Strange Session count problem with 900mhz system

I have a strange problem with two connections off two seperate AP’s in different locations. Both appear to be excellent connections RSSI around 2300 jitter of < 2 power level < -50. But their session and reg counts climb continuously. No other connections on either AP seem to have any problems.

The only thing these two have in common is the distance from the AP. They are 440 and 445 feet from the AP.

Does anyone know if its possible it could be a wavelength issue with 900mhz.

There are connections that are closer and further from these AP’s with no problems. The connections have had all the level 1 possibilities looked at but nothing has helped.

I would try lowering your power levels to start with thats awful hot.

is it pointing over a metal roof?

They both have direct line of sight but one is on a metal roof. We did try pulling back so that the power levels were around -60 but it didnt seem to make any difference. I will pull them back again though. What sort of power level should I be aiming for?

The strange thing is they might be fine for a few hours than jump up by 50 within half an hour.

is the one on the metal roof aiming over the top of the roof or away from the metal roof?

cause i have had to relocate my antenna many times due to metal roofs

also correct me if i am wrong guys but would it not be a good idea to have all his sm’s set at -60 to -65 as long as it is 10 db under the noise floor.

you’ve checked for sources of interference? 900mhz cordless phones or possibly baby monitors? we’ve had quite a few issues with both of those circumstances.


I had a similar issue with a 900 cluster. The only way i could get it to work properly for those close people was to have them associate with an AP in the higher end of the frequency range. Since they were so close to the tower, that was easy to do (they could see almost all the APs).

Not sure if moving the SM to a higher frequency AP is a universal solution for the problem in question: It might have been specific interference at our site that dictated which frequency worked best for those clients close in.

Still . . . grist for the mill.

You might try lowering the power levels on all SM’s and AP’s.

Everything is very close and you are getting alot of high level multipath, especially with metal buildings in the area. If you were able to view the RF spectrum from the radios’s perspective, it would look like a huge cloud with no discernable source.

If you don’t know what the RF background noise was to begin with, pick a safe low end like -70dB. Start lowing each SM’s output until they all fall between -70 and -60 at the AP sessions page.

Once you have the SM transmit levels balanced on the AP sessions page, look at a number of your SM’s status pages and see how much signal they are getting from the AP’s. Pick the weakest SM’s and see how much above -70 they are. Start lowering the AP level in 1-3 dB increments until the SM’s see something closer to -70dB and higher. A large variation from SM to SM is not a problem, just the overall level.

This will quiet things down.

The metal roof thing is brutal. You need to make sure you get all the way to the end of the building - don’t try to look over it unless you elevate the SM about 20’

The AP’s should also be elevated 20’ above any metal roof.

I have lowered the power level so that on the sm side its now showing -64. But is it the cae that this only effects the transmit power and the recieve power is set through the AP ? If this is true i cannot lower the AP transmit power as there are some connections that are a fair distance.

I tried replacing the 7db yagi antenna with a 7db panel and on the metal roof connection it seems to have totally settled down. I have been watching it for a day and a half now and no problems. I thought the panel would struggle more with multipath but it has definately helped. I tried the same thing on the second connection today and will let you know how it goes.

If you have distant SM’s, you would only want to lower the AP enought to keep the weakest SM registered (and then maybe not at all - I thought that this was a small network.)

The SM power levels as read at the AP’s sessions page are more important. Here’s why:

The SM only needs to be able to “see” one RF source - the AP. As long as that level is continually +3dB above the noise floor, then the SM will register with the AP. The “continually” part is important as the noise floor will breathe throughout the day. This is why it’s recommended that you have a +10dB margin to allow for variances in the RF level. It does not really matter if one SM’s status page shows -42dB, and another’s shows -72dB.

The AP needs to be able to see many RF sources. If one SM is hitting the AP with a much higher power level than the other SM’s the AP will not be able to see the weaker SM’s (referred to as De-Sense). Lowering the individual SM power levels so that they all fall within a 10dB window improves the AP’s ability to see all of the SM’s. The closer all of the SM’s power levels are to each other in the AP’s session page the better the AP will perform.

AP transmit power affects SM status page value
SM transmit power affects AP sessions page value

SM status page = not important as long as it’s ~10dB above your noise floor
AP sessions page - each SM’s level should fall within 10dB of each other with the optimum being all SM’s exactly the same AND +10dB above the noise floor.

Interesting that the 7dB panel works better than the Yagi, but hey some of this is pure magic.