Please help me interpret noise in the Santa Cruz mountains (roughly -75dBm across the 2.4GHz spectrum)

Something is wrong. The signal levels are much too low. See sample photos at:

You might try pointing it at a wireless router some place nearby. If it doesn't go nearly full scale, you have a sensitivity problem.

Also, you might try to drag your Nanostation over to the neighbor with a similar system, and point it at the same access point. That way, the signal levels should be the same. Then, compare AirView screen.

I just downloaded Airview 1.0.11 for my XP desktop. After installing, it won't start. I'll troubleshoot later.

you do know that the 2.4GHz band is the ISM band and is used for microwave ovens among other things besides WiFi?

Mark

Stopping is a good idea, but I don't think those Simpson Strong Ties are THAT strong. Your supposed to think WTF and give up, but the poster has dulled two bits! I think he is hitting something else.

BTW, would an antenna using a dish have significant lobes?

Yeah, but what else? Drilling through a nail stopper is difficult but not impossible. However, someone mentioned that all the metal hardware is on the inside of the framing, which might eliminate most of the framing hardware. He might have hit a bolt head used to secure the collar ties. Maybe earthquake retrofit hardware. Maybe someone decided to use steel beams. The house is new enough that it should be detailed on the plans. In any case, whatever he's drilling is probably not going to be suiltable for attaching an antenna, unless he goes all the way through the wall with a through bolt.

It depends on the disk. If the feed projects above the edge of the dish, as in a nearly flat dish, then there's going to be plenty of side lobes. If the feed is buried under the edge of the dish, as in a deep dish, then there will be substantially fewer sidelobes.

Looking at the photos of the NanoBridge M2, the feed projects well beyond the dish edge, so it's going to have some side lobes. How bad is difficult to guess without a proper model or a test in an RF anechoic chamber. The data sheet shows a few at -15dB, but nothing really serious. As usual, I don't believe the data sheet.

Wow. Some of those noise levels are huge by way of comparison! And the delta between average and peak were humongous on some! For example:

I also noticed the 5GHz noise levels were lower than the 2.4 Ghz graphs so I only looked at the 2.4 Ghz ones for comparison.

I'd have to point it straight down to reach the wireless router in the center of the house. :)

That's probably the best idea yet. There is one problem though, as he has his mounted on the top of a redwood - which is a pain to climb.

I didn't realize the tool works on the laptop. In my case, it's part of the software for the radio.

Wow. Some of those noise levels are huge by way of comparison! And the delta between average and peak were humongous on some! For example:

I also noticed the 5GHz noise levels were lower than the 2.4 Ghz graphs so I only looked at the 2.4 Ghz ones for comparison.

I'd have to point it straight down to reach the wireless router in the center of the house. :)

That's probably the best idea yet. There is one problem though, as he has his mounted on the top of a redwood - which is a pain to climb.

I didn't realize the tool works on the laptop. In my case, it's part of the software for the radio.

That's exactly why we're looking at the noise levels.

One theory was that there was industrial, scientific and medical (ISM) band interference.

Four holes! Drilled about 4 inches apart up/down and about 6 or 7 inches apart side-to-side.

I can't imagine what it is - but there is no way my drill bits were making 'any' headway. I was on a ladder, so I couldn't use 'all' my strength, but I've drilled enough to know that I was hitting hard steel.

I didn't even see any metal chips coming out, although they could have been lost in the inch of stucco as I was pulling the bit out.

At the moment, I can only think that there is some kind of earthquake tie of some sort since I'm very near the fault line.

Looking at the Ubiquiti Nanobrdige M2-18 datasheet at:

The lobes don't look too bad:

Must be fun to watch the signal level as the redwood sways, especially near the top. I did that many years ago with a 19dBi dish. I couldn't keep a stable signal. We got lucky and found a "truncated" redwood, with a clear view. It was effectively half a redwood, but it was stable.

I use a telescoping window washing pole for antenna tests. Some ropes on top to 3 cinder blocks is usually sufficient for a temporary mast. Aiming it is a problem and does take some practice and profanity.

I've never tried it before. So, I downloaded the XP flavor and immediately found that it doesn't work with the current Java 7.5 version. So, I installed Java 6.0.33, which worked. However, the program complained that it didn't find the copy protection dongle errr... USB wireless device, that Ubiquiti sells for $40. Since I have a Wi-Spy, I really don't need yet another spectrum analyzer, especially when the web page says it's "end-o-life".

I was going to grind some numbers converting 1mw/cm^2 at 5cm max radiation leakage into what you would expect to see at various distances, but got sidetracked. Maybe later.

However, interference is not your problem. From the spectrum display, your receiver is comatose and can't hear either signal or interference. Got a warranty?

Pick an interference source, any interference source:

I can add about 4 more sources that I've found recently. However, the ones that I keep running into that are giving me the most trouble are wireless streaming video, wireless security cameras, and wireless point to point telemetry.

Hi Jeff,

Here's how my friend Simon explained it to me today ... so I hope to have more data this weekend.

... ... ... ... cut here ... ... ... ...

If you look at the antenna profile for the NanoBridge M2, it shows you how much loss to expect if you are a few degrees off target. If you are 15 degrees off-target, you lose about 7 dB.

15 degrees is a whole lot. The sun moves 15 degrees in an hour. Your thumb and fist (like a hitch-hiker's signal) at arm's length is 15 degrees.

At 5 degrees off-target, you lose 2dB. That's three fingers at arm's length. That is still quite a lot of miss. The entire mountain is 5 degrees across when looking at it from my house.

The sun and moon are each a half of a degree across. If you are off- target by the apparent width of the moon, or even twice that, you won't see a single dB of loss. The entire property where the antenna lies is about a degree wide as viewed from my house.

In other words, aiming is not your problem.

We see random fluctuations in signal strength that dwarf anything like a

5 degree aiming error.

I have the tripod, and a bunch of brand-new NanoBridges. You can ride along with me and we can set it up at different places along the road and see what kind of signal and speeds we get, with and without the power lines and towers in the way.

I'll be bringing it to Joe's house and Bill's house today or tomorrow, and you can help set them up if you like, and we can get measurements from those places as well.

I've yet to see anyone question the pattern/lobes/nulls of the _transmitter_.

Jonesy

It's a bit more complexicated than that. Just because you can see the other end, doesn't mean that a 2.4GHz link will be reliable. Any objects inside the Fresnel zone will cause problems. Reflections off the ground will cause nulls and peaks. Moving objects (cars) close to the path can cause fades. I use Radio-Mobile software:

to produce path profiles like this:

This one is at 420MHz and has some errors in the numbers, but gives a tolerable idea of what can be done. All you need to do is supply some numbers.

A simpler version is just the path loss.

that simply calculates the fade margin, which directly translates to reliability and uptime. I'm fairly sure you're close enough that there are no path loss issues. I'm not so sure about obstructions along the path.

6dB loss is the same as cutting your range in half at the same signal levels.

Another way to think about it is how much of your dish antenna can you cover in order to produce a -6dB loss? Well, covering half will be

-3dB loss. Another half will be -6dB. So, how well does your dish work with a dish only 1/4th the size?

Probably true. Something else is wrong which might be indicative of defective equipment. The easy way to determine that is to compare with a similar piece of hardware.

Did you ever settle on vertical or horizontal polarization?

Random fluctuations as in fades and nulls is exactly what you get when there are multiple reflections and objects in the Fresnel zone.

Great idea. Do it.

Even better. Remove your Nanobridge and bring it along. Compare signal levels.

Good point. That may be the problem. I run into it all too often. The WISP decides to use a very high gain omnidirectional antenna. Something like 15dBi with an 8 degree vertical beam width. If your house is at the exact same altitude as the WISP antenna, everything works just great. However, if the house is at the same distance, but at a different altitude, the house will be outside the antenna pattern and you'll get little or no signal as most of it will be going overhead.

Even worse, end fed colinear antennas tend to have some uptilt in the pattern. A 15dBi omni (which incidentally is about 8ft long), might have ALL the RF above the horizon if nothing is done about the uptilt. If the house is below this omni antenna, you'll get very little signal.

Ease of downtilting is one reason why WISP's use panel or sector antennas with patterns like:

120 degree horizontal beamwidth. 8 degree vertical. 14dBi gain. Three of these antennas, wrapped around a pole, pointed 8 degrees downward, put all of the signal on the ground in a wide pattern, with very little being lost sending RF above the horizon. Note the adjustable downtilt:

Hint: If it doesn't work on paper, it's not going to work when installed.

It isn't the signal strength as much as the waveform fidelity. If you are off a bit, you get more reflections of the transmitted signal.

When I use a rather fat beamed antenna, I move it from side to side and note where the signal begins to drop. then pick a spot in the middle. This is something like direction finding, where null are easier to detect than peaks. In this case you aren't using the null, but rather you reach the edge of the main lobe and the signal falls off rapidly.