By the time you read this, the FCC is expected to propose a set of rules by which unlicensed devices will be allowed to share TV Channels 2–51.
The good news would be if these proposed rules took into account all of the findings of the FCC Laboratory’s March 30, 2007 report, “Interference Rejection Thresholds of Consumer Digital Television Receivers Available in 2005 and 2006.” If so, the commission would recognize the interference potential of signals on certain pairs of channels which result when such signals overload the tuner of a DTV receiver generating third-order intermodulation products (IM3) which fall in the Desired channel (D). For example, this shows that at D = –53 dBm, if there is an undesired signal (U) on channel N+3, the median receiver would be “jammed” at a D/U of –49.8 dB, which is a U signal 49.8 dB stronger than the D signal; that is; –3.2 dBm. Such a high level for a received DTV signal is quite unlikely to be found, even near a transmitter operating on channel N+3. For an undesired signal on channel N+6, the threshold D/U power reported was –50.8 dB, so the U power would be –2.2 dBm. No problem here.
But if there were undesired signals on both N+3 and N+6, received at equal power levels to the median receiver, jamming would result at any D/U > –34.1 dB, or a U power per signal greater than –18.9 dBm. There is a problem here. This pair of undesired signals would jam DTV reception at about a 16 dB lower received U power than either undesired U signal alone. But that is just the tip of the iceberg.
The above data is for the median receiver, meaning that 50 percent of receivers were better, and 50 percent worse than the median receiver’s performance. I consider this criteria to be unacceptable. A large portion of viewers live where the received power is below –53 dBm.
VALUES TO CONSIDER
The FCC report also provided the arithmetic average or mean D/U and the Standard Deviation (Sigma) for each measurement. You may recall that the mean value +/– 1 Sigma covers 68 percent of cases, and that to get down to a 2.5 percent probability of interference we must increase the mean U values by 2 Sigma. Sigma for the N+3, N+6 case is given as 6.1 dB and the mean value for U = –53 dBm + 34.4 dB = –18.6 dBm, which when increased by 2 Sigma = –6.4 dBm. This means that only about 2.5 percent of the potential viewers would be unable to receive the Desired signal at –53 dBm, if the U signals on N+3 and N+6 are greater than –6.4 dBm each.
The FCC report also provides data for D = –68 dBm and for D 3 dB above its noise-limited value. Both of these other data sets are important to broadcasters and the viewing public. For the sake of brevity, I will use only D = –53 dBm in this article. Interference may greatly reduce service area. Noise-limited coverage extends, according to the FCC, to where your field strength is 41 dB uV/m. Interference limited coverage area may not support the (DTV) station, once these unlicensed devices become widely distributed throughout your market.
(click thumbnail)Table 1: Maximum U power/channel for channel pairs to realize a 5 percent probability of interference (jamming).Table 1 shows the maximum U power per undesired signal for two undesired signals on the channel pairs indicated. The maximum value is computed from the mean Threshold U power and the Standard Deviation from FCC Table A-5 of the above referenced report. The probability of interference is 25 percent.
In my opinion, these values should be used by those designing/testing DTV receiving appliances. These values should also be the basis for any single U test conditions. There need to be values for N-7 and N-9 and also for N+7 and for N+9. Next month, I’ll publish a complete set of U max for 25 percent probability of interference and explain how I got those numbers.
SIGNAL HOPPING
Notice that I never said that the undesired signal is a DTV signal. No, power is power. Jamming will result from either two DTV signals on such channel pairs, of which there are many or one DTV signal and one unlicensed signal, or by unlicensed signals on any of these channel pairs of the form N+K and N+2K which generate IM3 in the desired channel (N). That is bad news, but it is not all the bad news.
Now we have the FCC permanent channel allotments so you can see if your ox may be gored, or your competitors’ ox is at risk, or if both oxen are at risk. Before management asks you about this, why not make these simple calculations for yourself by reading the “Seventh Report and Order and Eighth Further Notice of Proposed Rule Making, re: MB Docket No. 87-268 released Aug. 6, 2007.”
I will use Denver as my example here. On page 70 of this NPRM we find that the FCC proposes to allot channels 7, 9, 13, 18, 19, 32, 34, 35, 40, 43 and 51 to Denver broadcasters. What if an unlicensed device (UD) operates on either Channel 17 or 20? Receivers tuned to Channel 18 will be subject to jamming from a nearby U/D on Channel 20, while a receiver tuned to Channel 19 may be jammed by a nearby UD on Channel 17. This is because third-order IM “hops over” the other signal of the two undesired signals and falls as far from the other undesired signal channel as the distance between the two undesired signals.
Likewise, consider Channels 32 and 34. A UD signal on Channel 30 will hop over the DTV signal on Channel 32 and land in Channel 34. A UD signal on Channel 36 would hop over the signal on Channel 34 and may jam Channel 32. A UD on Channel 33 will hop over Channel 34 and may jam Channel 35. What about a UD signal on Channel 11? Well we have no data on the RF selectivity of the tuners in the receivers tested by the FCC, but I am deeply concerned that receivers made this year and next will have little or no RF selectivity, in which case a UD signal on Channel 11 will hop over the DTV signal on Channel 9 and may clobber reception of Channel 7.
Come to think of it, why can’t the DTV signal on Channel 7 hop over the one Channel 9 and clobber reception of UD signals on Channel 11? This is quite likely, as I suspect that tuners in receivers of UD signals will be readily overloaded by strong DTV signals as we have seen with tuners in modern DTV receivers. So, I conclude that Channel 11 may not provide reliable operation of unlicensed devices when introduced, and the same applies to Channels 30 and 36 for the same reason.
Of course robust tuners could be designed for UD receivers, but will they? Will robust tuners be developed for DTV? That does not appear to be the direction in which the industry is going. I will be addressing these questions at the International Conference on Consumer Electronics, which follows the Consumer Electronics Show in January 2008 in Las Vegas. My audience will be mostly receiver design engineers and their engineering managers of consumer electronics firms worldwide. I hope that some broadcasters will also be in my audience.
As an exercise, figure out what could happen to DTV reception of Channels 40, 43 and 51 in the Denver area where a single UD operates on any of the channels where it can hop over a DTV signal and attack another DTV signal as I’ve tried to show above. A triplet of signals and their third-order distortion products is shown in Fig. 1 below. Note that the spectrum occupied by these three signals and their distortion products is 20 channels wide.
(click thumbnail)Fig. 1: Triplet of DTV signals on Channels 31, 34 and 37 with third-order intermodulation products on Channels 24-44. IM3 generated by this triplet may cause interference over up to 20 channels centered on 34.
THREES A CROWD
Welcome to the world of the future in which channel pairs become channel Triplets and the horrible consequences this poses.
Speaking of Triplets, I counted 29 instances of Triplets of DTV station allotments in the above NPRM. What happens when a third signal, say from a nearby UD, is received at the same power as the pair of DTV signals? This situation is called a “Triplet.” These signals do not have to be at the same power. Next month, I’ll explain how to deal with unequal received powers (or field strengths).
If one signal is received at, say –15 dBm, the pair is received at –12 dBm if their powers are equal hence 3 dB higher than either signal. Now add a third signal also at –15 dBm, and we now have a total received power in the Triplet of –10.2 dBm. That increase may not sound alarming, but remember that we are dealing with the IM3 which falls into our Desired channel and IM3 increases by 3 dB per each dB increase in the signal power generating it. Going from –15 dBm to –10.2 dBm is an increase of 4.8 dB and the resulting increase in cochannel noise (IM3) is 14.4 dB. Triplets are going to be fierce sources of interference which were not anticipated by the FCC. The FCC must have thought but never said that it assumed receivers would not be overloaded. Dr. Oded Bendov is publishing a paper in the December 2007 issue of IEEE Transactions on Broadcasting in which he will present a very comprehensive analysis of the fierceness of such Triplets. I hope that my readers will read this important document.
Let us hope that the FCC will recognize not only the proven fact that undesired signals on those channel pairs which generate IM3 falling in a desired channel can be the dominant source of interference to DTV reception, as shown above, but also that such channel pairs may readily become triplets when a “vacant channel” is used by a nearby UD. Look again at the Denver channel allotments. Most of these would be subject to this kind of interference.
To protect DTV reception for interference from unlicensed devices, the FCC should not allow triplets to appear where its DTV allotments have created channel pairs as shown above. For example, in Denver we have DTV channel allotments of Channels 18 and 19, therefore Channels 16, 17, and 20, 21 should not be considered as being vacant (suitable for use by unlicensed devices). This would encourage receiver designers to provide adequate RF selectivity to suppress interference from unlicensed devices operating on channels such as 14, 15, 22 and 23.
Stay tuned.
