Can you help me interpret this spectrum analysis noise plot?

Right. That's the symptom. Now, what problem are you trying to solve? Lack of speed? Variations in speed? VoIP jitter? Disconnects?

Why is your WISP using Ch 10. Is there a problem with Ch 11 or does he have his AP setup for the dreaded "auto"?

If you look at the lower graph, in the area of Ch 10, you'll see that the baseline "noise" level for Ch 10 is about -110dBm.

Unplug your antenna and see what manner of "noise" level you get. The -88dBm "noise" level is what the Rocket M2 hears between bursts of RF, when there's presumably nothing to hear. With an all digital receiver, the Rocket M2 is going to hear about 25 MHz or more of RF spectrum. I put "noise" in quotes because it's not really noise in the sense of measuring receiver NF (-174dBm/Hz). My guess(tm) is that it's mostly other users on Ch 10. That could be other client radios, such as your neighbors talking to the same WISP on Ch 10 or the 2-3 adjacent channels. It could also be additional WISP access points on Ch 10. I know you have a rather directional antenna, which is a big help at removing such "noise", but if you have the bad luck of having another client radio in the direction of the main lobe, or nearby to the side, you'll hear their "noise".

Duh. I didn't look at the pictures very carefully. Sorry.

Looks like you have interference from someone nearby using Ch 9. I marked the area on a copy of your screendump in red. I also took the liberty of tweaking the contrast so that the Ch 9 junk can be more easily seen: There's a 20dB signal level difference between your WISP and the interference, but that's probably enough to cause problems. It also explains the -88dBm "noise" level.

If you post any more images, please do NOT stretch the image. Just leave the aspect ratio the same as the original.

I found a good description of the three (actually four) plots here:

The Waterfall chart is a time-based graph showing the aggregate energy collected over time for each frequency, where colder colors indicate lower energy levels & warmer colors mean higher energy levels at that frequency bin.

My Waterfall chart shows relatively cool colors (low energy) over all the bands except those surrounding 802.,11n channel 2 & 10, which are the two channels of the WISP antennas facing me:

The Channel Usage chart, each Wi-Fi channel is represented by a bar displaying a percentage showing the relative "crowdedness" of that specific channel. This percentage is calculated based on both the popularity and the strength of RF energy in that channel.

My Channel Usage chart shows that the channel usage is between

5% and 10% where I live. I don't know if that is a high or low number, but, it *seems* low, simply based on percentages.

In my case, the channel usage is again highest in the channels that the WISP has an antenna aimed at me, namely channel 2 and

10; but the chart also shows relatively high usage in channel 1 & 3, and in 8, 9, & 11. The lowest usage is in channels 4, 5, 6, and 7, so, I guess that's where I'd put my home broadband

The middle chart in both those screenshots above is the Waveform chart. Just like the Waterfall chart, this time-based graph shows aggregate energy where the color of the energy indicates amplitude.

It also shows the signal strength humping around WiFi channels

2 and 10, which are the two antennas from my WISP facing me. Oddly, it also shows red areas in the middle bands, which makes no sense to me unless they're instantaneous and short lived bursts of energy (perhaps from cellphones or microwaves?).

The bottom chart in both those screenshots is the Real-time chart which shows what a traditional Spectrum Analyzer would display in energy (in dBm) as a function of frequency.

The blue trace is called "Max Hold" of maximum power levels across the frequency. The green trace is the average energy across frequency. And the yellow line is the real-time energy.

That's aptly shown by the fact that my WISP has two antennas facing me, one at channel 2 and the other (which is my "signal") on channel 10.

So, his channel 2 is creating noise for me, based on these spectrum analysis results yesterday.

In the middle chart, you can see high signal surrounding channels 2 & 10. So, if I select channel 10, channel 2 is noise to me.

I see what you mean.

I'm only looking at external noise (AFAIK) with these spectrum analysis runs.

These plots are all taken in a direction looking *away* from the city of San Jose. I'm pretty sure I'd get quite different results had I pointed directly at the Fairmont Plaza building!

Hi Jeff, Thanks for chiming in, as I know you're a God in the wireless forums, and, your information is especially useful since we're both on each side of the Santa Cruz mountains.

My speeds are actually OK, as shown in this speedtest, which shows 10ms, 19Mbps down, 18Mbps up, even though the antenna system spans over 20 miles until it gets to a bona-fide wire:

I just want to make it faster. Whatever I do to decrease noise makes the signal that much better. Notice, at 56% and 66%, my signal quality figures aren't close to maxing out yet:

Although, sometimes they get up into the 86% to 90% range:

Hmmm.... I don't have a clue as to the answer. All I know is that he had texted me how to set up my radio, and I set it up just like he had texted me to. I just looked at the settings in the radio, and I see the channel width set to 20MHz and channel shifting disabled, but I don't see where the channel is actually set.

Hmmm... I see the green "average" area to be around -90 at channel 10. Now I don't know if that's a good average or not, but, I'm beginning to think it's actually low signal strength, even though there's a radio aimed right at me on channel 10.

That's an interesting test to run. The radio & antenna are on the roof, so, I will have to climb up there to unplug it, and then climb back down, and then back up. So it will take a while. (Luckily it's something like 70 degrees outside today!)

That's interesting. It makes sense.

I think all the neighbors are on the same channel as the WISP radio, so, from that, I can see there will always be noise on that channel (since we're all using it).

Thanks Jeff for that insight. There are a couple of houses right under my signal, but a few hundred feet lower in elevation than I am. Perhaps some of their signal is bouncing to me. There are also a few to the side, and behind me, and above me, so, they're going to be the most problematic, I would think.

Thanks for the insight. I think everyone should post what they see from their radios so that we can all compare to each other.

Lady friend means untrash house. You should have thought about this :-).

Hi Jeff, Thanks for taking a closer look. I would *never* have been able to see that channel 9 interference in that chart unless you had outlined it as you did in your screenshot below:

A site survey from that same radio doesn't show anyone on channel

9, but, of course, it can be coming from portable devices & microwaves:

To see if that channel 9 pollution was coming from the house, I downloaded InsSSIDer freeware for Windows & Android:

Roaming the house, I could easily see my channel 1 home broadband router polluting channels 1 & 3; and I could see that 4 through

8 had "something" on them at much lower levels; but 9 seemed clear:

The caveat here is that the radio measures things from miles away while the Inssider program only measures what the absolutely puny antenna and crippled radio on the laptop (or Android phone) can see - which is miniscule.

So, all I can tell from that is that the channel 9 interference isn't coming from *my* house! :)

PS: I didn't stretch any of the screenshots, so, if they're stretched, Flickr did it, not me. :)

Nope. It was my idea. I spent my first 50 years or so accumulating all this junk. I'll probably spend another 50 years or so getting rid of it. Do I really need 5 bicycles? I'm making weekly runs to the recyclers, thrift shop, or municipal dump to get rid of the stuff. (I really miss my pickup truck). At my present rate, I should have the house and office cleaned up in about 50 years.

my friend just retired and moved from a huge custom house in the cold, snowy, Chicago area down to the middle of FL - sitting out on the lanai....

He coined the term ---> de-crapification

Just prior to his leaving, we had our little college ski group lunch gathering (4 guys) with him presenting us each with a token from our past as part of his de-crapification. Some were FORTRAN puch cards, from our CS days, some were check/receipts from our skiiing in Austria, some were photos from frat parties - with fuzzy hair and long sideburns.

Or at least clean underwear.

You can by RF shield spray at Frys. Not cheap, well at least compared to a can of spray paint.

Most plastic cased RF devices either have a little metal of some sort mixed into the plastic or are coated on the inside with RF shielded paint.

You would have to replace the antenna with a 50 ohm terminator to see the actual noise floor of the receiver.

There is probably a little black body radiation at 2.GHz, but the gain of these wifi receivers is not like a C-band feedhorn. [Since black body radiation has an exponential tail, it is technically everywhere.] You can point a C-band LNBF at the ground and see the noise floor rise. Some of the instruction manuals will warn about hiding dishes with dirt berms just because of the noise.

But channel 2 is out of band of channel 10. The only issue with out of band noise is the "desense" of the receiver due to the noise floor effecting the AGC (presuming it has one). The AGC generally sits at the front of the receive chain, potentially with little or poor filtering. [There is a trade off between the loss of the filter versus the loss of the AGC gain, so the filter is usually very simple if made from lumped compoents to reduce the loss, or they go balls out with cavity filters to get low loss.]

I don't really like the wiki on radio desense. Even weak signal do a little desense due to the effect on the AGC, but you really notice it with strong signals, so this wiki is kind of correct.

You should replace the antenna with a 50 ohm terminator.

Agreed. That's easy if there's an antenna connector available. Not so easy if the radio and antenna are conglomerated into a single assembly. Wrapping the antenna in aluminum foil is going to mangle the antenna impedance. Wrapping it in conductive foam might work better.

Your equipment is the client, so you don't get to set the channel. The WISP sets the channel and all you can do is follow their lead.

I beg to differ somewhat. The selection of frequency for desense testing is suppose to NOT be anything that makes it through the IF system or to the detector. With detector activated AGC, the AGC should never be active when testing for desense. However, in the real world, massive amount of RF shoved into the antenna port might leak around the IF and make its way to the detector, thus providing some desense from AGC action.

Thanks for clarifying that. I had poured over every page of the router setup, and I didn't see WHERE the channel was set.

I did see the channel width, and modulation, but, not the channel itself.

You explained why that is.

Thanks!

What I love about this WiFi gear is that we can connect to an access point five miles away, and it works as well as if the access point was in the next bedroom over!