We are in the process of installing Canopy for use with our SCADA system and need help fixing registration (interference) problems. We have (3) 900MHz AP’s and (1) 5.7GHz BH on a water tower all mounted approximately 90 degrees apart (looking down on the water tower) at 120’ on the hand rail. The BH link is working as expected but we are having trouble with AP’s. We are using advantage AP’s with built-in filters, CMM Micros and 7.2.9 firmware on all deployed units.
We tried to register two SM’s to the same AP and had trouble from the beginning. The first SM would register but immediately drop off and re-register; we had over 3000 re-registrations in one night. We then turned scheduling from hardware to software and the SM registered and held. The power on both sides is approximately 72dBm with a jitter around 3-4.
The next radio’s dBm was the same but jitter bounced from 3 – 10 with an average of 4. The radio registered in software scheduling but I couldn’t get SCADA to connect. I then decided to do a link test and received almost perfect download efficiency but nearly 0% upload efficiency. The first SM showed the same results.
The spectrum from the SM view is completely clean while the spectrum on the AP side is in the high to mid 70’s from 910 MHz to 924 MHz; thus I have set the frequency on these units to 906MHz where the spectrum is in the mid 80’s.
My questions are as follow:
1) Does this sound like an inference issue? The first AP we installed on this CMM port had to be powered off before entering the GUI after a reboot; thus I replaced it.
2) With upload efficiency at 0% do you think interference is very close to the AP site?
3) With out moving to a different frequency are there any ways to over come interference? Maybe re-route trouble paths with installing several daisy chains (doesn’t this slow down the speed after several hops)? I’m thinking it might be hard to get LOS at all of our substations with higher frequencies.
4) Do you think the water tower causes interference since the AP’s back side is facing directly to the tower? Would you recommend mounting the radios above the top of the tower?
5) I’m assuming you must use an omni for a 5.7 AP??? How do you connect a dish to receive maximum throughput? What is the maximum range with a 5.7 point-multipoint configuration?
Yes it sounds like interference. You can daisy chain as mutch as you want were are only talking about miliseconds. The connection i am on typing this goes threw 7 diffarent hops to get to the isp with 20 to 30 ms ping times. What kind of antennas are you using on the 900 ap’s internal or external? In general you dont use a omni on 5.7 or a reflector. a reflector on the ap would narrow the 60 deg sector. Cyclone does sell a 5.7 ap with an omni. I have used the 2.4 model. We have modified our own 5.7 bh to use on 4ft dish to get a 120 mile shot. and some 2.4 ap’s for distribution to some sites. If you are using external antennas on the 900ap’s make sure they are horizontal.
We are using internal antennas. Do you have to have a CMM every 2-3 hops for timing??
rconner wrote: We are in the process of installing Canopy for use with our SCADA system and need help fixing registration (interference) problems. We have (3) 900MHz AP's and (1) 5.7GHz BH on a water tower all mounted approximately 90 degrees apart (looking down on the water tower) at 120' on the hand rail. The BH link is working as expected but we are having trouble with AP's. We are using advantage AP's with built-in filters, CMM Micros and 7.2.9 firmware on all deployed units.
I don't know that either of these points will help but your problem sounds a bit like one I had...
All AP's on the same sync timing source which operate on the same band must be set to the same maximum range.
All AP's which are within five miles of each other and are on the same band should be run from the same sync timing source.
One other thing. I don't know this to be correct but I did it as I suspect it may be true. If you distribute your cluster across a wide geographic area (you forward sync), your maximum range should be large enough to encompass the entire area. What I mean is if I have two AP's that are five miles apart and the furthest client from one AP is seven miles and the furthest from the other AP is three miles, I would set my maximum distance to 7+5+3 as a mimimum. Personally I would double that, but I have NO evidence that this thinking is correct.
The manual says 3 hops. I havn’t gone any further. Yes it is good practice to use cmm’s especially if your sites are close together.
micers is correct in that all ap’s must be set to same max range. What the max range does is tell the ap how long to waite for the furthest sm to respond before it transmits again. In a cluster all ap’s need to be synced together and tx at the same time.
another reason it is good to use cmm’s on sites that are close is the same reason. if all sites are synced off the gps together all ap’s at all sites will tx together.
From your description of your spectrum i dont think timing is the main issue. need more info? when you ran the test were all ap’s changed to sm mode or turned off? If not they can give you false readings. What freq are they set at? give me a little more detail of your spectrum analisys? The water tank they are on is it a large cylinder type like a coke can or is it a tank up in the air on top of a tower?
All APs were turned off during spectrum analysis.
The radios are operating at 906 because the rest of the spectrum is around 80dbm.
The water tank is like a donut sitting on top of 5 large legs.
I went to the tower Friday hoping to move the radio around to avoid interference. After some time trying this and failing I decided to try registering with another AP on the tower but was very skeptical due to the water tank shielding the SMs from the AP. This AP registered both SMs with no problem with hardware scheduling @ 2x rate. I tried the bad AP at this location and again received registration but with lower link test results. I then tried the good AP at the original location and again received registration but with higher jitter values.
I have concluded that the tower must be shielding the good APs side from interference and that the original AP must not be as strong as the other. Is this possible???
Not only is this possible it is the norm as far I have been able to determine.
If you go to http://ip-address-of-AP-or-SM/mac.html you will find the real power output of the radio at a configuration setting of 28 and 21. This is the calabration of the radio at the factory. At a setting of 28 I have some radios that only provide an output of 25 and others at over 28.
This means that if you set a power output of 25 in the configuration on a radio that can only produces 25 at a setting of 28, you will get considerable less at the setting of 25.
I don’t know if the power spread is evenly progressive in the change between a setting of 21 and 28, but if it was a radio with a power output of 25.09 at the 28 setting and 18.88 at the 21 setting will give you a actual power output of 22.4 at a configuration setting of 25.
I have this particular radio that shows an output of 25.09 at a configuration of 28.