Is there any progress in updated AREDN firmware that will support Ubiquiti equipment manufactured in 2014 that have XW firmware/boards? I ordered some Ubiquiti Bullets and Nanostations, but they have Ubiquiti's most recent rev levels, manufactured in 2014.
Thank you.
73, Bob
Yes, we have M5 Nanostation XW working in alpha-test. It will be in our next release. Other XW devices will follow in a subsequent release. I have 2 M5 PowerBeams I'm waiting on myself... check that, I had to sacrifice one to Conrad, KG6JEI's, destructive testing to figure out the underlying issue with XW ;-(
Andre, K6AH
Andre,
Thank you. That is good information. I hope it is not more than a month or so before release?
73,
Bob - N7RJN
Bob, we are running the testing process like a formal commercial development team including a written test plan. While a formal release date has not been set, this is all volunteer work. The beta test is on target to start in April. Based on the results of executing the tests, we'll soon have a good idea of a formal release time.
Note, as in the past, there would be 'experimental' builds available to everyone while this is being tested. Our concern is not that people have early images to try out, it has the risk of increasing support when we are trying to do a release.
Also the XW hardware updates bring along an upgrade to OpenWRT Barrier Breaker with massive platform improvements and other fixes. Thus, all functionality is subject to re-test. Let us all hope we find only minor issues in the testing to get this released sooner rather than later.
Joe AE6XE
According to Ubiquiti website there is no XW firmware for bullets. Are you sure your bullet is an XW device ?
In addition the XW should mostly be a 5GHz issue (none of the supported 2GHz devices show as having XW firmware either)
All of my units were manufactured in the last half of 2014, so based on the info on the AREDN web site, I assumed all 2014 versions had the XW software. So far the only unit I actually tried was the Nanostation M5, since a 5GHz link of about 10 miles is my highest priority. This had the XW firmware. I will be able to check the Bullets on Monday or Tuesday.
Thank you.
Bob - N7RJN
Of the devices we officially support it is only a known issue on 4 of the 17 devices we support, the AirGrid M5 HP, Rocket M5, NanoStation M5 and NanoStation Loco M5 (it also affects other devices but they are not on the supported hardware list yet) which is really only a small subset of the devices available.
It also doesn't affect all 2014 boards but it has the potential to affect all 2014 and newer boards (I actually received a board dated the middle of last year on a Rocket M5 that turned out to be an XM board, which was annoying because I wanted an XW for testing) hence the suggestion to avoid anything in 2014 as it COULD be an XW but you wouldn't know until you you inquired.
I'm actually only just in the past couple months starting to hear of large numbers of the XW's showing up which maybe is just I've not been looped in, but would seem to track well with hardware support development in other resource channels that they are just now starting to really make it onto the market floor in the past 4-5 months.
Relevant Ubiquiti Forum Posts
04-02-2014 06:59 AM - edited 04-02-2014 07:00 AM UBNT-Matt_B_ Ubiquiti Employee:
Yes, I can clarify. NanoStationLocos and airGrids are shipping now with the new boards. I think the same day of my last post I saw the first report of the NanoStationLocos with new boards and I just checked, airGrids are shipping with them now. The whole line will be switching to the new board over time.
04-03-2014 08:01 AM UBNT-Matt_B_ Ubiquiti Employee:
Excellent questions. Here's what I do know based on a post @UBNT-Ben put in another thread. We're currently waiting on the NanoBeams to be cleared for DFS, Ben said that we'll continue to manufacture and ship the NanoBridges until DFS is cleared for NBE and then the NanoBridges will be no more. TBH, I do not know if the NanoBridges will/are getting the new boards. It may just be a timing thing with the FCC clearance.
All currently manufactured airMax devices will be getting the new boards as far as I know. Perhaps Ben can correct me if I'm wrong.
04-04-2014 01:26 PM - edited 04-07-2014 06:41 AM UBNT-Matt_B_ Ubiquiti Employee:
The device states the firmware version on the Main tab, it will also reject the wrong format. For example, if you try to upload XW firmware to a XM device, the image check will fail and you'll see an error. It works the opposite way too, trying to upload XM firmware to a XW device, you'll get an error.
Ted,
The primary thread in the OpenWRT community on this issue is tracked here: https://dev.openwrt.org/ticket/19085 . The failure is in the Linux kernel to work around a defect in the physical Ethernet port, to catch and reset the physical network port on the device. The symptoms are that the physical port stops working--a catastrophic issue for DTDlink and LAN traffic the node's basic function depends on.
If you look at the link of the OpenWRT ticket above, someone has created a crowd source funding drive to get the attention of experts that know linix kernel and drivers. The fix bounty is now at $215 as of Sept 28, 2015. I would recommend everyone that reads this, to add at minimum $15 on this bounty. We will be looking at this issue after releasing 3.15.1.0. Unknown timing still.
https://www.bountysource.com/issues/12898488-nanostation-m5-loco-xw-lose...
In regards to building a backbone or major links in an area, I would recommend looking at MIMO or dual chain/antenna devices. The airGrids are single chain (vertical or horizontal polarization based on how it is physically mounted). The MIMO devices (NanoBridges, NanoStations, Rockets, etc.) offer improved throughput and ability to work in challenging environments (noisy and/or lots of fade paths) with both Vertical and Horizontal Polarization used at the same time. But of course, this is the situation where you get what you pay for--MIMO devices are generally higher cost for additional performance.
Joe AE6XE
Thanks for the pointer Joe, I'll follow up on it. I do realize that Rockets are probably the best bet for performance, but this is a low cost first attempt. The 5G AirGrid software just isn't there yet, but at least I can get some testing done with AirOS in bridge mode (connected to co-located 2.4G NanoStations).
Ted. VE3TRQ
Ted - I've had real good luck with two of these Nanobridge M5 units:
http://www.amazon.com/Ubiquiti-Nanobridge-NB-5G25-Outdoor-25DBI/dp/B004M...
They are right around $100 plus shipping. Just make sure you get this exact one, looking just like this. The newer style won't work with the current firmware. I'd highly recommend them for point to point links, or for just a quick experiment/proof-of-concept. I've also been able to get a useable signal off of one while being at a 90 degree angle from the beam and 2 miles away.
Mike, KG9DW
Ted, To avoid an unstable link using an AirOS point to point bridge... If you implement this, take a look at the olsrd config paramenter: "option Mode 'ether' ". Best to comment this out on the AREDN nodes on both ends. A DtDLink over a cat5 cable will drop out if the LQ drops below something like 95% due to this option. Because you are looking to create a DtDlink over RF, best to comment out this option so that this link won't drop out until something like 30% LQ appropriate for an RF path.
Joe AE6XE
http://www.aredn.org/content/nanostation-loco-m5-xw-status
Joe AE6XE
One thing would help us once we start working on support more XW devices is if you could...
1) Set your PC's ip to 192.168.1.2 / 255.255.255.0
2) Telnet/SSH/Putty into the Nanostation (default user is ubnt; password ubnt; default ip 192.168.1.20)
3) Run "cat /etc/board.info
4) Save the results and post it in response to this post.
Thanks, Darryl
This device is just sitting here waiting for something I can load - I got caught early on with the XM-to-XW move.
XW.v5.6.3# cat /etc/board.info
board.sysid=0xe835
board.cpurevision=0x00001122
board.name=AirGrid M5
board.shortname=AG5
board.subtype=HP
board.netmodes=7
board.hwaddr=0418D6F8BA7E
board.reboot=40
board.upgrade=150
board.phycount=1
board.phy.1.maxmtu=2024
board.phy.2.maxmtu=2024
radio.1.name=
radio.1.shortname=
radio.1.bus=ahb
radio.1.subsystemid=0xe835
radio.1.subvendorid=0x0777
radio.1.txpower.max=25
radio.1.txpower.min=-4
radio.1.txpower.offset=6
radio.1.ieee_modes=5
radio.1.antennas=3
radio.1.def_antenna=3
radio.1.antenna.1.id=3
radio.1.antenna.1.name=Feed only
radio.1.antenna.1.gain=3
radio.1.antenna.1.builtin=1
radio.1.antenna.1.chains=1
radio.1.antenna.1.chain.1.id=0
radio.1.antenna.1.chain.1.name=Vertical
radio.1.antenna.2.id=1
radio.1.antenna.2.name=11x14
radio.1.antenna.2.gain=23
radio.1.antenna.2.builtin=1
radio.1.antenna.2.chains=1
radio.1.antenna.2.chain.1.id=0
radio.1.antenna.2.chain.1.name=Vertical
radio.1.antenna.3.id=2
radio.1.antenna.3.name=17x24
radio.1.antenna.3.gain=28
radio.1.antenna.3.builtin=1
radio.1.antenna.3.chains=1
radio.1.antenna.3.chain.1.id=0
radio.1.antenna.3.chain.1.name=Vertical
radio.1.distance_limit_km=0
radio.1.chanbw="5,8,10,20,30,40"
feature.rssi.leds=4
feature.ext_reset=0
You need a spectrum analyzer and the ability to do constellation analysis on this. A bird watt matter is NOT sufficent for working with the RF power of a mesh node. If you manage to peak the RF power it doesn't mean the transmissions will work. Linearity is critical. This is why most cheap external amplifiers for wifi gear actually degrade the signal.
For those who don't know the nodes decrease power to go to higher data rates to keep liniearity. Trying to push these past the specs set by the manufactures specs will very likely degrade performance.
2) PLEASE PLEASE PLEASE take this off your things to do list. These radios are very well tuned and you will most likely just be introducing much more noise onto the band. ;-)
3) It's probably a good way to burn up your finals. ;-)
My measurements indicate that the M2 units are already at or beyond the max. Due to saturation, they need to back down to less than 28 dBm to properly reproduce a complex OFDM signal, according to UBNT spec sheets (22 dBm at the highest data rate see page 13 on
https://dl.ubnt.com/datasheets/rocketm/RocketM_DS.pdf ).
Here (attached image) are my measurements with an HP 8900C peak power meter and 84811A power sensor. Real output is about 1 dB more due to losses in a jumper cable I was using. But you can see the trend ...
The box may be generic labeling and I would not trust what it says. If AirOS reports 5.6.3 XM, I would trust that. This should be a standard install of the AREDN firmware. It would be low risk to proceed directly to use the AREDN factory image upload in the AirOS GUI. But, always a good idea to run the uboot test program and create a backup of the device unique information to have a copy before doing anything.
Re-run the AREDN U-Boot Tester afterwards to verify..
It's near the bottom... under Experimental Builds
Documentation on the Nightly Builds can be found in the ReadMe file: http://downloads.arednmesh.org/snapshots/trunk/readme.md
Andre, K6AH