Mounting orientation of 1920

[joshintheco]

I am currently experiencing range limitations issues with Bluesocket 1920 APs that are vertically (wall) mounted.  Does the mounting orientation of a Bluesocket 1920 (internal PIFA antenna array) have any impact on reception?

The quick start guide gives instructions to mount a Bluesocket 1920 or 1925 to a ceiling or wall; however, I didn't know if this was applicable for both the 1920 and 1925.  With a 1925, obviously, it would be trivial to change the orientation of the rubber duckies.

[joshintheco]

Bump

[jayh]

Two considerations here, antenna pattern and polarization.

A dipole antenna or "rubber duckie" has a pattern that looks kind of like a doughnut with the antenna stuck through the hole.  That is, it radiates equally well in all directions perpendicular to the antenna but has very little coverage off of the end.  The internal antennas are PIFA, and the pattern is dependent on the internal design of the unit.  I don't think Adtran has polar charts on the data sheets showing antenna patterns.  PIFAs are versatile and can be designed for a variety of patterns.  More info here:

http://www.qsl.net/va3iul/Antenna/PIFA/PIFA_Planar_Inverted_F_Antenna.pdf

Polarization depends on the receiving device, turning a smartphone sideways will change this.  You really can't control what your users will use but I'd mount the units with the labels reading correctly rather than sideways.  This will be vertical and match the majority of users.  A specific site survey is your best option, but slightly above head height for a small room and closer to the ceiling for a large room are good starting points.  About where you would mount a wall clock.

[joshintheco]

Thanks Jayh.  The lack of polar charts available for the 1920's is what brought me to the forums.  My APs are located along a hallway and my hope is to serve the multi-units off that hallway - it's a hotel.  The units were installed vertically in the plenum.  I'm finding that the reception in the rooms leaves much to be desired.  If I can get hard data to show that the antennas would perform better if they were mounted horizontal (still above the plenum), then it would be worth the added work of re-mounting the access points.

[Anonymous]

When I think of plenum, I think of the space above the ceiling with active air flow. So I assume these devices are mounted near metal air conditioning ducts. That could be causing interference issues, so it may be that the mounting orientation is bad not, but rather just that placement of the APs is bad. But you should definitely consider mounting the APs with the logo pointing down.

And you mentioned mounting the APs in the hallway. You might consider the mounting the APs in a zig-zag pattern in every other room. For example, say your hallway runs north to south. Then you put an AP in the first room on the west, move to next set of rooms, and mount an AP in the second room on the east. That pattern would possibly provide a better coverage area for the hotel patrons.

Just start off mounting the APs with the logo facing down, and see if that helps. If not, consider the placement suggestion, and then you might also want to consider a site survey.

Also, could you describe the mounting in more detail, or possibly post pictures and a map/layout of the floor and APs? I want to be sure that my suggestions are not something you have already done.

[Anonymous]

I believe Josh is dealing with concrete interior walls as well. Concrete has high attenuation for Wifi.

another point, and this is just a general comment, but mounting immediately next to a concrete wall can be worse than backing away.

Two reasons.

• going through a wall at an angle increases the distance through the attenuating material. so if the destination on the other side of the wall is off on an angle from where the AP is, the attenuation is worse. Backing the AP away from the wall MAY help improve the angle.
• backing away from the wall also gives the devices a chance to find better paths through the wall (thinner spots, doorways, avoiding pillars, less rebar...)

Of course the compromise is that backing away also impacts attenuation.

and some ceiling tiles and insulation have metal foil built in, also a source of attenuation

-Glen

[jayh]

A bit more description would help.  "Plenum" to me implies metal ductwork, although I've heard others use the term to describe the area above a drop ceiling.  The other suggestions are good for zigzag in the hotel corridor, mounted vertically on the wall.  I'd be more inclined to mount them directly below the drop ceiling so that they're visible in the hallway unless there is a huge cosmetic objection.  This reduces attenuation due to the ductwork inside the space above and also brings the typical user device closer to the main beam of the antenna rather than below it.

Are the interior walls concrete or drywall, doors wood or metal?  Concrete and metal doors will be the worst possible combination.

Back in the day when I installed two-way radio systems in buildings such as hospitals there was a form of "leaky cable" called Radiax.  It is basically a coaxial cable with a very loose or slotted outer shield, and is used as a continuous antenna.  This works well in that type of environment at VHF/UHF but I haven't heard of it being used for wi-fi.  Perhaps there are some issues with the higher frequencies involved, but it might be worth exploring.  Essentially you attach one end to the radio and string it along the area to be covered, put either an antenna at the other end or a 50-ohm terminator depending on the scenario.

[Anonymous]

You should never use a DAS or DAS-like antenna system with 802.11 technologies. The reason is that the CSMA/CA access method defined in 802.11 works best when clients can hear each other. With DAS, there is potential that clients can be located very far from each other, and the physical or virtual carrier sense mechanisms in the CSMA/CA process will not work. A DAS almost always causes hidden node problems, and should be avoided at all costs.

[jayh]

OK, scratch the Radiax idea then. 

[joshintheco]

I tested vertical vs. horizontal orientation in my lab with a 1920.  The AP was mounted vertically to a concrete wall and measurements were taking using 'WiFi Analyzer' on my Samsung S3 in a room twice removed (two concrete walls).  The AP was then mounted horizontal at a height of approximately 2.5m and again measurements were taken from the same location.  Mounting horizontal resulted in approximately 3dBm gain.  This answers my initial question of whether or not they are meant to be installed vertical.

Then I took the show on the road and went on-site at the customer location.  First, to clarify, when I said 'plenum' I actually meant space above the sheet rock ceiling.  This area is 30cm in height and has cabling, water pipes, ceiling hanging hardware and wire, etc.  The APs are accessible via small metal access doors in the ceiling.  Contrary to what I initially thought, most APs are actually mounted horizontal using zip-ties to the insulation water pipes.  The space is very tight above the ceiling so our installers did what they could.  Mounting the APs to the actual ceiling, so that they are visible, is not an option due to aesthetic restraints imposed by the customer. 

I compared a 1920 with a 1935 with dipole antennas.  I tested with the 1935 in the same place as the 1920 (1) as well as exposed at ceiling height (2).  All dividing walls are concrete and the room door is very heavy wood.  The following are my findings.  Again, I used WiFi Analyzer on my Samsung S3.  I realise that there are better ways to calculate signal strength, but the majority of end users are using a similar mobile device, so I thought it good enough for testing.

                               1920      1935 (1)      1935 (2)

  Inside closed door      -48       -42              >-40     

  Midway                     -60       -54              >-40

  Far end of room         -76        -62              -48

The meter sensitivity starts at -40dBm.

So, not only does the 1935 have considerable gains over a 1920, but the above-ceiling mounting placement is really attenuating the signal. 

That being said, we current had one AP allocated to serve up to six rooms (AP in the hallway in the middle between three rooms on either side).  The AP was directly in front of the room I was testing in.  As I moved over to the furthest room, the signal was unusable at the far end of the room.  We would need to double the number of APs installed in the hallways.  At this point, the real solution is to put third-party in-wall APs in the rooms and use the 'unified access' feature on the vWLAN controller.

Adtran should introduce the Motorola 6511 into their product line.  Extreme Networks is also re-branding this unit. 

[Anonymous]

You should note the 1930 series is a 3x3:3 device meaning it has 3 RX antennas, 3 TX antennas, and it can support 3 spacial streams. The 1920 series is only 2x2:2. So the 1930 has more antennas to utilize spacial diversity meaning it can better handle the negative effects of multipath. The 1935 can also support higher data rates assuming the client devices support 802.11n and are also capable of 3 spacial streams; most consumer grade devices do not support 3 spacial streams. You should expect better performance out of the 1930 series APs.

Our Unified User Access feature allows clients on 3rd party APs to have the same experience as clients on Bluesocket APs which limits the need for us to OEM an AP from another vendor. If you have autonomous APs or an existing wireless infrastructure and controller, then you can simply leverage that through Unified User Access.

You really should consider the zig-zag pattern described above as opposed to lining all the APs down the hallway, and This might even limit the need for an AP in every room. At the very least, I would definitely suggest a professional site survey prior to installing new APs in every room.

[joshintheco]

I realise that it's not an apples-to-apples comparison but it was what we had to work with.  Hopefully the information can help someone else.  Our investigation was to look at replacing the 1920 with a 1925, so if a 1935 can't do it, then we won't bother looking at the 1925.  Unfortunately, relocating the 1920s to the rooms is not an option.  We'll be using Tamograph with a Wi-Spy to do the site survey.

I appreciate that Adtran added the unified access feature - we would be in a bad place without it.

[jayh]

With those numbers, it sure looks like the best solution is to put them below the sheetrock ceiling away from all of the pipes and the like.  Maybe some kind of non-conductive decorative cover to appease the cosmetic issue.  I'd think that a hotel environment would have visible equipment in the hallways such as smoke detectors, fire alarms, exit signs, sprinklers and the like where access points wouldn't be particularly obtrusive.