From what I understand, unfortunately toggling the country on the bullets doesn't work because Ubiquity just used a lowpass filter preventing any channels below 149 from being used.
It'll be great if I can find the alternate version for sale because it's a great value as you indicated.
Thanks for explaining the near end cross-talk thing. Given what @Howard is sharing it sounds like that's something I need to be more aware of for the years to come.
I appreciate your sharing all this helpful info regarding Ethernet NEXT (Near End Cross Talk) and how cheap cable testers no longer cut it.
I very much wish the problem was the units being operated on the same channel. I did that once just the for heck of it and accordingly saw the SNR drop like a rock such that there was just as much noise as signal.
Unfortunately no DAS running here.
It may have been an L-Com's or Fluke report, instead of Tessco. Here is a piece I found today from Beyondtech/L-Com:
One of the CCCA papers on the subject:
One very interesting paper on the CCCA site is from Fluke:
It describes copper clad aluminum wires, and how some bogus manufacturers are selling boxes of these cables with extra "ballast" in the shipping boxes to make them appear to the casual observer to be legitimate.
P.S. You are probably fortunate that you DON'T have any DAS around.
Saw the postings this morning. I believe the discussion concern 5 GHz spurious emissions for indoor environments.
For my Friday presentation at CWNP conference, I believe I may have a slide that shows what occurs when multiple RAPs all on same 5 GHz Backhaul channel compete (no separation between noise floor and SNR). This is for an outdoor environment.
Andi over at LEDE forum explains what's going on.
New units Linksys shipped me behaved the same. I opted to try a pair of Ubiquity Unify UAP-AC-PRO instead.
At the same 10' distance the achieved throughput is only 20% diminished instead of 90% diminished with the Linksys units.
For those interested, I even placed the Unify units immediately beside each other and the achieved throughput was only 54% diminished.
My hat is off to Ubiquity!
I'm glad you got it working. I have found Ubiquiti gear to be quite good especially for the price. The UAP AC Mesh Pro would have a better antenna pattern for mounting on a wall, though. Otherwise it is the same device as UAP AC Pro I believe.
Andi's words could have been mine:
You can try a metal sheet between the APs but I’m not sure if the clients aren’t radiating out-of-band noise also.
Hard to judge without a spectrum analyzer
I sound just like my mom used to: "Didn't I just tell you?" :o)
So, what did we learn?
The theory about side lobes can't hold, because there is no way the side lobes could extend from ch36 to ch165. That would make the whole channel division useless.
The theory about transmitters inducing currents directly into each other's circuitry doesn't make sense either, because dropping the transmission power didn't have any effect.
Since replacement with Linksys units didn't help, it must be a flaw in Linksys design. One of:
- The transmitter filter is bad (but the FCC should have caught this)
- The receiver filter is bad (but this would make them useless in any crowded environment)
- There is no shielding whatsoever on the devices
Only the first flaw would be visible in spectrum analysis.
@Petri I greatly appreciate your keeping me on track with sound logic.
Petri...The theory about side lobes can't hold, because there is no way the side lobes could extend from ch36 to ch165. That would make the whole channel division useless.
That's pretty much what I thought, yet I corresponded with SmallNetBuilder staff that was pretty assertive that the side lobes could extend that far.
SNB...The 802.11ac spectral mask for 80 MHz bandwidth requires the signal be down only 40 dB 120 MHz either side of center channel. That doesn't mean the signal goes away after that or even gets lower.
If I understand correctly, SmallNetBuilder's contention is that the 5G radio on devices that go from 36-161 aren't filtered like dual 5G radios actually have one radio with a high filter and one with a low filter to prevent them spilling on each other and that devices doing VHT80 on 36 or 161 may actually have side lobes that don't fall off.
I look forward to the coming decades whereby radio/chipset/antenna/filter engineers will devise solutions to make appliances far more precise/reliable so that people only have to worry about adjacent channels and nothing 2 or more channels away.
Am I blind ? I did not see anything in the original posts about using wide channels.
That's a mistake. Even 40 MHz wide channels can be a serious problem.
However, if the Ubiquity works.with the same wide channels, it definitely means their "secret sauce" is better.
Either way, I recommend you forget the wider channels - they won't be usable 90+% of the time anyway..
My solution for any type of spectrum analysis is not inexpensive. Anritsu MS2760A.
Various antennas up to 110 GHz.
Ultraportable (see video: https://www.youtube.com/watch?v=TvYQN-y1wWI)
Also useful around elevator shafts, utility pole transformers, leaded room (healthcare).
Another product, that is much less inexpensive, runs on Surface Pro (have not tried it on a MacOS Pro with VMWare) is: veex
Good stuff. Yeah if I was going to do spectrum analysis I was thinking I'd get a Ubiquity Bullet M5 for $100 yet I imagine it's not as good as the solutions you're proposing.