SpectrumTalk

Ofcom on
TV White Space

As many readers know, Ofcom is the FCC's UK counterpart, with roughly comparable jurisdiction. It is much less politicized than FCC is and is run by what we would call a "single administrator" like EPA. As opposed to FCC where virtually all decisions have be be voted on by 5 political appointees,

"(Ofcom) is based upon a model which is familiar to the commercial sector but which marks a departure from the past.
Ofcom has a Board with a Chairman and both executive and non-executive members. The Executive runs the organisation and answers to the Board, whilst the work of both Board and Executive is informed by the contribution of a number of advisory bodies."

(As I have pointed out in my reply comments to Docket 09-157, FCC - without additional legislation - could move to this model for noncontroversial spectrum matters and probably increase its productivity significantly - but I digress.)

On November 17, Ofcom issued a "discussion document" entitled "Digital Dividend: Geolocation for Cognitive Access A discussion on using geolocation to enable licence- exempt access to the interleaved spectrum". After you translate the UK-jargon (Why can't they learn to speak English like us?) it turns out that this is basically the same issue in Docket 04-186, "TV whitespaces". Like the original FCC NPRM, Ofcom is thinking about white space devices using either geosensing or LBT (listen-before-talk) technology. FCC in its November 4, 2008 2nd R&O selected the "belt and suspenders"
approach of requiring both geosensing and LBT - further compounding the situation by basing the geosensing on a 1966 propagation model that FCC itself had already discredited for all but its original context of TV allotments.

In Ofcom's geosensing proposal they state

"Over many years broadcasters have carefully predicted the signal levels that will be received from their transmitter networks and have refined and validated these predictions. This information is held by Arqiva, which conducts the modelling on behalf of the broadcasters. There should be little difficulty in providing such signal strength information to the database. This implies that the database will not need to perform propagation modelling on behalf of DTT (digital terrestrial television - not necessarily HDTV)."
(Arqiva is the private entity in the UK that provides the TV transmitters for ITV, Channel 4, S4C and Five. UK private broadcasters generate content, they do not have the equivalent of Part 73 licenses as US broadcasters do. DTV/DTTV signals for private broadcasters are multiplexed together and transmitted from common transmitters and towers on common frequencies.)

Thus the assumed coverage of TV stations under the Ofcom proposal will be based on "validated predictions" of coverage accumulated over many years. So what does FCC require? Check out the new §15.512(a):

"TVBDs must protect digital and analog TV services within the contours shown in the following table. The contours are based on the R–6602 curves contained in §73.699" of this chapter.

These R-6602 curves were developed in the pre-desktop computer age of 1966 as a convenient way of calculating nominal coverage of TV stations, using 8 radials to consider terrain. The primitive algorithm can be seen in Figures 4 and 5 of §73.333 . (On these 8 radials only the difference between the 90 percentile altitude and the 10 percentile altitude is considered and even this is done on a rough grid appropriate for the manual calculation of the 1960s.) Note that due to the wording of §15.512(a) it is not clear if the limited terrain corrections of Figues 4 and 5 are even to be used! That is because the details for using even these primitive correction factors are in §73.684 which is never cited or incorporated by reference. FCC has warned about the limitations of this technique even when the correction factors are applied:

"Under actual conditions, the true coverage may vary from these estimates because the terrain over any specific path is expected to be different from the average terrain on which the field strength charts were based. Further, the actual extent of service will usually be less than indicated by these estimates due to interference from other stations. Because of these factors, the predicted field strength contours give no assurance of service to any specific percentage of receiver locations within the distances indicated.”

The limitations of the FCC technique are clear in the following coverage map of KTNC-TV/DT, Concord CA from the FCC's dtv.gov website.  Note the nearly circular shape of the contours for both the original analog service and the new DTV service map of this station.  Also shown is a terrain map of the area with a circle marking about the same spot as the official service contours.  (Note the presence of a few orange dots showing people within the contour who lose coverage after the switch to DTV.  Those who never had analog coverage are not shown.)


Hint -- why is it called "Napa Valley"?  Do you think the wine growers of Napa Valley can get this signal from Concord CA with antennas?  Anyone who has been to San Francisco will recall it is a hilly city.  Yet according to the FCC coverage prediction, people on the Pacific Coast north and south of SF can get KTNC.  These are the realities of the R-6602 model.  It was appropriate for determining allotments prior to the computer era and it may be appropriate for "administrative certainty" to keep it for allotments today.  But it is unrealististic to use it to define actual coverage of a TV signal.

So why is the UK proposing a realistic propagation-based geosensing while  FCC is using "belt and suspenders" with unrealistic propagation which overprotects the few people who actually use antennas for TV reception?