The axiom that “everything is bigger in Texas” probably originated as a reference to the state's vast geography. In square miles, Texas is second only to Alaska in the United States. When discussing Wi-Fi 6E, “everything is bigger in 6 GHz”. Especially when talking about bigger and wider channels. The new 1200 MHz of frequency space truly is a spectrum bonanza.
Actually, large Wi-Fi channels are not a new concept. 802.11n (Wi-Fi 4) introduced the use of 40 MHz channels in the 5 GHz frequency band. 802.11ac (Wi-Fi 5) introduced 80 MHz and even 160 MHz channels in the 5 GHz band. Wi-Fi 6 radios that operate in the 5 GHz band can also use these wide channels. But the truth is that there simply is not enough frequency space in 5 GHz for an effective 80 MHz channel reuse plan. As a matter of fact, 20 MHz channel reuse plans are still dominant in 5 GHz instead of 40 MHz.
As shown in the first figure, 40 MHz channels are created by bonding together two 20 MHz channels. Channel bonding effectively doubles the frequency bandwidth, meaning double the data rates. And doubling the data rates means double the throughput.
Primary and secondary channels are used together only for data frame transmissions between an AP and a capable client. For backward compatibility, all 802.11 management and control frames are transmitted only on the primary channel. Additionally, only the primary channel is used for data transmissions between an AP and clients
Figure 1 - Channel bonding
Although 20 MHz channel reuse is the norm, 40 MHz channel reuse patterns can indeed be effective in the 5 GHz band with good design best practices:
On a side note, channel bonding should never be used in 2.4 GHz in the enterprise.
Now, very large 80 and 160 MHz channels are created by bonding together multiple 20 MHz channels. Even though 80 MHz and 160 MHz channels are available in 5 GHz, they should not be used in the enterprise. Bottom line, 80 MHz and 160 MHz channel deployments do not scale in 5 GHz. There is simply not enough 5 GHz frequency space, and co-channel interference is guaranteed to occur. Performance will drop significantly if 80 MHz channels are deployed on multiple 5 GHz APs in the enterprise.
But remember, “everything is bigger in 6 GHz,” so the potential of large channel reuse plans for 6 GHz is quite astounding. Because of all the available frequency space in 6 GHz, it is expected that 40 MHz and even 80 MHz channel reuse plans will become standard practice. As shown in Figure 2, in the United States, 29 new 40 MHz channels can be used in a channel reuse plan. If all 29 of the 40 MHz channels are used, performance degradation due to co-channel interference should be a non-issue. In Europe, there will be 12 new 40 MHz channels available for a reuse plan.
Figure 2 - 6 GHz for Wi-Fi – a spectrum bonanza
And for the first time, the use of 80 MHz channels will be a reality for enterprise deployments. In the United States and other countries, 14 channels are available for an 80 MHz channel reuse plan in 6 GHz. Co-channel interference can probably be limited if all 14 of the 80 MHz channels are used together with careful planning and design. As a matter of fact, I expect that most enterprise vendor APs will have 80 MHz enabled as the default channel size.
However, you might ask, “What about the noise floor?” - Another problem with channel bonding is that it also results in a higher noise floor. The noise floor rises 3 dB when 40 MHz channels are used. If the noise floor is 3 dB higher, then the signal-to-noise ratio (SNR) is 3 dB lower, which means that the Wi-Fi radios may shift down to lower modulation data rates. In many cases, this offsets some of the bandwidth gains that the extra frequency space provides. If 80 MHz channels are deployed, the noise floor is 6 dB higher, and the SNR is 6 dB lower. A 6 dB hit on SNR is significant.
Would this problem still apply to 6 GHz? Yes, but there is a solution. Instead, the FCC has defined new transmit power rules that actually favor the use of large 80 MHz channels.
An intriguing difference in 6 GHz will be the new power spectral density (PSD) rules that should offset rises in the noise floor caused by channel bonding. PSD is the measure of signal-strength (energy) variations as a function of frequency. A unit of PSD is energy per frequency (width), for example, 5 dBm/MHz.
For 6 GHz channels, the FCC limits radios and antennas by power spectral density (PSD). The FCC will allow a maximum radiated power spectral density of 5 dBm per 1 megahertz for low-power indoor (LPI) APs. Under the control of the LPI APs, clients are permitted a maximum radiated power spectral density of -1 dBm per 1 megahertz. Confused yet?
Let me try and make this a little easier to understand. Signal-to-noise ratio (SNR) is a great metric for a quality RF signal. Figure 3 shows that the SNR is simply the difference in decibels between the received signal and the background noise (noise floor) measured in decibels (dBs). If a Wi-Fi radio receives a signal of –70 dBm and the noise floor is measured at –95 dBm, the difference between the received signal and the background noise is 25 dB. Therefore, the SNR is 25 dB.
Figure 3 – Signal-to-Noise Ratio (SNR)
An SNR of 25 dB or greater is considered good signal quality, and an SNR of 10 dB or lower is considered very poor signal quality. An SNR of below 10 dB will likely result in data corruption and retransmission rates as high as 50 percent. To ensure that frames are not corrupted due to a low SNR, most enterprise Wi-Fi vendors recommend a minimum SNR of 20 dB for data WLANs and a minimum SNR of 25 dB for WLANs that require voice-grade communications.
In the past, equivalent isotropically radiated power (EIRP) has always been a constant and the SNR a variable. Every time you double the channel size, the 3 dB rise in the noise floor lowers the SNR. However, the new FCC indoor rules in 6 GHz instead keep PSD constant. And every time you double the channel size, the EIRP is also increased. As shown in the table below, the noise floor still rises 3 dB, but so does the EIRP by 3 dB. The result is an effective EIRP that stays at the same level, and even more important is that the SNR remains constant. Please note that SNR is dependent on the RF environment, and the table assumes that we have a quality SNR of 25 dB.
The conversion between PSD and EIRP can be calculated with this simple logarithmic formula:
EIRP = PSD (dBm/MHz) + 10log (channel width in MHz)
The bottom line is that the 5 dBm/MHz rules will compensate for the 3 dB rise in the noise floor when 6 GHz channels are bonded. As a result, 80 MHz channel reuse patterns in the enterprise could become common in countries with the entire 1,200 MHz of 6 GHz frequency space available. In Europe, 40 MHz will probably be more common because there are twelve 40 MHz channels available for a reuse plan, but only six 80 MHz channels.
Keep in mind that 6 GHz Wi-Fi has not been field-tested yet; therefore, it remains to be seen if 80 MHz channel reuse patterns in the enterprise become prevalent. Some high-density verticals might still be better served with 40 MHz channel plans. So, will 160 MHz channels be used in the enterprise? Probably not at any scale, however, APs deployed in isolated areas could use a 160 MHz channel.
Depending on the vertical and Wi-Fi uses cases, we expect some creative channel-planning to be used with the additional availability of the 6 GHz spectrum. For example, 2.4 GHz is already considered a best-effort band reserved for legacy client devices and IoT Wi-Fi devices. Critical applications, such as voice over Wi-Fi (VoWiFi), have already been designated for use on 5 GHz only. For indoor Wi-Fi, we might see certain mission-critical applications being used over both very specific 5 GHz channels and specific 6 GHz UNII bands. Segmenting application traffic by frequency bands will not be the model that most people use; however, the 1200 MHz of 6 GHz frequency space provides room for creativity.
Another thing to consider when deploying big channels in 6 GHz is the selection of the primary channels. When channel bonding is used, one 20 MHz channel is selected as the primary channel, while all the other channels are known as secondary channels. As previously stated, the primary channel is used for the transmission of management and control frames. On the other hand, data frames are modulated across the entire bonded channel. In 5 GHz, almost always, the first 20 MHz of any 40, 80, or 160 MHz channel is the primary.
This will be different in 6 GHz; instead, the second 20 MHz of any 40, 80, or 160 MHz channel will usually be the primary. (Although this is not a requirement) - Take a look at Figure 4, which depicts the U-NII-5 band in 6 GHz. I refer you to the earlier blog discussion about preferred scanning channels (PSCs) used for active discovery by clients in 6 GHz.
Figure 4 - Preferred scanning channels and primary channels.
If PSC is used, the PSC channels must also serve as the primary channels when channel bonding is used for 40, 80, or 160 MHz channels. Please note that the positioning of the PSC channels is the second 20 MHz of a bonded channel. If enabled, the PSC channels also function as the primary channel for half of the 40 MHz channels and all 80 MHz channels, as shown in Figure 4.
In another previous blog, I discussed how out-of-band AP discovery mechanisms will be the primary method for Wi-Fi 6E clients to find Wi-Fi 6E APs. However, because PSC may be enabled on many APs, my recommendation would be to always use the second 20 MHz channel as the primary in any 40, 80, or 160 MHz channels in 6 GHz.
I am fond of saying that Wi-Fi 6E provides us a new spectrum bonanza - So, saddle up and get ready for Wi-Fi 6E because everything is bigger in 6 GHz!
Portions of this blog have been excerpted from the free eBook: