Signal quality criteria for WiMax networks

In the article “Testing Wi-Fi antennas for receiving WiMax”, I cited the CINR and RSSI parameters as the results of my experiments, only approximately imagining what these values ​​reflect. I was ashamed to be in ignorance, and I understood this question in detail.
I have no doubt that there will be Habralyuds (such as shogunkub ), for whom this topic will not reveal anything new, but I am sure that many people will be just as interesting to find out, just like me.



On Habré, this issue was addressed in the article "WiMAX Base Station" , but I would like to highlight it a bit more fully.

RSSI (Received Signal Strength Indicator) - literally: an indicator of the strength of the received signal. The power in this abbreviation should be understood as the power of a radio signal.
In the IEEE 802.11 standard, RSSI is an indicator of the power level received by an antenna, expressed in “parrots” - abstract units chosen by the manufacturer of wireless equipment.
The physical meaning is carried by units of mW (mW) or dBm.
Yota engineers made sure that Yota Access recalculated the received signal into good, usable dBm units, which I will write below.
The higher the RSSI number, or the less negative it is (just the case of Yota), the stronger the signal.
Before moving on to decibel and its derivatives, you need to talk about CINR.
')
CINR (Carrier to Interference + Noise Ratio) also called SINR (Signal to Interference + Noise Ratio) - literally: the ratio of the (useful) signal to the interfering noise. The established Russian term is the signal-to-noise ratio. Measured in dB.
The value 0 dB means that the signal and noise are equal, a positive number - that the signal is stronger than noise, and a negative number - that is weaker.
The user, of course, is better that the signal was as strong as possible noise :)

Consider the decibel value in more detail.
Decibel (decibel, dB) is a logarithmic quantity that reflects the ratio of a physical quantity (usually power or intensity) to a given reference level.
The ratio in decibels is a tenfold decimal logarithm of the ratio of two powers:

The double power difference will be written in decibels as ~ 3dB.
The ten times power difference will be recorded as 10dB.
A 1000-to-1 power ratio in decibels will be recorded as 30dB.
The beauty of the decibel is that it is convenient to express very large and very small values ​​in it, due to its logarithmic nature.
Often a decibel receives a suffix that indicates the reference value used.
The suffix m in dBm is milliwatt. That is, L (dBm) = 10 * lg (P / 1mW)

Knowing the RSSI and CINR for the WiMax signal, we can fully appreciate the signal quality and please the user with a picture with sticks:



We decipher.
CINR 32dB - the received signal is more than 1000 times stronger than noise.
Is it a lot or a little?
Even if the signal and noise are equal, with good error correction work is possible. Of course, you should not count on a high data transfer rate in this case.
RSSI -58dBm– the received signal is almost a million times weaker than 1 milliwatt, that is, it is one nanowatt (1nW). A microwave oven operating at the same frequency roasts 800W.
Values ​​differ by 12 orders of magnitude. WiMax radiation is so safe :)
These are ideal parameters (which are observed in my home, on a UMPC with an Intel 5150 controller and a good antenna), the speed in this case reaches a theoretical maximum of ~ 10Mbit / sec.



For comparison, a less joyful picture (the controller works without an antenna at all, an unstable reception zone is simulated):



CINR 14dB, RSSI -81dBm - the “only” signal is 40 times stronger than noise (25 times worse), the signal power is 1000 times less.
The speed with such “blatantly bad” parameters is about 1Mbit / sec.

Instead of a conclusion, I will answer the following question:
“Full signal 4 sticks, RSSI -60 - -65 dBm. CINR from 0 to 5 dB. Why speed is disgusting? ”
The answer is simple. RSSI and CINR are not directly related. With a normal RSSI, we see that the user encounters noise (interference from a jammer, or something else), which prevents him from using the Internet.