RSSI vs. dBm

4 posts by 2 authors in: Forums > CWTS - Enterprise Wi-Fi Fundamentals
Last Post: October 18:
  • I'm using am Intermic wireless scanner. It shows RSSI: -60 dBm.

    Since RSSI is an arbitrary number, set by manufacturer, what the heck does dBm mean?

    Let's say the scanner requires RSSI of, I don't know, let's say -70 dBm.

    What coverage do I give that area when surveying with AirMagnet? AM shows in dBm.

    AM and Intermic don't know the relationship.

    And what it's actually being measured? RF power only when the beacon is transmitting?

  • By Howard - edited: November 23, 2015

    It is true that RSSI can be an arbitrary number.  Usually however, the radio manufacturer does have some idea of the true power, in dBm, that the radio is sensing. However, that does not mean that they publish these numbers.  

    One of the companies reports that I  do trust are those published by Cisco regarding some of their AP's.  These charts show the dBm value versus the 0-255 RSSI value, sometime even in one RSSI increments.  Reading these charts you can see that even with a range of 255, some values are ill defined as to real power level.   This usually happens at the far extremes of received power, say above -40 or below -85 dBm.   

    Many manufacturers use a smaller range for RSSI, say 0 to 31, and these numbers should be considered even more suspect.   Unless you have a utility that reports the real numbers, and you can interpret them correctly, you can't know the true values.

    Their is a new number that is supposedly more reliable than RSSI, and is mentioned in the /n/ac specifications.   Unfortunately, I can't remember the name of it at the moment- it is not RSSI.   I can tell you that it is only accurate within plus/minus 5 dB.    That's not anywhere accurate enough for my work at the PHY level.

    One problem we face is that most modern utilities do not give you the actual RSSI number as reported by the chipset.   It has already been converted to a dBm number and we have no way knowing how accurate that is.   An even worse number, in my opinion, is a signal report in percent.   What in the world does 50 % signal mean ? Percent of what ?

    This is not to say that manufacturers are lying, although many are suspect when publishing transmit power levels , which is a whole other conversation.   Some are trying to give you the most honest report they can, but this is sometimes influenced by how knowledgeable they think their users are.   For example, a report in % !  

    The best way to run a site survey is to become VERY familiar with it using the radio promoted by the survey tool manufacturer.   Short of that, I suggest you run your own comparative tests against different adaptors, and find which radio works best for you.

    Hopefully, you'll let us know what you find.

    Regarding "Beacon Only" Power levels - In my opinion this might be a valid metric, depending on how fast the client is travelling.  But of course, at the PHY level everything can be measured.   I'm sure it depends on how much time & CPU power the survey tool is willing to spend reporting it.

  • By Howard - edited: November 23, 2015

    Update to my previous reply on RSSI  vs. dBm.

    The name of the field that I could not remember earlier is the Received Channel Power Indicator (RCPI).   Although it has some issues itself, it sounds more useful  than RSSI to me.    

    BTW, I am betting that some manufacturers intentionally misstate which value they use.

    The RCPI is an 8  bit number that ranges between 0 and 220.   The values 221-254 are reserved, and 255 means the value is unavailable. 

    It is updated in increments of 0.5 dBm.   0 corresponds to -110 dBm and 220 represents > 0 dBm.   Measurements are rounded (up/down ?) to the nearest 1/2 dBm.  For my earlier example of -40 t0 -85 dBm that corresponds to values of 140 and 50 respectively.  It is supposed to be accurate within 5 dBm (bah).

    Significantly, it is the average of all received streams, over the data portion of the frame.   This detail is important, as earlier RSSI measurements sometimes varied over the length and/or over which portions of the frame that were included.   Specifically excluded in this measurement are the frame preamble, header, and trailer bits.

    The 0.5 dB step size sounds wonderful at first, but the plus or minus 5 dBm accuracy sounds awfully wide.  (My Anritsu has spoiled me.)   From a range perspective this is a variation of over twice the theoretical range between radio peers.

    It also takes some (unspecified ?) amount of time before the value can be reported, and this may also affect its utility, especially for survey software.

  • By Howard - edited: October 18

    Part of the difficulty in specifying RSSI or RCPI has to do with the techniques used to convert power levels to either of these values.

    Logarithmic Amplifiers (Log Amps) are what are used to make these measurements.   

    This Analog Device's document clarifies the idea behind Log Amps, and explains why "logarithmic converter" would be a better description:

    If you want to understand why neither of these numbers are very accurate, or at least consistent between devices, this document can help.

Page 1 of 1
  • 1