Missing channel(s) at bottom of 5GHz UNII?

Products ePMP
1

Hello all,

Our old Ubiquiti AirMAX M radios that we are slowly Elevating away from :-) allow us, even with current firmwares and "Revised UNII rules", to choose a center freq as low as 5165MHz with 20MHz channels.  When our ePMP APs are set to 20MHz channel width, the lowest channel center we can pick is 5180.

Why is that?

Everything else between AirMAX M and ePMP appears to match up, spectrum-rules-interpretation wise.

-- Nathan

2

I wondered that same thing... why is the lowest allowable 20mhz channel at 5180, but it's 5160 when using a 10mhz channel? Shouldn't the lowest allowable channel be 5170 when using a 20mhz channel width, and 5180 when using a 40mhz channel width? I know it's just 10mhz difference, but when using wider channel widths, every little bit counts.

3

Eric, 

Due to band emissions requirements, we had to exclude some channels on the band edges. 5.1 is right next to a licensed band and this was needed to keep the licensed spectrum clean when operating near the edge. 

Sriram

2 Likes
4

Thanks for the explination Sri... I figured it might be something like that.

1 Like
5

This makes sense and is what I suspected, but might it be that you picked too conservative of a band edge on the bottom end of the spectrum?

My understanding is UNII starts at 5150.  Older Ubiquiti radios used to allow for 5160 w/ 20MHz channels as a center, which would go from 5150-5170, and also 5840 as a center on the high end, but after the revised rules came out, Ubiquiti changed this and brought both ends in by 5MHz...so, 5165 on the bottom end and 5835 on the top.  Presumably for the same reason you cited (some measurable bleedover below 5150).

Your 5835 on the top matches Ubiquiti's, which is why I am confused.  If ePMP cannot go below 5170 without that edge bleeding into the band below it, why is the top end not similarly bounded (e.g., to 5820)?

Also I have a hard time believing that UBNT has cleaner/tighter rolloff at channel edges than you do...but they at least seem to THINK they do...

-- Nathan

EDIT: Also, I think Eric's earlier point, which I had not thought about when I first posted, still stands: it doesn't make sense that with 10MHz channels, you can pick 5160 as a center.  That would be 5155-5165.  So...why with 20MHz can you not pick 5165 as a center, which would be 5155-5175?  The "left" channel edge is the same in either case, but for 10MHz it's allowed and yet not for 20MHz?

6
@nathanawrote:

EDIT: Also, I think Eric's earlier point, which I had not thought about when I first posted, still stands: it doesn't make sense that with 10MHz channels, you can pick 5160 as a center.  That would be 5155-5165.  So...why with 20MHz can you not pick 5165 as a center, which would be 5155-5175?  The "left" channel edge is the same in either case, but for 10MHz it's allowed and yet not for 20MHz?

It does make sense. A wider channel will have wider "bleedover" (IMD's and such). 

1 Like
7
@CWB wrote:
@nathanawrote:

EDIT: Also, I think Eric's earlier point, which I had not thought about when I first posted, still stands: it doesn't make sense that with 10MHz channels, you can pick 5160 as a center.  That would be 5155-5165.  So...why with 20MHz can you not pick 5165 as a center, which would be 5155-5175?  The "left" channel edge is the same in either case, but for 10MHz it's allowed and yet not for 20MHz?

It does make sense. A wider channel will have wider "bleedover" (IMD's and such). 

Correct. Also 20 MHz and 40 MHz channel behave the same because 40 MHz is really two 20 MHz in the 802.11 world. 

1 Like
8
@Cambium_Sriwrote:
@CWBwrote:
@nathanawrote:

EDIT: Also, I think Eric's earlier point, which I had not thought about when I first posted, still stands: it doesn't make sense that with 10MHz channels, you can pick 5160 as a center.  That would be 5155-5165.  So...why with 20MHz can you not pick 5165 as a center, which would be 5155-5175?  The "left" channel edge is the same in either case, but for 10MHz it's allowed and yet not for 20MHz?

It does make sense. A wider channel will have wider "bleedover" (IMD's and such). 

Correct. Also 20 MHz and 40 MHz channel behave the same because 40 MHz is really two 20 MHz in the 802.11 world. 

Okay, thanks; I am not an RF engineer (as if that probably wasn't painfully obvious) so all of this is helpful. :-)

Does the width of that "bleedover" really grow exponentially like that?  A 10MHz channel only needs to maintain a 5MHz guard band (between 5150 and 5155), but a 20MHz channel needs to maintain a whopping 20MHz guard band (between 5150 and 5170)???

In the U.S., it looks like 5000 - 5150MHz is set aside for aeronautical navigation, and 5850 - 5925MHz is for DSRC. So I'd also still like to understand why it's necessary for a 20MHz-wide channel to maintain a 20MHz "guard" on the bottom end of the spectrum (lowest center 5180, 20MHz gap of 5150 - 5170), but for the same 20MHz-wide channel to only maintain a 5MHz "guard" on the TOP end of the spectrum (highest center 5835, 5MHz gap of 5845 - 5850).  I assume the "width" of the emissions rolloff is roughly the same on both edges of the channel, so if everything that's been said so far is true, shouldn't 5820MHz be the highest allowed 20MHz channel instead of 5835?

Finally, does UBNT really have cleaner/tighter emissions, or are they risking a slap by the FCC by allowing their customers to use 5165 as a center freq with 20MHz-wide channels?

-- Nathan