OK, we’re finally here. The NetSpot Site Survey is done. There is a bunch to unpack here. The pictures below show the results. First, my WiFi signal isn’t terribly good on the Signal to Noise front. For the uninitiated, there are a few weird concepts in WiFi that bear explanation, which I will try to explain so I understand them better myself. Please note that I am an enthusiastic layman, and not an expert in radio wave signal propagation, so I hope I get this right.

Results of the NetSpot Site Survey

Signal-To-Noise

This concept, abbreviated SNR, is, like it sounds, the ratio of signal to noise. The problem with WiFi is that we think of WiFi as a bright shining signal, but it’s hard to think of the environment it’s in.

Imagine you are in front of a waterfall, talking to a friend. The waterfall is the noise, and the speaking is the signal. You can easily imagine that you have a hard time distinguishing what your friend is saying from the background noise. WiFi is more like being at a party where lots of people are talking. Near talking is louder, and far talking is softer, and people with similar voices mask each other. WiFi is further complicated because the channels overlap, making WiFi design in any congested environment a bit of a crapshoot.

My survey of the noise in my attic looks like this:

Pretty bad. The only place that does not have all red noise readings is the basement. That, however, is a mixed blessing, since the same dirt and material that blocks noise blocks the signal, too.

SNRs are expressed in decibels, so let’s explore that.

Decibels

Any time you are working in waves (voice, radio, etc) you see the notation in decibels. It basically is trying to relate power to perceived power. The number (20dB) is relative, and basically shows the number of zeros you add to show the increase in power (divided by 10 since it’s a ‘deci’ scale).

So if you are in a quiet background, say at 40dB, and it becomes 60dB, then it’s a 20dB difference, or 100 times louder. If you’re in a loud place, say 80dB, and it becomes 100dB, it’s also 100 times louder.

The trick is that your ear only perceives a 20dB as twice as loud, but it’s 100 times more powerful.

Signal Strength

I’ve always been baffled when I see that my signal strength is -39dBm. I’ll try and put it into words, since all tools measure signal strength this way.

Basically (and I’m leaning on my good friend, wikipedia), the number is the strength of the signal, in dBm (deciBel-miliwatts). The reason that the number is negative is that it’s the logical result of the relationship above: as signals get weaker, the exponent goes negative:

A power difference of:

• 100 means a difference of 20dB
• 10 means a difference 0f 10dB
• 1 means a difference of 1dB
• .1 means a difference -10dB

And that’s why it’s negative.  Basically, this -39dBm means that you’re dealing with very weak signals (-40dBm is 100 nanoWatts),  and each increment you go up is a reasonable amount of power. Here at my desk, my attic access point is at -38dBm, and my farthest one is at -77dBm, meaning that the signal is 1000 times less powerful, but what the antenna ‘hears’ is around 4 times less ‘loud’.

The dB component is that it’s a relative, exponential relationship, and the ‘m’ means we’re dealing in milliWatts.

What does it all mean?

When you put it all together, you see that there is a lot of noise, but my Access points are able to cut through it. Not super-well, but pretty well. Our houses are full of electronic noise. In my attic I have a fan motor running, AC, lots of passive electronics, lights, etc.

I tried to turn off a lot of it, and it didn’t move the noise level.

So when we look at this map, you see that the signal levels are a good ways above noise. My competitive nature makes it so that when I see a scale like this I just want to get better, so that’s what I interpret here. We’ll see if it’s possible.

What makes WiFi work well?

WiFi, as far as I can tell, likes two things: Line-of-sight and no other WiFi. Also, the orientation of the antennas matter. Ubiquiti has a whole web site dedicated to this – which is the coolest thing I have ever seen. Long story short: the radios don’t radiate backward.

I think that putting my AP on the floor of my attic may not have been a wise choice given that I want the signal on the floor below it.

What’s next?

I feel that I have my arms around the WiFi in my house, and a good baseline to see if I can improve it. All my equipment is several years old, so I think that it’s ready for the upgrade. Next I’ll install the Unifi controller software on my Centos Server, then we’ll start putting in APs.

My basic plan is to start by swapping out one AP at a time, seeing the effect, then once that’s done, I’ll start to tinker with new SSIDs and such.

What I’m listening to as I do this:

Trevor Moore’s music: Notably The Story of Our Times. It’s all pretty funny stuff.