Our old system just got hit by lightning and we’re running a totally new system (Except for the SM’s obviously)…We’ve got a new CMM/AP’s x6/GPS/2 brand new backhauls.
Most of our links have 1600+ RSSI and 1-3 Jitter. They were all running at 99.8% uptime previously and now they’re running @ ~90%. After being at a few customers locations I’ve noticed that the SM’s will randomly drop off and rescan (which was somewhat normal) but they’re doing it more often now than before. From the looks of everything - all the configurations are the same as before. (I saved configs into a .doc file previously)
All AP’s/SM’s are running HWS w/ 2x disabled and 7.2.9 - Max range is 14 miles on the AP - our longest shot is currently 10 miles and change. We aren’t using Prizm/BAM and AP’s are set to SM as configuration source
It seems that when one person drops off, multiple people do. Sometimes even every SM on the AP which makes me think its a timing/sync issue so I’m looking into that now. Even when all the customers on the AP drop off I can still ping the AP with 8ms ping consistantly. When my customers are up my ping times are only 13-15ms to them.
Noticed something else odd…when I go to the AP eval data page it shows all my customers that were registering fine to now have power levels of -75 through -81. Which is weird considering some of them have 1500+ RSSI and were working perfectly fine before.
The customer above is on another AP and has 82% availability (according to intermapper, not polling SMTP by the way) and 41% packet loss. Some other customers have 100% availability with 0% packetloss, others have 100% availability with 20% packetloss, etc. Kind of weird?
So basically I have two problems. 1 AP that customers are dropping off of consistently and another AP with some SM’s having packet loss and others not. I had a ton of registration problems but I think I fixed that as most of the time now when an SM is rescanning it only takes 21 seconds instead of 1minute+. It’s just annoying to my customers if they’re using VOIP etc. and the module rescans. I know it’s going to happen but it’s happening 100x more now than with the system before (that -looks- the same)
According to the tower guys they moved the AP over a degree or two to align it where it should’ve been (Due North etc.) To me I would think over the course of a few miles this would throw it off.
What’s the next step to get it right however? They’re halfway across the state and to get them to come back and reclimb we’re looking at an easy 3 grand. Is there any other solution?
The coverage radius isn’t wide enough to just pick it up? Since the tower hasn’t moved I would’ve thought that since we have it pointed at the tower, as long as the signal from the SM is in the 60 degree radius from the AP that it would still pick it up?
That’s with the other AP’s on so I know it doesn’t give a 100% accurate measurement. The frequencies we’re using are 906/915/924.
For some reason the SM’s were rescanning every 1-2minutes. Now they haven’t rescanned in over 30 minutes. Wtf?
Edit: Yeah I tried rescanning AP’s etc. they were all on the right AP still.
What I don’t get is why the sudden all the SM’s stopped dropping. Maybe just luck of the draw with interference or something? Now I’m monitoring my intermapper to figure out why there is any packetloss (Assuming it’s reporting correct)
I need to do the color codes - just haven’t done it yet. It makes more sense than the way I currently have it.
I have the SM’s locked into the frequency they’re on. So when I need to rescan frequencies I have to readd the other 2 frequencies and reboot, etc. It’s a pain compared to color codes. Also I’ve experienced SM’s logging into the AP that is directly behind the AP I want them to register onto.
All of your symptoms point to interference. I was thinking self interference but it seems the problem has cleared up so it looks like either the interference went away, or Canopy adjusted to it (the magic of Canopy).
FYI - re-registering is not normal operation. You may have had self interference from the beginning and just worked around it. Installing all new AP’s just exaggerated the problem. The packet loss might still be a symptom.
Here is how I would diagnose: - Turn off all AP’s except one. - Do a spectrum analysis and make sure you are not trying to use a noisy frequency - Run link tests and ping tests to several SM’s on that AP. - Repeat for each AP to get a baseline (1 AP on at a time) - Start adding AP’s one at a time and retest - When you start seeing re-registering SM’s or packet loss that’s your first clue.
What are your channel assignments? Are your BH’s in the same band as the AP’s? Check that all AP’s and BH Master are receiving sync 100% of the time (possible flaky GPS)
As clueless mentioned, in general you never want to change alignments once a system has been up and running. There are enough mysteries with RF that you don’t want to add more mystery…
with that said, that is probably not your problem.
I don’t know if how we operate is correct. is there a better way propably.
We check all freq in the sm: incase we need to make a change we don’t haft to go to each sm and change the freq.
My Ap’s have a different color code to keep the sm’s locked to a certain one.
Not knowing how far your ap’s were actually turned from their original direction, and the problem started after the change.
I would check the ap aval in a few of the lowest pwr level sm’s. I bet they turned them further than you think. Or mis labeled them and the Ap you thing is north west may be north east
I would check the ap aval in a few of the lowest pwr level sm's. I bet they turned them further than you think. Or mis labeled them and the Ap you thing is north west may be north east
I'm tellin ya, you need to change your name to Not-So-Clueless.
We set up the same way: - Each AP get's it's own color code - All SM's can scan all frequencies
At the install, we use the AP eval data to pick the best AP and set the color code for that AP.
This way all the SM's follow the color code regardless of the frequency.
I have a set of both, at different locations of course
That’s correct we started at 57,67,77,87,97ect
backhauls are 5,10,15,20,25…ect
Our Ip address are fixed with the color code 172.17.5.52 (cc code 57) or 172.17.6.52 (cc code 67) This helps me locate a trouble customer I Know which Ap they are on. So on the Customers ledger in Notes we have CC-57, last two numbers of the Sm and the last numbers of the ip address
Doesn’t matter what you use as long as each AP has it’s own unique number.
the color code only applies in the band so you could have 10, 20, 30, 40, 50, 60 in 5.7GHz and the same color color codes in 2.4GHz and again in 900MHz
I’m an idiot. There are just some days you’re better off stepping away and coming back to things.
I walk away and come back thinking…God this -HAS- to be some sort of sync issue if it’s not an alignment issue.
Sure enough I check the AP and for some STUPID reason it’s set to generate it’s own sync instead of receiving sync on the power port.
AFK hanging myself.
At any rate I got the issue with them rescanning fixed…but now I have to figure out why the power levels are so low. That may require rolling a truck but we’ll see if anyone still drops off as often now.
Edit: Applying color codes at the moment - I’ve been meaning to do this for a while now.
