Saturday, September 27, 2014

Wi-Fi SNR to MCS Data Rate Mapping Reference

I previously posted a picture of an SNR to MCS data rate mapping table that I have compiled based on various sources of credible research. Keith Parsons has kindly put this information into a printable format for reference. You can download them below.

It should be noted that individual devices perform differently. These tables are simply generic estimates that are a good approximation for many Wi-Fi devices. In other words, it's not perfect.

Click to Download Full Version (PDF)

This table maps client SNR values to MCS indexes for the purpose of determining the data rates that clients can achieve based on the signal quality of their connection to the AP.


SNR is also related to RSSI. Two RSSI values are of importance: the Minimum Receiver Sensitivity and the Expected Receiver Sensitivity. The 802.11 minimum receiver sensitivity tables often referenced in research and testing material are the required minimum RSSI values that a radio should be able to decode a given modulation type and encoding rate (MCS index) with a packet error rate (PER) less than 10%. Most 802.11 radios provide better receiver sensitivity than the minimum requirement. Therefore, the "Expected Receiver Sensitivity" reflects the typical receive sensitivity of clients with the ability to achieve any given MCS index at a lower RSSI than the minimum receiver sensitivity required to pass testing. For example, the minimum receiver sensitivity for an 802.11ac 20 MHz PPDU at MCS 9 is -57dBm, but most 802.11ac radios can decode this PPDU at a lower RSSI such as -62dBm. 

It should also be noted that a receiver's ability to perform Maximal Ratio Combining (MRC) across multiple receive antenna chains is not reflected in this SNR chart. MRC can allow a device to receive the incoming signal at a lower energy level at each of the individual antenna inputs to the RF front-end radio circuitry which are then combined using digital signal processing (DSP) to provide additive gain. This effectively increases the SNR the client experiences. MRC is based on each client device's receive antenna chain specifications and the number of spatial streams being used for the link between the client and AP, with extra receive radio chains being used for MRC. After MRC gain is added, you can use this table to lookup the MCS rate the client may be able to achieve given it's final resulting SNR . Also be aware that many manufacturer receive sensitivity specifications will list RSSI and SNR values 3-6 dB lower than what is specified here because they list the signal level at the antenna input prior to DSP and MRC gain.

Some of the references used to help compile this table (not an exhaustive list):

  1. IEEE and Realtek - Receiver Sensitivity Tables for MIMO-OFDM 802.11n (PPT) - See tables in appendix
  2. Heegard - Range versus Rate in IEEE 802.11g Wireless Local Area Networks (PDF)
  3. IEEE 802.11-2012 Standard - Sections 16.4.8.2 (802.11 DSSS), 17.4.8.2 (802.11b HR-DSSS), 18.3.10.2 (802.11a OFDM), 19.5.2 (802.11g ERP), 20.3.21.1 (802.11n HT)
  4. IEEE 802.11ac-2013 - Section 22.3.19.1 (802.11ac VHT)
  5. Aruba 802.11ac In-Depth (PDF) - See figure 19, page 25


Cheers,
Andrew von Nagy

6 comments:

  1. I'm confused on the MCS value in the chart, doesn't MCS go up to 32?

    ReplyDelete
    Replies
    1. Hi Jeremy,
      For 802.11n the MCS numbers were 0 through 31, with 8 MCS rates defined for each spatial stream 1-4. So you had MCS 0-7 for 1SS, MCS 8-15 for 2SS, MCS 16-23 for 3SS, and MCS 24-31 for 4SS. However, with 802.11ac that changed and there are only MCS 0-9 now and they numbering is the same no matter how many spatial streams are being used. The MCS really describes the modulation type (BPSK, QPSK, 16-QAM, 64-QAM, 256-QAM) and the encoding ratio (the number of bits used for data versus error correction). Essentially, the older method used by 802.11n was redundant since every MCS above 7 was simply the same as the lower MCS rates just numbered higher to account for the number of spatial streams.

      Hope this helps.
      Andrew

      Delete
  2. Hey Andrew,

    What is the meaning of the difference in color (lighter/darker) for the same MCS rate?
    ex. 29dB, MCS = 8 (color is light green)
    30dB, MCS = 8 (color is dark green)

    Best Regards,
    Koen

    ReplyDelete
    Replies
    1. Hi Koen,
      That is a coloring typo. Thanks for pointing it out, I'll get it updated. The color shading is used to indicate the different encoding ratios used with the same modulation type (e.g. 1/2 versus 2/3 ratios).

      Thanks,
      Andrew

      Delete
  3. Hey Andrew,

    What contributes more to SNR, wider BW or modulation scheme(64 qam to 256 qam). Also if one day we see 1024QAM, would it be almost impossible to send at an 80MHz channel due to SNR?
    thanks, Love the blog!

    ReplyDelete
    Replies
    1. Hi Kenny,
      The SNR is based on the noise floor. There can be many contributing variables such as the thermal noise floor (based on the channel width), noise from indistinguishable Wi-Fi frames (e.g. on an adjacent channel), and non-Wi-Fi noise (e.g. microwaves). The radio circuitry in the receiver also plays a part in regards to receiver sensitivity.

      Regarding the channel width, the noise floor will go up by 3dB for every doubling of channel width.

      Regarding modulation scheme, SNR impacts what modulation can be used since it has to reliably recover the encoded data. It's not the other way around (modulation type doesn't affect SNR). 256-QAM is already very tough to reliably decode and requires a very high SNR (approx. 30-35 dB SNR depending on channel width).

      Cheers,
      Andrew

      Delete