ePMP 2000 question from webinar about the WISP test

Guy on webinar says he saw an approx 10db stronger signal level when the 1000 AP was swapped with a 2000 AP.  I'd like to know how that is possible?  I only imagine one of three things - new software allowed him to run stronger tx than what is legally allowed (a bug / pre-release issue), the new antennas are that much better (I thought he used same antenna though - and remember this was improvement on client side I think so beam-forming would not have healped), or the tx power on the new 2000 platform is higher.

It is possible he meant the client signals - on the AP in shich case the beamforming could have improved the signal of each but I would have expected the newer technology to reduce other noise rather than improve the signal by 10db.

Either way, just looking for how any receive level could have improved by 10db.....    Thanks

The EPMP-2000-AP ( E2K ) AP has only the better "Adjacent" channel filtering.

The really diffrerent thing is the beam-steering antenna.

The beam-forming operates in the RX of the AP side, therefore not affecting the maximum-regulatory permitted power both at the AP and the stations. 

the +10db signal is obtained by changing the antenna-array characteristics for every recieve frame, making the AP listen more directionally  (the signal of the specific client is improved at that frame-time).


0- AP that transmits via the traditional sector antenna, at the same TX power we're used to today;

1- The clients recieve the signal at the same levels seen today;

2- Clients reply at the same regulatory power they can today;

3- AP listens via special beam-steering antenna, tells the antenna where to listen from, picking more signal from that region in space in that time-frame. 

I get that, but I do not believe the "smart" antenna has 10db more  gain than the primary tx sector.  Therefore how would it show up 10bd stronger?

@carullos wrote:

I get that, but I do not believe the "smart" antenna has 10db more  gain than the primary tx sector.  Therefore how would it show up 10bd stronger?

Patened Cambium 'Magic Sauce' :-)

I hope its got lots of the sauce and have my fingers crossed....

"Gain" is the same as "Ability to concentrate signal to/from"

A parabollic Dish is the same for RF as a parabollic mirror is for light.

A element-array antenna  is a bunch of little squares in a matrix, that pick the signal, and a bunch of transmission-lines that are size/impedance matched so the bunch of signals that were picked by each element arrives at the feed-point (SMA / N ) connector in coherent phases, therefore summing equally the signal picked by each element (and with antennas, whatever is true for RX, is true for TX).

this has the disadvantage that when your signal comes at an angle relative to the array , the signals are summed partially out-of-phase, this imperfect phase-matching results in lower RX / TX signal from the direction / AngleX,AngleY that gives a certain phase-delay (that is different for each element and polarization V/H in the array, because each element is in a different piece of space, and the signal/ "wave" arrives first at one element, and last at another ). 

A beam-steering array is when you add an adjustable delay to each transmission line, and create the inteligence to adjust these delays in a way that makes the phases of the signals from each element cohere for every AngleX;AngleY.   

You store those delays in a table, and since the EPMP running in TDD mode is a TDMA system, the AP knows when in time each client is transmitting, so it tells the antenna to load from the table the delays necessary to make the array phase-matched to the direction of the signal of that client. 

The antenna now behaves as the center of its beam is pointed exactly at every client, but only during the periods of time when the AP expects to hear from that specific client/direction.

Check the animation at this page: (and read the math if you're into RF-VooDoo-math)



Thank you for the info Guilherme I appreciate your time.  I did not realize it worked that way.