(moderator: moving to the new BETA TESTING forum)
I did a factory load v3.15.1.0b01 on a NanoStation M2 (XM), which worked fine (noting
the Ch. 0 and Ch. -2/254 issues found/reported earlier). Operationally this version seems
to be working great on several other M2 and M5 devices.
NanoStation M5 factory load initially seemed OK but only allowed frequency setting up to 5825,
not up to 5920. After saving/rebooting the menu only displayed frequencies from 5182 to 5485.
I also got an "Untested Hardware" warning at the top of the page, pointing to a "sysinfo" text file.
Rick VE3CVG has confirmed he has similar issues on another NanoStation M5.
I've backed off to 3.0.2 which works up to 5825, but it's telling me I can select TX power
between 6dBm and 32dBm (a bit high for the 27dBm NanoStation). This version also
gives the "Untested Hardware" warning.
Upgrading back up to 3.15.1, results were same, so backed off again.
I am very certain correct firmware was used: non-XW, factory load for NanoStation M2 and M5
airOS version before first load was XM v5.5.6 .
I can enter a ticket, just wanted to post this first.
Dave
Thanks for the report Dave. Please open a ticket for this.
Note, there is an identified bug on some of the new channels 5Ghz. I don't believe there is a specific bug entered to make the details visible just yet. The default kernel code on channel selections does not calculate the right freq settings--goes to neverland. No one before us has ever mapped channels to these frequencies before (a non-ham has never had cause to map channels to a ham freq...).
Joe AE6XE
Sounds like an older unknown series of device. File a board.info ticket with a request to support as sounds like it's an unknown untested all the other flaws are somehwat not relevant.
The fact it was AirOS 5.5.6 means it's certainly older, Ubiquiti makes minor changes every now and than and changes a board is file we use to watch.
Below is the first part of the "sysinfo" file. I'm not yet up on various SoCs used
in various models/vintages. We got a good deal on used gear, taken out of service.
NanoStation M2 also had airOS XM v5.5.6 and no issues with it.
Dave
node: VE3KMV-NanoM5-1
model: Ubiquiti Nanostation M
!!!! UNSUPPORTED DEVICE !!!!
boardid: 0xe005
Device has not been tested. Please file a ticket with your experiences.
eth0 00:27:22:33:33:6C
eth0.1 00:27:22:33:33:6C
eth0.2 00:27:22:33:33:6C
eth1 02:27:22:33:33:6C
wlan0 00:27:22:32:33:6C
wlan0-1 00-27-22-32-33-6C-00-44-00-00-00-00-00-00-00-00
/proc/cpuinfo
system type : Atheros AR7240 rev 2
machine : Ubiquiti Nanostation M
processor : 0
cpu model : MIPS 24Kc V7.4
BTW I'm a bit fuzzy on how non-developers would submit a ticket; have not done
one yet. If it's done through Bloodhound, looks like my AREDN login doesn't work,
so I need to register, or.... ?
Dave
At the moment its a controlled setup, mostly the beta team and active developers have accounts for bloodhound, I'm sure that will change as we continue integrating the systems together, but for the moment they are indeed separate.
Go ahead and file the details here and we can create the appropriate ticket for it.
What we actually need to move forward on this is a board.info file, which I realize as I search for the procedure to link to has never actually been needed yet so here it is:
The file contains details needed about the hardware offsets so we can program the gui to correctly display them.
I only have one unit and it now has AREDN firmware. If it's feasible to re-install airOS of
the right version, I could try; or check if any in the local group has a similar unit which has not
been converted yet (sounds easier).
Dave
absolutely.
Here is the generic procecure to doing a TFTP load.
http://bloodhound.aredn.org/products/AREDN/wiki/HowTo/FlashUbiquiti
Essentially:
Darryl
OK, thanks Darryl, done. Here is the board.info file listing:
board.sysid=0xe005
board.cpurevision=0x000000c2
board.name=NanoStation M5
board.shortname=N5N
board.subtype=
board.netmodes=7
board.hwaddr=00272232336C
board.reboot=30
board.upgrade=150
board.phycount=2
board.phy.1.maxmtu=1524
board.phy.2.maxmtu=1524
radio.1.name=
radio.1.shortname=
radio.1.bus=pci
radio.1.subsystemid=0xe005
radio.1.subvendorid=0x0777
radio.1.txpower.max=27
radio.1.txpower.offset=5
radio.1.antennas=1
radio.1.def_antenna=4
radio.1.antenna.1.id=4
radio.1.antenna.1.name=Combined
radio.1.antenna.1.gain=16
Any other files I should grab/list etc. while I'm at it?
BTW how does the SoC listed earlier (Atheros AR7240 rev 2) fit in with the one(s) currently supported?
Dave
"BTW how does the SoC listed earlier (Atheros AR7240 rev 2) fit in with the one(s) currently supported?"
Ultimately, we just have to test and try it to confirm everything works. This looks like a NanoStation M5 with a cpu 'rev' vs. a new cpu that comes with the 'XW' ubiquit boards. The XW boards come with the AR9342 chip. There's also some AirGrids out there with ar7240 rev 2 boards. The signs are good this will work. The 32dBm selection showing are probably just the txpower.offset=5 that the UI doesn't known about. I would assume rev 2 is just bug fixes and fully compatible.
Interesting this hasn't popped up before, maybe limited shipments and then XW boards came through. If Ubiquiti has already upgraded to the AR9342 chip in the NSM5 models, this implies the ar7240 rev 2 boards came before it.
"Interesting this hasn't popped up before, maybe limited shipments and then XW boards came through. If Ubiquiti has already upgraded to the AR9342 chip in the NSM5 models, this implies the ar7240 rev 2 boards came before it."
Actually what the cause is is that this is an EARLIER board than what most people tested with. The majority of beta testers bought new hardware from authorized resellers, this means they were all recent boards, this production change happened some time in the past and is actually 2+ hardware revisions ago trick is most of us have been buying new which is why this doesn't come up that often (XW=Current, XM that most people have, and than these devices XM predating the majority of XM devices beta team had)
I'm sure we will see more of this pop up from ebay sales and similar as gear gets cycled out of its environment to be updated with new hardware as part of a hardware refresh.
Added as an "in-testing" support device in changeset:f9152254595301cd6f43a9e82a505be3bd76267c/aredn_ar71xx
Should be in 3.15.1.0b02 when it comes out to test with.
Make sure to let us know how it turns out when its published so we know to promote it to supported.
Two of the older NS M5's (mine and Rick VE3CVG's) with 1b02 have been able to mesh with other
5.8G nodes, on one of the extended channels (5910MHz). Without formal testing, I'd say the
problems appear fixed. Great work and responsiveness to this somewhat unexpected issue!
Is there anything else we should be checking, other than the obvious setup issues in 1b01 ?
Dave