Get four used connectorized P10 2450APs from SWG or PDM. Put them on
90-degree v-pol sectors. Two channels. I have done a ton of 900
sectorization this way, like at sites where we never expected to have
more than 10 SMs now have 40-50, and on 900... you get the idea. At
least with FSK it has a chance of working. There are very few places we
can use 2.4 anymore. Have you thought about moving to 5.7? Even just a
5.7 AP or two for those 10 subs that are having issues?
On 12/1/2014 6:32 PM, Heith Petersen via Af wrote:
I am pretty confident that I am having an interference issue. I could
isolate customers that were bad and customers that were good. I
couldn�t log into a radio that rendered a bad link test, however I did
not get a chance to visit a customer in person and run the test that
way, like say while it was having the throughput issue. Out of the 39
subs I have, I planned on moving 10 to 12 of them to different tower
sites. Some would take remounts and cable runs, others would take
wireless repeater from different structures on their property. I did
in the mean time swap to a different, odd channel, tweak down power
even more on some of the closer subs, and was able to kind of
stabilize the troublesome 10 or so customers, but of course the issue
still exists.
Needless to say that I am dealing with an Omni. After 11 years of
using Canopy we just recently performed our first Omni to Directional
conversion on a tower with 70 Canopy subs. I was having very little RF
issues, mostly just needed to break down the users for bandwidth. I
had to do a UBNT conversion from one Rocket Omni to 3 Rocket panels a
year ago due to a competitor. One of the panels I later had to swap to
a 365 due to the issues, however I am able to maintain solid RF
control and decent bandwidth to my subs. Any UBNT site over the last 3
years we just went right to doing sectors.
Question is do I look at doing sectors here? I understand the benefits
of doing it, I am just afraid that if the issue is around and its
nothing that I can control that I will have a sector with continuing
issues. With 40 customers nowadays that�s a decent load for a Canopy,
just trying to justify costs. More justifiable for 70 subs over 40,
but if its got to be done, I got to keep the customers happy, or just
keep them period.
I appreciate the input.
Heith
*From:*Af [mailto:[email protected]] *On Behalf Of *Heith Petersen
via Af
*Sent:* Friday, November 28, 2014 2:52 PM
*To:* [email protected]
*Subject:* Re: [AFMUG] Suggestions on Canopy PtMP Issue
This is also in response to Gino. 1 AP. I added several color codes
to all the subs for several combinations of allowed subs. So I would
select a CC and maybe 8 subs would come on. Then broke those to 4 & 2
total subs were the only subs associated. If I thought I found the
perp I would set him totally out to a dif color code so he couldn�t
associate with AP. I did that with several subs. But I would pick 4
users (meaning they were the only subs associated) then pick another 4
and would have the same results.
On the 1x/2x I just scanned my session list. All users are 1x/1x. My
AP is set to 1x/2x, I want to say, just by looking at a few subs, that
they are all coded for 1x in the subs.
I am seeing very few re-regs from the subs. And they are all passing
data with link tests above 90%.
I am contemplating sectorizing this AP here this spring. I have had
some of these Canopy APs with 70 subs before sectoring and haven�t
seen this issue. I have been trying to run torch via our Tik core
router here recently. I am not running any bandwidth shaping via
protocol here, just throttling via SM, which I want to avoid. Maybe we
have created a monster due to poor design.
*From:*Af [mailto:[email protected]] *On Behalf Of *Ken Hohhof via Af
*Sent:* Friday, November 28, 2014 2:30 PM
*To:* [email protected] <mailto:[email protected]>
*Subject:* Re: [AFMUG] Suggestions on Canopy PtMP Issue
Can you clarify what you mean by segregating the subs to multiple
color codes? Are you setting a new secondary color code on just a few
SMs and then switching the AP color code so those few register and the
remainder of the 40 cannot? Or do you have multiple APs at the site
and you are using color codes to force most of the SMs to another AP?
What up and down modulation are you seeing? Are most of the SMs at 2X
both directions? What do you get on linktests, are they at least 90%?
Like George says, run spectrum analyzer, also run remote SA on some
subs in various directions.
What about upstream traffic? If you are saturating the upstream
(perhaps a subscriber running Bit Torrent or cloud backup or infected
with a spambot) you could run out of upstream capacity or control
slots, and downstream capacity will go unused because requests and
acks cannot get back up the pipe.
In the 2.4 band, interference is always the first thing I think of.
Interference at the AP will screw up everyone�s upstream and could
cause the symptoms you are seeing. If it�s another Canopy operator,
you might be able to sync with them. Otherwise, it�s look for a
clearer channel. We used to avoid 2435 like the plague, but these
days it can actually be the clearest channel, although some customer
routers might need reprogramming if like us you have them manually set
them to channel 6.
*From:*George Skorup (Cyber Broadcasting) via Af <mailto:[email protected]>
*Sent:*Friday, November 28, 2014 2:09 PM
*To:*[email protected] <mailto:[email protected]>
*Subject:*Re: [AFMUG] Suggestions on Canopy PtMP Issue
Ethernet/CRC errors? What do linktests to the SMs return? Is it
possible you have SM to SM traffic on the same AP overloading it?
Enable the throughput monitoring tool on your APs.
Besides all that, what about interference, have you ran the spectrum
analyzer? This is a two-way street. You can have noise at an SM which
will get it very crappy link efficiency. Like their router wifi
channel is the same frequency (or close to it) that the SM is trying
to listen on.
On 11/28/2014 1:50 PM, Heith Petersen via Af wrote:
Earlier I posted that I was having some through put issues. I have
a 2.4 AP with 40 subs. I started having issues where customers
were having horrific ping times on the AP, the majority anyways. I
would still have customers that were stellar most of the time.
Firmwares were all over the place, all on the same level as well
as AP with the same issue. I had segregated the subs to multiple
Color Codes, as times I had as few as 1 to 4 subs registered at a
time. If I swapped color codes to a different code and allowed 4
different users on, at times the issue would follow them as well.
�
I had a tech swap out the AP 2 evenings ago, and to my dismay I
noticed the issue right away. Today I re-routed a back up back
haul to that site and the AP is the only device pulling service
from that BH. I have 2 more sites that feed off of this tower and
they are not experiencing the issue. This issue crept up on us
last Friday night. I have only heard from a handful of users so I
don�t know if everyone is experiencing the issue, or at least
noticed. For the most part its really bad at time during prime
time, however today and during the day Monday it was real
noticeable. From what I can tell on my MikroTik is that its really
only handing out about 3 meg right now. This network is bridged, I
have some Tiks in place to start routing and also for management
purposes. I had the same results from just using a standard
switch. I am not experiencing any other issues on that network.
All radios, except for 4, are NAT enabled. The other 4 at times I
would have totally disabled or not associated with the same results
�
We were suspecting a bad Canopy radio mucking up the AP, which we
have seen but its rare. But like I said we have seen it with only
a few users associated. I cant imagine that I had several bad
radios at once. The AP is swapped out. Could it be a network
issue? In all my years of using Canopy I have never seen this
issue. I fear that my BS skills may be wearing thin on my
customers, but I am unsure of where to look
�
Thanks
Heith
�
