I know this is going to be a little late for my company, but I'm hoping this conversation can help others.
We're currently deploying a lot of Baicell LTE stuff (much to my dismay). What I am the most upset about is the amount of bleed into neighboring frequencies. I used a spectrum analyzer on the Baicell we have, and then another on the Cambium 3.65 that we have in a different location. Holy crap. The LTE looks like it bleeds heavily into about 20mhz on both sides of the 20mhz channel. The 450, on the other hand, is much cleaner.
Has anyone else done some comparing to see how these differ?
These are a little different situations, because atc is a guy-wire tower with three 120deg KP Performance sectors running 10mhz channels, and the elevator is a single Baicell AP with a 90deg KP Performance sector.
The scans were done with unused Ubiquiti APs that are waiting to be removed. The atc has Ubiquiti APs within 3 feet of the Cambium sectors, and the elevator has a Ubiquiti AP within 20 ft of the Baicell AP.
We are on the opposite path. We have abondoned all Cambium 3.65 and moved to the Baicells stuff. Far better receive sensitivity. The 3.65 LTE stuff works where PMP450 900 MHz stuff won't. I have no info on the spectrum efficiency. Sorry.
That's what my company is doing. The problem is all the excess noise that seems to be created by the LTE stuff.
Have you analyzed the spectrum to see how ugly it is around your APs? This is why I'm disappointed at the massive deployment of LTE in my area. It's alarming that management is putting so many eggs into the 3.65 basket, rather than planning a cambium 5ghz answer for LOS customers, and far away customers who use a lot.
The way airview works can cause those 20mhz-wide chunks of seemingly noise. That may or may not be there, will depend on the relationship between the actual center frequency the LTE stuff is in, the actual center frequencies the airview driver is programmed to hop arround, the channel size of the RF source(in relation to the channel size airview is listening in), the time relation between the presence of energy and the scanning cycle of the airview device.
Overall, airview is a good tool to tell you if there is stuff talking in the band, and in which frequencies the stuff is talking in(and therefore the stuff-free chunks of spectrum, if there are any). You can educatelly guess about the actual magnitudes of the signals being amostrated, but without an 802.11 frame(or preamble) to calibrate against, the power levels you're seeing are an estimate by the airview algorithm.
If you need a measure of how dirty a signal really is, and how much it bleeds out of the intended channels, you should take a look at the FCC certification plots for that device.
If 3.65 Baicells gear is outperforming 450i 900Mhz gear in terms of NLOS penetration or even throughput in my experience, then you are truly doing something wrong or there's interference.
Our problem is interference. We have a ton of oil and gas stuff around here with every instrumentation company using 900 MHz gear for SCADA. It just blows us away at a few of our tower sites. Our old PMP100 FSK stuff could hold on well, but the sites where we have "upgraded" to 450 just can't keep some subs registered, even with signals around -65 dBm sometimes. Super frustrating. Lightly licensed 3.65 has been the answer for us, but the Cambium 450 series just doesn't have the receive sensitivity to connect our customers that used the old PMP100 900 Mhz stuff reliably at -80 dBm. LTE has been working well in those cases down to -100 dBm (our lowest).
Well you can't really blame cambium for the 900 gear issue if you know its a spectrum issue, as for FSK working of course it did it was FSK :P Moving to OFDM means dealing with higher link budgets and modulations. Your solution wasnt going from cambium to baicell it was going from 900mhz to 3.65ghz lol
As for the LTE vs 450 efficency, i'd have to agree but not even just LTE, on wimax we saw some crazy bleed as well into adjacent channels sometimes up to a full 20mhz adjacent.... Cambiums not magic but its a lot better than our old stuff when it comes to adjacent channel bleed.
Who is blaming Cambium for our area having a lot of interference? Did I read something incorrectly? Also, our solution path went Cambium 900 to Cambium 3.65 to Baicells LTE 3.65. Cambium 3.65 would not work at -88 dBm. The LTE solution works very solid at that and lower. lol
Well, we have done a few Biacels deployment... and when it works, It goes fast. I have seen 50mbps where 900 or other 3.65 would not hook up.
I have a few big problems with it... Management of the CPE and troubleshooting are a complete joke. and yes the Biacels is extremely noisy RF wise... but the worst part is we have had competitors ubnt and 450 3ghz totally knock the Biacels out. we started losing clients and we could not tell wich ones were gone because of the lack of CPE management. then the interference got so bad that 0 clients could connect back up.
Yes, when it works it is surprising what it can bounce off of... but it can not take the interference... nor is the management of the CPE anywhere near where it needs to be.
Well, we have done a few Biacels deployment... and when it works, It goes fast. I have seen 50mbps where 900 or other 3.65 would not hook up.
I have a few big problems with it... Management of the CPE and troubleshooting are a complete joke. and yes the Biacels is extremely noisy RF wise... but the worst part is we have had competitors ubnt and 450 3ghz totally knock the Biacels out. we started losing clients and we could not tell wich ones were gone because of the lack of CPE management. then the interference got so bad that 0 clients could connect back up.
Yes, when it works it is surprising what it can bounce off of... but it can not take the interference... nor is the management of the CPE anywhere near where it needs to be.
Yikes man. That was our biggest problem with PMP-320, competitors with Ubiquiti and I think Telrad LTE. The 320 impressed at first with performance on meagre link budgets but when problems appeared (often) we couldn't even figure out what they were.
Wish you guys the best of luck. We gave up on the 3.65 GHz band. Everybody loaded onto it too fast with too much junk.
We have a wimax network we're phasing out as well and have to say it's not equipment/vendor specific, motorola, ericsson, cambium, huawei, baicell, i find all of the WiMAX/LTE gear to have that same fault, its great when its working but its one of those systems designed from the old guard of "carriers have prestine channels, and never touch/change anything once its up"
With our wimax (not 320's but honestly wimax is wimax which is only slightly different from lte at the rf layer) ... The visibility into our customers and what was actually going on was always an issue, i quickly realized how spoiled by FSK/Ubiquiti/Etc... I mean the fact our wimax didn't even have a spectrum scan made me feel blind lol. I mean sure their were ways to ssh into the modem cards and pull raw data and sorta figure things out but come on thats not an honest way of diagnosing issues on a day to day basis lol.
Though i really really really wish cambium would go out and take a look at ubiquiti's web interfaces dear god i was playing with one recently and their UI designers make some REALLY nice interfaces... i mean i love cambium for the simple fact their rock solid, but i feel like the UI and some features like being able to get spectrum without disconnecting customer traffic, are just obvious features that should have come over that still havent :(
Not to mention mimosa with their auto-mapping via the radios builtin gps's... its sad most of the issues i wish cambium would fix are software related, or software stuff they said are coming but are in a SOONtm sort of way lol I really don't mind the premium on the 450i/450m units over say ubiquiti for their stability but it's painful when i see mimosa/ubiquiti moving so much faster on the interfacing and management of the system...
Sure the theirs hardware tweaks i wish they would make (gps in the SMs, gyro/compass chips in the AP's for reporting, secondary radio for spectrum scans, better recieve sensitivity, more power in the stock SMs) some of which they have said are looking into but it still feels like "SOONtm" that could be 100 years from now lol
The latest beta UI's from Ubiquiti look like a circus. can't find anything quickly, everything moves all the time, no color-coding of graphs and their designers think we install stuff on roofs comfortably sitting with 4k screens in front of us.
Kudos to Cambium for keeping their UI's professional and civilized.
LOL old doesn't mean professional and civilized :) the fact ubiquiti doens't properly detect a mobile or smaller screen and adjust according might be an issue, but honestly an easy way to see how old the UI is from cambium, install the chrome extension for cambium and notice how much more useful the UI is and thats done externally without actually having access to the clean info from the radios.
Ubiquiti might have gone too far, and no ones saying cambium should go that far, but seriously the fact i need to click through 2 pages to get to change and "Save" the refresh time is an example of pre-html5 things that should have been fixed, lack of graphing and aggregating stats, "gamifying" as the html5 generation likes to call it things like the AP Eval page to have some color and quickly see good signals vs bad signals, or hell making the ap eval page sortable are things that in this day and age shouldnt be aftertoughts
The chrome extension made by a user takes the UI pretty far towards being much improved/good, but theirs only so far he can take it.
