I have over 10 years of experience with wireless and I have never been so disappointed and frustrated.
lately nv2 protocol is just a big disaster.
I already changed most of SA units to 802.11 protocol where I got far better throughput
For example SA unit p2mp 20 clients all SXT Lite5 signals are from -42 max -57 max distance 600-700m in the peak hour when nv2 is selected
throughput is about 15-20mbits + I do some btest to some clients throughput can go up +2-3mbit, but when I select 802.11 protocol throughput immediately jump 50-60mbits after a while throughput stays above 40mbit on the top of that when I do btest to some of clients total throughput is above 70mbit.
What’s the catch I already try changing cell radius, period size, simple all nv2 situation sucks.
After 6.40.x Mikrotik introduced option where you can set downlink ratio I mean what’s with that does anyone got better throughput
Oh yes and synchronization in what scenario is tested? Does anyone get it to work?
Just for fun I already made a test two SA mounted on opposite sides same channel, 802.11 protocol I got 2 times more throughput than when nv2 synchronization is in play.
At the end one simple question - m I the only one hеre whit such a bizarre test results?
Testing SA as AP 15 clients connected SXT5 location is on remote side where interference is very very low 802.11 vs nv2
SA ethernet port is limited to 100mbit.
nv2 protocol actual clients throughput plus btest(TCP 5 connections) to one client
nv2.png
802.11 protocol actual clients throughput plus btest(TCP 5 connections) to one client
802.11.png
I registered just so I could agree with you. NV2 for a WISP is unusable and its really hurting my business. We now try limit our Mikrotik sectors to 15 clients and even that doesn’t always work. In general we get even worse results than you do, I’m guessing because we run larger distances.
I’ve obviously spent time looking for alternatives, which was never something I wanted to do before, and found that NV2 could quite easily be the worst performing TDMA protocol out there. I’ve seen test result online for opensource TDMA protocols that look better than NV2 let alone the actual competition.
Like you said, very disappointing. It would be nice to hear Mikrotik acknowledge this and provide some kind of timeline for a resolution.
I have this problem
But problem is on AC devices and nv2 ex: SXT 5 AC
nv2 was big disappointment 2 year after first submission, now several years after that disappointment have epic size.
nv2 in same conditions have almost double lower TCP troughput than next comeptitor TDMA protocol and others is mutch better (in ptmp).
even worst nv2 in congestion area dont even scaling troughput by widening bandwith from 20 to 40MHz, in other word we have almost same troughput on 20MHz and 40MHz, its ridiculous.
I am convinced that the problem is in the poor connections of certain clients, but competing tdma protocol are mutch less vulnerable to a similar thing, and they scaling bandwith just nice (20 to 40mhz)
It seems to me that the mikrotik use referent atheros board design, others have less or more modificated and upgraded their radios, maybe that makes a difference also.
Unfortunately, this is not a matter of fines and a few percent, differrences is humongous, and they put us in an unbearable position,
Hi,
we have the same problem with AC and NV2 bandwith!
Point to Point with boths QRT5/AC and 80Mhz:
802.11 = 250 Mbit
nstreme = 160 Mbit
NV2 = 120 Mbit
What does mikrotik say about this problem??
Hi all,
what do you expect? Follow the forums and see that this problem is no news at all. MT does not comment these posts. So do you expect MT to say “our priority is elsewhere go and find a better solution” ? Do you want to be a WISP without gps sync in 2018? Do you know your business?
MT is great at rooting. The new mixed switch/router looks good and they seem to do some changes to ROS to make switch configuration more straight forward. But outdoor wireless …
So do yourself a favor, test equipment and select the appropriate for each job. Time for complaining is long gone.
my 2 cents
my fullest consent
Nothing, and this is the same like on Spectral Scan
Outdoor Wireless is not Main Focus for Mikrotik
@ste
you are sure the mikrotik works on GPS Sync?
Are the wishes or facts?
You misunderstood my post.
In last two years mikrotik put lan routers an switches in front on routerboard.com
Послато са GT-I9195 уз помоћ Тапатока
Good day
I have always used nv2. i tried switching from nv2 to 802 after reading post and in meduim-high noise area 802 does not work for ptmp.I get way better throughput with 802 and slightly better latency in ptp link, but in ptmp clients just keeps on disconnecting because of hidden node problem(extensive data loss in log). All clients is between -55 to -65. While nv2 is not optimized i agree, it still works better in ptmp setup.
thanks
smittie
To mikrotik:-
we only can request only please solve this problems with wireless as soon as possible we understand mikrotik to they are working hard and trying to solve .but our problem is we are loosing clients pls fix this issue as soon as possible .nv2 is killing us with low speed and performance .we all love mikrotik system and we dont want to give up
raj singh
Cambium Elevate - only potential solution.
It’s obvious by now that Mikrotik has just given up on supporting wireless lol
Still only 11n …
Mikrotik outdoor wireless is not more usable Cambium N support for mikrotik elevate 'll be available for Q1/2018 and the AC support for Q2/Q3 2018
The nightmare is close to end 
We believed in mikrotik for log time but we can’t wait more features like GPS sync, mu-mimo, ecc helps to stay on the market
Giuseppe
Cambium is not the holy grail. We remove epmp 1000 due to its bad behavior with interference. Performed better than nv2 but gave us big trouble with disconnects in dense deployments. In this situations nv2 is more stable than epmp. There are better solutions out there. We are already burnt with their pmp320 System so we are no big fan …
For me NV2 works better in high noise environments then nstream and 802.11.
The latest nv2 sync and downlink settings is not the Miracle we where hoping for but a step in the right direction(would be better if I could disable DFS on 6.40.x)
If mikrotik can implement some kind of mu-mimo tech into nv2 in the next year or so it would be great, but only time will tell.
Personally i doubt that the cambium elevate for MT will be worth the hassle at the end of the day.
to be honest nv2 really approaching, and become comparable with even cambium in extreamly noisy areas, with still lower troughput but maybe better stability than others.
It does not make things better, its not really about quality of nv2 but the circumstances make the others same bad.
but of course this applies only to extremely bad conditions.
Hello
I posted on this two years ago.. After much experimenting I found a few things that can help mitigate issues.
- Do not let anyone connect with less than mid 60’S signal strength but it is not just signal strength the CCQ of each connection is vital do not have a connection below 75% CCQ. NV2 will only provide the speed of its worst connection.
- TDMA period …while longer periods mean more data can be sent… it also means a longer time till the AP gets back to each client. Longer times means less data… it is a trade off… if you use 4ms and you have 15 clients that is 60ms plus your guard time… so in any given second you only get half the receive or transmit time of say a 2ms period. More Clients, larger Period, = less speed.
- Number of clients DO NOT GO OVER 16 CLIENTS EVER. The performance drop of NV2 after 16 connections is fast and furious. Go over 16 and by 20 you will see a 90% reduction in speed. 12 clients is even better.
By keeping to this formula I can deliver 25Mbps to each client 95% of the time.
We all use Mikrotik because it is cheap and most of the competition is either way more money for little improvement or it is UBNT and UBNT ptmp sucks hard. In addition the other features of mikrotik give us options that we would not otherwise have. Low density AP’s is the price we pay for price and features.