Crown Castle’s Coverage Map Games

This post gets into technical details regarding existing AT&T coverage in Rancho Palos Verdes.  In it we show that a recently submitted Crown Castle coverage map has a 30 dB discrepancy (that’s off by 1000 times as decibels are logarithmic) that misleadingly depicts existing coverage being much worse than it actually is.  This continues a very troubling pattern in Crown Castle’s submissions.  We’ve tried to make the technical aspects readable to the layman and put the legal implications in context.

 The Importance of Propagation Maps

Rancho Palos Verdes’ comprehensive wireless facility ordinance requires applicants provide propagation maps depicting existing, proposed, and combined/existing proposed wireless service.  Propagation maps are sophisticated computer simulations that depict signal levels and cell tower coverage over a wide geographic area.  They are an industry-standard tool used to assess wireless service coverage.

att_palos_verdes_propagation_map

AT&T propagation map example, submitted to the City of Palos Verdes Estates in October 2015.  Note continuous coverage levels shown over a wide area.  (As an aside, this map was found to have significant errors as it omitted coverage from numerous existing cell towers.  This has been a common and continuing problem with Crown Castle submissions.)

Cities require propagation maps for good reason.  Proposed cell towers are often highly intrusive as the industry wants cheap sites, usually resulting in “cookie cutter” designs in highly prominent locations.  Fortunately, states allow municipal regulation of cell towers, including those in the public right of way.  In our state this power derives from the California Constitution which reserves for municipalities vested police powers to regulate these sites (see this post regarding  T-Mobile et al vs San Francisco, 2016).

Often highly intrusive cell towers are proposed for speculative reasons or to give a wireless carrier a competitive advantage.  Based on all we’ve documented that certainly seems to be the case for much of Crown Castle’s Palos Verdes deployment.  Cities are under no obligation to approve such sites.  Under federal law, municipalities may deny a cell tower if it is not required to close a “significant gap” in service OR if the cell tower proposed (i.e. design/location) is not the “least intrusive means” of doing so.  Cities protect themselves by requiring propagation maps proving the site is truly needed.

The cell towers installers don’t want municipalities challenging their proposed cell towers.  Any challenge can result in delays, increased costs, and could ultimately result in denial of the proposed cell site.  They like the good old days, such as back when Rancho Palos Verdes used to rubber-stamp these cell towers.  Our post on the proliferation of ugly towers in RPV (here) shows what happens without a strongly enforced ordinance.  The propagation map requirement was one of many long-needed reforms implemented to get this problem under control.

Coverage map games

Enter the coverage map games.  If the cell tower installers can create the impression that a significant gap exists (i.e. existing coverage is terrible) then cities will likely think they have no choice but to approve the site.  RF Engineering expertise isn’t commonplace, and few cities have the experience to challenge such claims so they usually go unquestioned.  Fortunately, we have both the experience and the required test equipment to scrutinize what we’re being told.  We documented multiple Crown Castle coverage map discrepancies in an earlier post here, and in every case the errors we found understated the actual existing coverage.  As we’ve seen over and over, Crown Castle’s “errors” almost always seem to be in their favor.  It certainly doesn’t seem random.

It looks like Crown Castle’s string of “lucky” errors continues, as their recent submittals have brought us a new gem. This is for site ASG53, proposed for the corner or Granvia Altimira and Monero right next to the Palos Verdes Estates city line.  We’ve seen this site before and it’s always been questionable as it’s only 800 feet away from an existing AT&T “macro” high power cell tower.  Here’s the existing coverage map they submitted.

crown_castle_asg53_alleged_significant_gap

Crown Castle coverage map submitted to Rancho Palos Verdes depicting an alleged “significant gap” in coverage.  Map is a public record.  Note it looks nothing like the Palos Verdes Estates propagation map shown above.

Right off the bat, it’s got lots of problems:

  • It’s not a propagation map as required by RPV’s ordinance but a series of Crown Castle field measurements known as a “drive map”. This is an unacceptable substitute for authentic AT&T propagation maps like the Palos Verdes Estates submission shown at the top of the post.
  • It doesn’t show the frequency at which it was measured. AT&T operates on four frequency bands in our area, which one is it?  (BTW, we know what it is from our measurements, why won’t Crown Castle say?  In October we exposed a coverage map in Palos Verdes Estates that was deceptively marked with the wrong frequency.  The apparent lesson Crown Castle learned is to not list it at all.)
  • It only shows one frequency band. Coverage is the aggregate of all four of AT&T’s operational frequency bands, not some single cherry-picked worst case band
  • It only shows LTE/4G (Long Term Evolution) coverage. AT&T also operates UMTS/HPSA service in our area it markets as “4G” that supplements LTE coverage.  It’s completely omitted.

In short, it’s a misleading and cherry-picked depiction of coverage intended to convey worse coverage than actually exists.  Unfortunately, that’s just the start of it.  It looks as if the data submitted may have been altered from the actual measurements.

 The 30 dB Discrepancy

In addition to submitting the ASG53 site (Granvia Altimira and Monero), Crown Castle also recently submitted a neighboring site at Highridge and Peacock Ridge (ASG15).  Due to their proximity, the submitted drive maps have a significant overlap.  These areas should show identical coverage, but they don’t, the coverage is substantially different.

crown_castle_asg15_asg53_comparison

Top: Drive map allegedly depicting existing coverage submitted with ASG15.  Bottom: Drive map allegedly depicting existing coverage submitted with AG53.  Maps resized to common scaling, red box added to show common area.  ASG15 original available here, ASG53 original available here.

In order to allow a clear comparison between the two, we’ve shown just the common areas between the two maps here.  Note how the signal levels are much lower in the bottom plot.

crown_castle_asg15_asg53_common_area

Top: Common area from ASG15 drive map.  Bottom: Common area from ASG53 drive map.  Scale submitted with drive maps shown on right (poor quality resolution was how scale was submitted.) Note dramatically lower coverage levels in ASG53 (bottom plot) despite supposedly representing the same existing coverage.

So let’s compare the two maps allegedly depicting existing service in the same area.  Remember these are negative numbers so as an example, -85 dBm is a lower signal level than -65 dBm.

  • Areas that are dark green (> -65 dBm) in the top map are either light green (-65 to -75 dBm) or yellow (-75 to -85 dBm) or red (-85 to -95 dBm) in the bottom map
  • Areas that are light green (-65 to -75 dBm) in the top map are generally blue (-95 to -105 dBm) in the bottom map
  • Areas that are yellow (-75 to -85 dBm) in the top map are generally black (<-105 dBm) in the bottom map
  • Areas that are red (-85 to -95 dBm) in the top map are black (<-105 dBm) in the bottom map
  • Areas that are blue (-95 to -105 dBm) in the top map are black (<-105 dBm) in the bottom map

It seems pretty clear the two plots shows comparable distributions (higher in the same areas, lower in the same areas) but that the signal levels shown in the bottom plot are shifted approximately 30 dB lower than in the top plot.  This is a huge discrepancy that can’t be dismissed by some hand-waving excuse.  These values are expressed in decibels and are logarithmic, every 10 dB is ten times lower.  So 30 dB equals 1000 times lower (10 x 10 x 10).  It’s not like they forgot to carry the one or something trivial.

Which is Right and How Did This Happen?

So which map is right?  Based on our field measurements we think the bottom map (ASG53) understates LTE coverage on this band by 30 dB and the top map is accurate for this band.  That doesn’t mean the top map accurately represents service as it’s still a cherry-picked single frequency when there are actually four active bands.

As to how this happened, we don’t know but we certainly aren’t surprised based on what we’ve seen from Crown Castle to date.  We do note that if the ASG53 map showed what we believe are the correct levels (i.e. 30 dB higher) it would be very difficult to justify this new cell tower.  The reason is that the proposed Granvia Altimira and Monero location is only 800 feet from an existing AT&T macro cell tower (shown as a big blue dot on the maps.)  It’s just another one of the coincidental “errors” that happens to help Crown Castle, we suppose.

One last note, Crown Castle has stated during  Palos Verdes Estates Planning Commission hearings that they used a third party to collect these drive drive map measurements.  They even had a speaker from that company address the Commission, the clear implication being that the maps were independent and therefore trustworthy.  We spoke to these guys when they were recently conducting measurements in our neighborhood.  We specifically asked them if they made the maps.  We were told no, they give the raw data to Crown Castle and they generate the maps.  Make of that what you will.

Like this post?  Click here to go to our home page for more.