What I Know About the Great White Spaces

Broadcasters are well aware that the spectrum of DTV transmitters extends outside of the allocated channel. They know this because the FCC has required that this out-of-channel emission be controlled with its RF mask specifications updated in 1988.
Image placeholder title

By now you know the FCC has released its Second Report and Order and Memorandum Report and Order concerning broadcast spectrum sharing. Now we know that the maximum average transmitter power output of the proposed devices for use in this service (as measured at the antenna) is 1 watt. If the transmit antenna has a directional gain greater than 6 dBi the transmitter power output must be backed off to the extent that the antenna gain exceeds 6 dBi. So the maximum average radiated power is 4 watts.

Some readers may well ask what a dBi is. A radiator of electro-magnetic energy is said to be isotropic if the power density at a given distance from the radiator is the same in every direction. A light bulb approximates being an isotropic radiator of light. The key word here is isotropic. For example, a dipole has some directivity; it concentrates its field along directions normal to its axis by not radiating any power along its axis. The gain of a dipole over an isotropic radiator is 1.64 times, or 2.15 dB so it is expressed as 2.15 dBi. If you purchase an antenna, its gain might be specified either in dBi (its gain over an isotropic radiator) or dBd (its gain over that of a dipole).

What were called white space devices are now referred to by the FCC as Television Broadband Devices (TVBD). TVBD of the fixed facility type will be limited to 4 watts effective isotropic radiated power. With a 1 watt transmitter and an antenna whose gain is 6 dBi, the maximum field strength at 1 mile can be shown to be 76.7 dB above 1 microvolt/meter (dB uV/m) at 1 mile. The power received by a dipole antenna resonant at 615 MHz 1 mile from a fixed TVBD is –54.25 dBm (76.7 dB uV/m – 130.8 dBm-dBu).

Field strength decreases by 6 dB for each doubling of the distance from the radiator over a direct line-of-sight path. For example, with an EIRP of 4 watts the field strength is 76.7 dB above 1 dB uV/m at a distance of 1.0 miles from a "fixed facility device" in nearly any "vacant channel" of the broadcast spectrum after Feb. 17, 2009.

Channel 37 is disallowed as it is reserved by an international treaty for Radio Astronomy only. Channels 3 and 4 are also prohibited to protect TV reception where the signal received is translated to channel 3 or 4 and fed to the receiver on that channel. This has been the practice in CATV systems for some decades. Channels 12 to 20 will be disallowed in and near the 13 major markets in which these channels may be shared with public safety radios or mobile telephones.

Field strength is the density of the radiated power, so it is independent of frequency. Of greater interest is the power captured from this field by a receiving antenna and this is where the dipole factor comes into play. This factor relates the power captured by a resonant dipole antenna to the field. For example, at 615 MHz (the center of the UHF TV band) this is –130.8 dBm/dBuV/m. A rooftop directional antenna with a net gain of 6 dBd would capture –48.25 dBm of power at 1.0 miles from the fixed facility device operating at 615 MHz.

The power captured varies inversely with distance over a line-of-sight direct path by 20 log 1/d where d is given in miles. It also varies with frequency by 20 log 615/ log F (MHz). So at 69 MHz, the center of the low VHF band, the same radiated power would be received at 1 mile by 19 dB more than at 615 MHz. This is because a dipole resonant at 69 MHz is physically large compared to one resonant at 615 MHz. I like to think of this like a sail—the greater its area the more power it intercepts.


So now we know all about effective isotropically radiated power (EIRP). But what does this portend for broadcasters? First, there is no such thing as an isotropic radiator, but it is a useful concept. By specifying EIRP, the FCC is, in effect, saying that the field strength of such devices will not exceed 76.7 dB uV/m anywhere 1 mile from the radiating device. The FCC is not saying that these devices will be fitted with isotropic radiators (just try ordering one). If one of these devices were fitted with a dipole antenna, its field strength would vary with the gain of the dipole. Along the axis of a dipole, there is zero field strength, but along a perpendicular to the dipole, the field strength will be 2.15 dB higher than if the radiator were an isotropic radiator. If someone wants to use a directional antenna with a gain of 10 dBi, the power fed to that antenna must be backed off by 4 dB.

As for personal/portable TVBD, they are limited to 100 mW EIRP unless they operate on a channel adjacent to a DTV channel in which case they must operate at 40 mW.

My understanding of how these TVBD will work is as follows: Each device has a GPS receiver which determines its location. When the TVBD is switched on, it transmits a message to a "third party" who (for a fee), will determine which channels are available to the transmitter at this site and automatically transmit back the authorized channel number and authorized power. The user is completely unaware of this transparent process, at least until he is billed.

The FCC certainly realizes that some people might want to increase their EIRP by feeding their transmitter output to an off-the-shelf wideband power amplifier—many of which are readily available. The FCC requires that the transmitter be connected to the amplifier with a permanent connection. Perhaps they regard as permanent a special connector with nonstandard threads. I found that such connectors are available commercially which meet FCC specifications given in Part 15.203 of the FCC rules. These are 50 ohm connectors in BNC, TNC and SMA styles.


Broadcasters are well aware that the spectrum of DTV transmitters extends outside of the allocated channel. They know this because the FCC has required that this out-of-channel emission be controlled with its RF mask specifications updated in 1988.

They know all about RF mask filters because they had to purchase such a filter or perhaps two, one for their transition channel and in some cases another for their permanent channel. RF mask filters are both costly and physically large. They are designed for a narrow range of channels and tuned for one specific channel. DTV emissions outside the allocated channel are limited to be at least 44.2 dB below the power radiated with channel.

So could the FCC impose a similar out-of-channel emission limit on these unlicensed devices? Both are digital transmitters, but one is frequency agile; covering channels 2 to 51 or 54 MHz to 698 MHz with a maximum power output (at the antenna) is 1 watt or 100 mw. The other is for one specified 6 MHz channel, but the power may be 60,000 watts average power, 240,000 watts peak.

Perhaps these low power transmitters don't need an output filter because of their low power; they can run the amplifier in class A1, which is the most linear mode, trading power efficiency for linearity. Certainly for vehicular use, this is a good tradeoff, but I'm not so sure for personal portable units. The 100 mW of power comes from a re-chargable battery carried around by the owner. Class A1 power amplifiers operate at 25 percent power efficiency, which is significantly lower than class AB1 or AB2.

Perhaps there won't be a power amplifier, perhaps the digital-to-analog converter will operate at the radiated signal frequency and provide the power needed. That would be a truly all-digital transmitter. I think one recently tested by the FCC was of this sort, but its power output was only –9.5 mW but it was certainly "clean power."

But how can this –9.5 mW transmitter be used by a TVBD? Simple, just add a power amplifier to bring its transmitter power output at the antenna up to the FCC limit. Not so simple, this power amplifier may generate second and third-order distortion products which may have to be filtered. This power amplifier will have to be extremely linear to protect to the FCC adjacent channel protection ratio (ACPR) of 55 dB or it will have to be followed by a tracking filter, which is linear even at the power peaks involved. If this tracking filter is part of the transmitter, we can assume a 4 dB line loss to the antenna so the amplifier must provide more than 4 watts average power output to the transmission line after the tracking filter. At these low power levels, the tracking filter could be permitted to have a significant insertion loss but I don't see how it can be tuned by varactor diodes.

Any TVBD that operates on either channel 36 or 38 will have to meet additional requirements for spectral purity.

Not to be an alarmist, but why are these emission limits important to broadcasters?

I will skip over the problem of interference from a TVBD transmitter into DTV receivers tuned to a DTV signal on a first adjacent channel because the FCC will not permit either co-channel or first adjacent channel operation of a TVBD within or even near the coverage area of a DTV station. This is not because of interference to DTV reception, but is due to the fact that reliable signal sensing has not yet been demonstrated when there is a DTV signal on a channel adjacent to the vacant channel that the TVBD might wish to use. The implication is that the FCC would consider allowing vacant adjacent channels to be used if and when signal sensing is proven to be reliable and accurate.


Let's look at second adjacent channels to DTV channels. Consider a DTV allotment on channel N. The second adjacent channels are N-2 and N+2. If either or both are not allocated they are vacant and usable by a TVBD.

Now let's look at the known DTV-DTV interference thresholds for second adjacent channels. The place to look is FCC/OET report 07-TR-1003 March 30, 2007. Table A-9 has the measured U thresholds for D = –68 dBm.

The FCC reported the threshold of interference for five performance grades: Best U, Average U, Median U, Second worst U and Worst U.

The median receiver U thresholds were –25.7 dBm and –27.2 dBm. The second worst receiver U thresholds were –38.2 dBm and –40 dBm.

Remember this was for one ATSC signal interfering with another ATSC signal. First, the mechanism for this interference is cross-modulation of the (weaker) desired signal by the (stronger) undesired signal. The cross-modulation took place in the front-end of the receiver.

Now cross-modulation is a third-order distortion product. This means that the average power of these signals is not the significant power the peak power is because it is the power peaks that overload the receiver front-end. OK, the transient peak to average power ratio of our ATSC signal is generally accepted as being 6 dB. If the TVBD employs a multicarrier modulation scheme such as COFDM, its peaks will be significantly higher than 6 dB above the average power of the undesired signal. We will come back to this point.

Table I gives the distances from a fixed TVBD to the rooftop, directional antenna needed to receive DTV signals at –68 dBm for interference on channels N-1 and N+2.

These channel pairs are of the N+K, N+2K form which generate third-order IM centered in channels N and N+3K. K is an integer. The FCC tested with K from –5 to +5.

Distances are for an undesired DTV signal. They will be greater for a COFDM modulated U signal. Distances are approximate only. The order of magnitude is what is significant.

The above distances around each fixed facility TVBD will be a black hole into which DTV signals vanish. As the receiving site moves from the –68 dBm contour towards the –84 dBm contour, these black holes will become even larger as receiver noise will add to the noise from these TVBD.

We've got a problem to solve. Stay tuned.

Charlie Rhodes is a consultant in the field of television broadcast technologies and planning. He can be reached via e-mail at cwr@bootit.com.