Troubleshooting Bridge Link
Last Post: September 9, 2010:
I've built a wireless bridge between two buildings located across the street from one another (about 600 feet apart). using the 5 GHz radios on a pair of Cisco 1242 APs. The antennas are TESSCO 15 dBi Yagi (484982). There is 50 ft of coax (LMR400) between the antenna and the radio.
My problem is that the signal strength is extremely low with everything connected. Approximately, -82 dB. Based on my calculations, the signal strength should be in the upper 50s to low 60s. Calculations shown below.
The 1242 has a receive sensitivity of -88 dB @ 6 Mbps and -73 dB @ 54 Mbps. 50 feet of LMR400 has a loss of 3.5 dB, plus approx 1 dB loss for for the ends (.5 dB each). The antenna has a 3 foot pigtail (approx 1 dB loss). So that's roughly 5.5 dB of loss through cables and connectors on each side. Round up to 6 dB.
Tx Power (17 dBm) - Tx Loss (6 dB) + Tx Gain (15 dBi) - Free Space Loss (5.8 GHz @ 600 feet = approx 92 dB) + Rx Gain (15 dBi) - Rx Loss (6 dB) = -57 dBm
Quite a difference from the actual. During my initial visit, I decided to take the coax out of the loop and ran a length of Ethernet cable to the roof on the non-root side of the bridge and connected the antenna directly to the radio. Signal strength improved to -78 dB and a SNR of 18/19. At this time I noticed the connector on the non-root cable was damaged. I patched it as best I could and moved over to the root bridge side. I verified the link was still at -82 dB as it was before. I replaced the coax with a length of Ethernet and connected the AP directly to the antenna. I had a signal strength of -64 dB and a SNR over 25. My diagnosis at the time was that the piece of coax on the root side was bad because I had a measured loss of over 20 dB when there should have only been a loss of 7 to 8 dB max. I also decided to replace the non-root cable since the connector was faulty. I left the root AP directly connected to the radio.
The customer had issues during the time that the cables were being shipped. When I arrived onsite to replace the cables, the signal strength on the root bridge side was -82 dB with a SNR of 12/13. Thinking that the connector on the non-root side had pulled apart, I replaced that cable first. When finished the signal strength was -72 dB with a SNR of over 20, which is better. However, I was expecting much stronger since when I left the last time, the signal strength was -64 with a SNR over 25. I went back to the root side and disconnected and reconnected the antenna and got a signal of -63 dB. However, when I introduced the piece of 50 foot coax, the signal dropped to -78 dB and directly connecting the AP to the antenna only yielded a signal of -70 dB with a SNR of 23.
This definitely appears to be an RF issue, I?m not sure if I have an additional hardware problem such as antennas/pigtails or if the mounting location of one or both of the antennas is causing some strange multipath issues. As close as the two buildings are, antenna alignment should not be an issue.
I thought with multipath issues, the signal strength jumps around quite a bit. I observed the bridge for over 30 minutes and the signal strength was consistent. One antenna is mounted on a tripod on the roof of the building about 25 feet in the air and the root antenna is mounted to the side of a building about 3 to 5 feet off the building and 20 feet in the air. This building does have metal siding along the area where the antenna is mounted.
Any insight would be greatly appreciated.
Lots of questions. I'll start with:
1. Who built your cables, and do they know what they are doing? Did you check the Pin Heights on the connectors? Were the connectors and cable in good shape? How about the radius of any bends in the coax?
2. Did you try using a spectrum analyser before you "switched on"?
3. Metal sides - How about metal walls behind your antennas? Are you creating nulls at your antenna with their reflections?
4. Are you using your eyeball to align the antennas, or something quantitative?
5. According to the [u]actual[/u] Tessco spec. page for that antenna, it's built for 802.11a channels 140 and above. That's 5.725 to 5.850 GHz. Is that the channel range you are using?
1. Cisco cables - AIR-CAB050LL-R. Original cables that were swapped out were Tessco cables.
2. Yes, according the Spectrum Expert, the 5 GHz spectrum, specifically the UNII-3 bands are clear. Not many buildings/homes in this area.
3. The root antenna has a sheet metal wall approximately 3 to 5 feet from the antenna. The antenna is almost perpendicular to the wall. Maybe 10 degree angle. My thought was if I could achieve a stable -64 dB, which I did after the first visit, that chances were good that it wasn't RF due to reflection. I very well could be wrong.
4. For 600 feet, I am eyeballing the antennas. The antennas have a 25 degree beamwidth, eyeballing it should be good enough.
5. Yes, using the UNII-3 channels.
First, is my math sound. Are my expectations of signal strength and SNR of the proposed link correct?
Second, my first thought was to rule out faulty equipment, which I thought I had done with my first trip. BTW, there are no severe bends in the cables. The root side runs from the antenna through a hole in the side of the building with a gentle curve and then a straight shot from there another 30 feet to the wiring closet. The non-root side is a straight shot through the roof to the wiring closet below about 25 feet. All the connectors look good. The only thing I haven't swapped out is the antennas and pigtails.
I've definitely entertained the idea that it could be an RF issue due to reflection/multipath. Just trying to figure out the best way to prove or disprove this theory.
Very appreciative of your time and input on this matter. Thanks.
I don't have to do the math to know that you have plenty of signal potential.
Check the polarity of your connectors. I know it may sound trivial, but make sure there is a "male" piece going into a "female" piece. With so many reverse polarity connectors out there, this mistake could easily occur.
Oh, one more thing. Have you selected either the left or the right antenna connector in the software and is the antenna connected to the appropriate connector?
Both radios are set to use the right antenna and both are connected to the correct side. I've accidentally done that before, so it's usually on my list of dumb things to check. :) I'm pretty sure the connector is correct, but unfortunately the client is over an hour away. Not something I can just run down and check over lunch time.
Trying to come up with a game plan of things I can check, or tests I can perform while on-site next time to either isolate or eliminate the problem. So I'll add check cable connector's polarity to the list, but I need additional tests. Somehow I need to move the root antenna (without putting another hole in the customer's building) to determine if reflection/multipath is an issue. I'll probably try getting a new antenna to eliminate that as a point of failure.
Any other ideas? Thanks for all the help.
Depending on how ?long? a roof you have, you can put the two radio systems facing each other on the same roof. Assuming no obstacles in the way and having appropriate in-line attenuation. This will help rule out path effects. This may or may not be possible in your case. I?ve done that many times. You can make two simple portable antenna mounts ( poles with flat bottom plates ) and use them for this type of testing. One thing I would suggest in future ( this depends on the physical layout of your work place ) is to try and do a 24 hour ?soak test? with two radios facing each other and using as much of the final kit as you are going to install. You would be surprised to see how much time this can save you in the long run.
In my last job, I was tasked with cutting down the number of ?failed installs?. I insisted that every piece of kit that went out the door had been tested in our ?lab? ( nothing more than a workshop with a wide roof ). A lab can be a simple setup or as complex as finances dictate.
This was greeted with great sneering by the accountants and many of the engineers at the first meeting.
Within a year the number of failed installs had dropped by over 70%.
Waterproofing ( tiny amounts of water vapor can cause large losses )??
Bye the way, interference can sometimes do strange things to RSSI readings. You might imagine that inteference would just affect the noise floor and hence SNR. Without going into all the details, depending upon the type of signal ( remember in the unlicensed band there are all sorts of non-wi-fi signals ), strange things can happen to the RSSI values.
You seem to have "block" drops in level, so the other factors mentioned should be looked at first.
The actual mounting hardware should be checked and re-pointing should be done. Without a diagram it's a bit difficult to see, but depending upon how far off from boresight each antenna is, the rate of change of signal level with respect to angular error ( dP/dTheta ) varies dramatically close in to the boresight and farther away ( still on main lobe ).
If you are de-pointed, small amounts of wind or vibration from traffic ( with not-too-tight hardware ) can cause quite severe drops.
Correct pointing is very important.