Yeah, I see these screen shots of people having backhauls hitting at -60 and it's not realistic to me in terms of power needed to actually make the link work.
Then again, UBNT won't publish their AF SNR requirements... ----- Mike Hammett Intelligent Computing Solutions http://www.ics-il.com ----- Original Message ----- From: "Faisal Imtiaz" <[email protected]> To: [email protected] Sent: Saturday, April 18, 2015 11:13:06 PM Subject: Re: [AFMUG] Very interesting post.. While I applaud the initiative to do a comparison test, there are quiet a few things in this test presentation that would cause one to raise an eyebrow... First one that jumps out is ... how come an 802.11n radio is grouped together with all of these other 802.11ac radios ? (EPMP is not the same category radio as the others) Additionally, information is lacking about what firmware versions were in use with these radios, and I believe choosing a fixed tx power for the test, as an approach to equalize testing parameters is bound to provided some skewed results.. especially with radios which are auto balancing power on the CPE side, e.g. epmp and also the mimosa's.. A few years back, Exalt ran a similar test for their Extend-air radios, they list out their test setup in detail here ... http://www.exaltcom.com/ExtendAir-vs-Wi-Fi.aspx I believe if the comparison between these different radios were done along the lines of Exalt test, the results would be much more meaningful to real life deployment... ... My experience with the couple of AC radios is that they need SNR to be as high as possible....which greatly affects performance, additionally, with the Mimosa's I see they are pretty comfortable in working with -30 signals... not sure if the other AC products behave the same way.... While it is nice to see which radio can deal with interference, I for one would like to know how these different AC radios perform when the SNR is reduced...along with how 'hot' of a signal they can deal with ... e.g. with the Mimosa, we see them operating, and performing well, with a -30/-34 signal when the noise floor is -60.... I for one would be interested in seeing what the other AC radios do in such conditions. Regards Faisal Imtiaz ----- Original Message ----- From: "Mathew Howard" <[email protected]> To: "af" <[email protected]> Sent: Saturday, April 18, 2015 8:52:37 PM Subject: Re: [AFMUG] Very interesting post.. This test ignores a few kind of important details... the UBNT and Mikrotik AC radios have no ability to sync, which gives them a significant disadvantage. also, the Mimosa radios are (theoretically) capable of higher throughput since they are the only ones with the ability to use two 80mhz channels... granted, it's pretty rare that is actually possible in the real world, but if you had synced Mimosas everywhere, it could be done. He's also using a $499 ePMP radio, when he should be using a $200 Force110 PTP. That said, the conclusion the the AF5x is the best is probably right :P On Sat, Apr 18, 2015 at 5:55 PM, Ken Hohhof < [email protected] > wrote: <blockquote> If your criterion is performance in the presence of a signal on a different frequency 30 dB stronger than the desired signal, this analysis is relevant. Also, this seems to be the scenario airPrism is designed to address. But how often would this occur? Even if the interference is from another non-synced transmitter on the same tower, you’d think directional antennas would knock the interfering signal down to less than 1000 times the desired signal. I guess this could be realistic if you have a point to point link in the same band as a sector, so that a giant dish at the other end is pointed right at your sector. From: Josh Reynolds Sent: Saturday, April 18, 2015 5:34 PM To: [email protected] ; Seth Mattinen Subject: Re: [AFMUG] Very interesting post.. Horseshit, read the article. Did you miss the portion where Jim said "it's the exact same chip that's in the RM5"? I would have liked to have seen the RM5 in this test as a baseline, but ignoring the results simply because it's N tech in the EPMP is silly. Not only does the throughput drop, but the LEVEL it degrades at is only "bested" by the B5C in a few of the tests. N or not, that's a very poor result. I would love to see other tests posted on this from other people, its always nice to have multiple sources to remove any potential level of bias. Jim did an excellent job on this and should be commended. On April 18, 2015 2:26:50 PM AKDT, Seth Mattinen < [email protected] > wrote: <blockquote> On 4/18/15 2:49 PM, Peter Kranz wrote: <blockquote> Very interesting shootout comparing AF5X, AC-Lite, AC PTP, EPMP-1000, B5c and RB922 https://community.ubnt.com/t5/airMAX-Stories/Radio-Shootout-Pt-2-let-s-try-a-whole-bunch-of-them/cns-p/1232309 Dude didn't seem to catch that the ePMP is an N radio and dismisses it as worst of the worst. Looks to me like it would probably hold up comparably to its AC counterparts if you take that into consideration. ~Seth </blockquote> -- Sent from my Android device with K-9 Mail. Please excuse my brevity. </blockquote> </blockquote>
