Netonix Forums

rebelwireless wrote:I'm watching mimosa closely. I'm getting pretty fed up with ubiquiti. I really want to see how the product evolves and firmware issues get handled etc etc to see what kind of vendor mimosa is going to end up being when they are up to volume. I think that the B5-lite might end up being a really good compromise at $600 for a lot of my links, and AF5X is still an ~$1100 link.

Just to quickly correct one thing, the B5-Lite is $300 PER LINK, not per radio. Product is on schedule as well, mass production on its way!

jaimefink wrote:
Just to quickly correct one thing, the B5-Lite is $300 PER LINK, not per radio. Product is on schedule as well, mass production on its way!

What's the ETA on the Lite?
Thanks!

Jaime, Chris corrected me in another thread a long time ago (on the $300 per link)

My real interest here is how a B5lite link compares to a ubnt PowerBeamAC link. The PowerBeams are higher gain, but they also need MUCH better CINR than anything else I've seen. At -65 with somewhat minimal noise (rural environment) the ubnt AC gear starts loosing steam, I find myself switching back to their NanoBridge M5-400 for >-65 because they perform better.

An AF5X link is more expensive, but it's not really fair to compare the AF5X to a B5lite. The fair comparison is obviously the AF5X and B5c.

Hey guys, we just got our first set of B5s up yesterday, and I was hoping to get some info on the comparison with AF5x here as I haven't been able to find any yet.

One big thing that bugs us on UBNT, which I thought AirOS 7.x would fix but doesn't, is that changing any setting resets all services. The B5s don't seem to have this problem, which is especially important in a backhaul application. They only reset if you change wireless settings. How do the AF5xes function in that regard?

The cost difference is pretty steep, we're looking at about 2x as much for B5s as AF5x + antenna, and I'm still a bit wary of whether it's really worth the cost bump.

The other thing I'm seeing is latencies of around 10ms on the b5s where the nanobridges previously handling the link only introduced 1-2ms of latency. On the other hand maybe I'm being too greedy with the channels as it's a noisy environment and retransmits are slowing things down? I'm not anywhere near the link capacity. I tested it out to around 300 in one direction and 350 Mbps UDP throughput and the latency is up around 8-10ms even at idle.

Mimosa's support chat is currently offline (really nice to have, btw), but I'll ask them about this on Monday. If the B5 proves reliable, we'll shut down the old link, which is on UNII-1, or move it to DFS and then the B5 can take over that clean spectrum.
EDIT: This is a 2-mile link, I changed the TDMA window to 2 ms from 4 and the latency is looking MUCH better.

Yea I was thinking your latency was ODD as I do not see that on my link.

Curious what you get on the clean spectrum when you switch.

I got my first AF5X link in last week and have yet to deploy it, waiting on our new RF Armor Rocket cover box that is made for the AF5X.

We also have a cover for the MIMOSA B5C coming out soon as well.

Yea the B5C is pricy but they have a B5C Lite coming out that is MUCH less than the AF5X. "PRICE WARS I GUESS"

I have not had much issue with my B5 link other than I have had to reboot the link on 2 occasions from the UI as it stopped bridging traffic but that was awhile ago and may have been fixed in firmware upgrades?

10ms is very out-of-character. Something is wrong.

the B5* have a few features that the AF5X doesn't, and if you need those features then you can make a point about them. If you don't need those features, the AF5X is significantly better. I have a number of AF5X links now and they all perform at <2ms consistantly. We talking 95% of packets at 1ms, 5% at 2ms, and nothing above that. Sync works fantastically.

It is VERY annoying having to reboot the AF5X for updates though. They take a while to reboot to, so tweaking a link takes a long time.

rebelwireless wrote:the B5* have a few features that the AF5X doesn't

Which ones are you thinking of?

rebelwireless wrote:It is VERY annoying having to reboot the AF5X for updates though. They take a while to reboot to, so tweaking a link takes a long time.

I'm really impressed, maybe it's only with dual channels, so it can switch one at a time, but I can even change some of the link parameters like TDMA window and channel on the B5s and the link stays up.

If they pull this off on their A5 hardware it'll be awesome.

Specifically the dual link. Af5x will switch channels if one goes down, but nothing like the b5s can. Plus the potential throughput of the b5 is higher but the af5x will get a higher modulation on the same signal level so once you get out to any distance that advatange kind of fades. Also, the huge channel width be realistically 80Mhz channels are not much more that 'look what we can do' because using them is a serious corner case.

rebelwireless wrote:Also, the huge channel width be realistically 80Mhz channels are not much more that 'look what we can do' because using them is a serious corner case.

We have some "green field" locations where we can use that, but the bandwidth demand is not there, so... yeah pretty much. The only case I can think of is we're looking at getting on a mountain top that can see for 200 miles or so, and if we wanted to branch out from there we could probably feed it with a B5c, but in that case we'd probably want a proper AF5 at a minimum for proper full-duplex.

Personally I am glad UBNT limited MAX channel size to 50 MHz on the AF5X, this shows responsibility on their behalf.

Now if we can get them to go back and change the AF5 to support UNII-1 full 100 MHz channel usage in that spectrum but limit UNII-2 and UNII-3 to only 50 MHz

I really do wish MIMOSA also limited MAX channel to 50 MHz except UNII-1

I would be happy if they [BOTH] just limited the channel width to 50 MHz in UNII-2 and UNII-3 and let it rip on UNII-1 as that is a PTP link spectrum only anyway.