldn't be an issue, but I have doubts now.
3. The link is very stable in this configuration, even in the rain, so I suspect all my issues with a low CSQ (re-transmitts) was due to something around channel 11 on the horizontal polarization.
01-04-2019 09:50 AM
One other thing that just dawned on me, is that I have some 1.9ghz 17dbi yagi antennas mounted fairly close to the AP nanostation for an LTE modem setup. I highly doubt this is causing issues, but they are both horizontally mounted like pictured below:
I know for a fact that the LTE antennas recieve centered on 1940mhz and transmit on 1865mhz (band 2, for those who know LTE speak). I've confirmed this with my SDR, unfortunately my SDR does not go above 2ghz (gonna fix that soon and get a different one). Anyways, I highly doubt this LTE modem is emitting spurrious transmissions that high above its designed transmit range, but I suppose its possible. I installed a high pass filter that blocked 0-600mhz on these LTE antennas, but I maybe should get a band pass filter for just band 2 in the future. More food for thought....
01-07-2019 10:14 AM
01-11-2019 12:15 PM - edited 01-11-2019 12:16 PM
i've noticed something interesting in the last month, but seen it a couple times this week. It appears that when the temperatures get below freezing the RSSI in both directions between the nanostations gets a good bit stronger 3 to 5dbm, plus the chains are very well balanced.
I also have noticed this with my LTE connection, which is currently only working at 1.9ghz (LTE band 2) which is close to my 2.4ghz nanostations in frequency.
Doing some quick googling, I found a research paper on the effects of temperature on signal levels specifically in the 2.4ghz band:
The research paper really doesn't say why this happens, but I suspect its due to moisture molecules slowing down at freezing temps or below, and causing less reflections/multipathing.
Below is a screenshot of my signal stats early this morning, when it was around 25 degrees farheinheight.
01-11-2019 12:51 PM
01-11-2019 01:05 PM
oh, interesting. I think I've heard of this, but haven't read much up on it yet.
I don't think this would be happening between my two nanostations (as they are only about 500 ft apart and the AP is maybe 50ft higher than the lower one), but I could see this happening on my LTE link, as that is 5 miles NLOS over a few mountain ridges.
01-20-2019 06:08 PM
I finally figured out what was happening.
TL;DR: This was interference being picked up from some neighbors 2.4ghz APs on Channel 1 and 11.
Quick summary of the link again:
500ft through trees.
The AP (on my house) is about 50-75ft higher than the station on my Mom's house. My home is high up on the side of a hill going up to the top of a mountain. The AP is facing down into this valley, the station is facing towards the mountain as it goes up.
A couple of days ago, I noticed again that the Transmit CCQ from the AP and the station were both flucating wildy. The weather was perfectly clear, no wind or rain. I kept suspecting interference, but my neighbors are at least 500+ ft away, most of them probably more like 1,000+ ft away. I've obviously done quite a few airviews, but I wasn't sure what the hits were that I was randomly picking up. I realized I could do a site survey to maybe see from the AP (or station for that matter) if there was some nearby networks being picked up. I had tried in the past doing some site surveys, but didn't see anything. I recalled seeing on the forums someone mention that in order to see SSIDs of standard APs (Which are at 20mhz wide channel widths) you have to have the station set at that channel width, so I decided to change to 20mhz wide channel, so I could do a site survey on the AP, and just see if I could see anything while this was happening...... and BAM (interesting tidbit here is, my cell phone could not pickup this network, because its antenna is so weak compared to the Nanostation's).
FINALLY some validation to what I was seeing. These two SSIDs are modem/router/wifi APs for Windstream customers in my area. The signals are clearly pretty weak, but I was running on channel 1. The interesting thing here is, I didn't pick up my Mom's AP inside of her house. I think this is because I turned her APs transmit power down to 10dbm, cause her house is so small, and everything was geting blown away. These windstream routers, must sure be blasting out 2.4ghz on some high power. I quickly changed my AP to be centered on channel 6, and changed my channel width back to 5mhz. Now, I see the following:
1. Transmit CCQ stays at 99% or above literally all the time (even had some heavy rain and wind the last day, and it bounced around a little bit, but didn't see drop below 90% and the modulation rates stayed pretty high too.) Before, it would stay above 99%, but then at seemly random times, it would start bouncing all around. I think this is when my neighbor was using their wifi network, and hence why it didn't occur all the time (or maybe the interferring signal was coming in better to the nanostation at certain times?)
2. The chains are actually balanced now, and they stay that way! The interesting thing here is, I think the neighbors APs must be transmitting on a horizontal polarization, cause that was the most effected polarization on my nanostations, and now its totally normal:
Also thought I'd share the site survey from the station (at my mom's house)
Top SSID is my AP, but interstingly, the main SSID and guest SSID (bottom two in the list) have a pretty strong signal in the site survey. This Nanostation has RFarmor on (and the AP is behind it, so the armor should be blocking the signal pretty good) and the AP inside has its transmit power only at 10dbm.
I think this interference issue didn't happen during non-winter months due to all the leaves/vegetation blocking the neighbors signals from their APs. Anyways, I'm glad to finally have figured out what the heck was going on, and wanted to share here from others to learn from my experiences.