The End of Analog Allotments - TvTechnology

The End of Analog Allotments

This is the first allotment table having no analog TV channel allotments; it is our future broadcasting universe.
Author:
Publish date:

In the FCC’s Seventh Report and Order and Eighth Further Notice of Proposed Rule Making, (MB Docket No. 87-268), it has given its latest DTV Table of Allotments in Appendix B. From it, I have looked for pairs of adjacent DTV channels. This is the first allotment table having no analog TV channel allotments; it is our future broadcasting universe.

It has by my count, 85 adjacent DTV channel pairs. At one time we had almost 300 first adjacent channel pairs where there was an NTSC channel allotment immediately above a DTV allotment in the same city. So a lot of progress has been made in reducing the number of adjacent channel pairs allotted to the same community.

In a few cases, there are triplets; N-1, N and N+1 in the latest allotment plan. For example, Los Angeles has Channels 41, 42 and 43. Miami has Channels 18, 19, and 20. Atlanta has Channels 19, 20 and 21. Honolulu has a quadruplet set of channels; 8, 9, 10 and 11. Topeka, Kan., has 11, 12 and 13; Boston has 30, 31 and 32; while Detroit has 43, 44 and 45; my hometown of Buffalo, N.Y., has 32, 33 and 34; and finally, Milwaukee has 33, 34 and 35.

(click thumbnail)Fig. 1: The spectrum plot Triplet 1, 1.
The spectrum of these triplets looks like the triplet of Channels 35, 36 and 37, shown in Fig. 1, except that they are shifted in frequency. In each case, the center channel will have triple-beat cross modulation (X-M) denoted by two red Xs in Fig. 1. This triple-beat X-M is generated in the front-end of receivers when they are overloaded by the total signal power present. Third-order IM products are also generated in overloaded front-ends and these are shown in blue as capital letter “I” in Fig. 1.

In addition to this triple-beat X-M there will also be sideband splatter into the middle channel (36 in Fig. 1), which is sideband splatter radiated by the other transmitters of the triplet, (Channel 35 and Channel 37 DTV transmitters in Fig. 1). If this triplet is received at say –25 dBm per channel, the sideband splatter into Channel 36 from Channel 35 and from Channel 37 will be –25 dBm – 46.5 dB = –71.5 dBm from Channel 35 and also from Channel 37. The total sideband splatter into Channel 36 is therefore –68.5 dBm. This is the noise floor in the desired channel (36) due to the other transmitter’s radiated sideband splatter. Any third-order distortion products generated in a DTV receiver would only increase the noise floor in Channel 36.

The total signal power received on just these three channels is –25 dBm (per channel) + 4.88 dB = –20.23 dBm at which there will be both X-M and IM3 third-order distortion products generated in receiver front-ends. Reception of the center channel signal at –25 dBm requires that the total noise in the center channel is more than 15.2 dB below –25 dBm. That is the total of sideband splatter from Channels 35 and 37, plus the receiver’s X-M and IM3 in the desired channel must be less than –25 dBm – 15.2 dB = –40.2 dBm. When both the desired and undesired signals are received at the same power as shown in Fig. 1, there is a lot of headroom for receiver generated third-order distortion, and it is almost certain that none of these signals is a threat to another signal in this triplet. As for third-order distortion products generated by this triplet in receivers, Fig. 1 shows that there are a lot of distortion products both IM3 and X-M in Channels 34 and 38.

RECEPTION FAILURE

However suppose the D/U ratio is –25 dB, which is close to but still within the FCC limits (–26 dB and –28 dB) so the desired signal on Channel 36 of Fig. 1 is –50 dBm while the undesired signals on Channels 35 and 37 are at –25 dBm. The sideband splatter into Channel 36 is still –68.5 dBm, while the desired signal power is only –50 dBm, so the maximum noise level in Channel 36 is –65.2 dBm, which is only 3 dB above the sideband splatter from the other transmitters. If the distortion products in the desired channel generated in the receiver total about –68 dBm, then the total noise increases by 3 dB so the SNR is only 15.2 dB, the threshold SNR of our ATSC signal. Put simply, reception fails!

The center channel of a triplet of three contiguous channels is more vulnerable that the others. You may have already concluded that an ideal receiver, one which is perfectly linear with two input signals at –25 dBm each, has a mere 3 dB signal level margin and you might say this is not a comfortable margin, and I’d say you are right.

Why I didn’t bring this forth earlier is that I wanted to be sure that there are going to be such triplets of contiguous channels in the Digital World, post Feb. 17, 2009, and in fact there are and you know where they are going to be found.

There is also one quadruplet. You can break this quad down into two different triplets: 8, 9 and 10; and 9, 10 and 11. The other channel only increases the U power thus increasing the receiver generated distortion products. There would be three undesired signals of –25 dBm each or a total of –25 + 4.77 dBm = –19.23 dBm. However this 4.77 dBm increase in U power will result in an increase in receiver generated third-order distortion products of 3 x 4.77 dB = 14.3 dB more X-M + IM. Good luck, Honolulu.

TUNER DESIGN CHALLENGE

I’ve heard from an unusually reliable source that Broadband over Power Lines is now dead. The FCC found only 4,776 subscribers in the country at the end of 2006, the latest data available. BPL is a scheme for broadband transmission using power lines which is similar to the scheme to provide a campus-wide radio service, (Carrier Current) on the low end of the AM broadcast band which is more than a half century old technology.

Now, if sharing of broadcast spectrum were to also go away, with as little harm done as BPL did, broadcasters should be extremely pleased, but I don’t anticipate such a blessing.

Designers of TV tuners face a really difficult set of design trade-offs. The RF selectivity of what is often called the tracking filter is an excellent example. Tracking filters are tuned (electronically) to the desired channel. All frequencies within the desired channel should pass through the tracking filter with the same attenuation, and that attenuation, usually called insertion loss, must be kept very small as the insertion loss adversely affects the noise figure of the tuner. The designer’s dilemma is that all frequencies outside of the desired channel should suffer more attenuation. Information about the RF selectivity of tuners is generally a trade secret but the physics is in many text books. In the good old days, circa 1981, Blair Benson published his “Television Engineering Handbook.” In the chapter on TV tuners, he showed a single tuned circuit which couples energy from the antenna port to the RF amplifier, and a double-tuned circuit between the RF amplifier and the mixer. Back then, the tracking filter consisted of a single tuned circuit and a double tuned circuit isolated from each other by the RF amplifier. The double tuned circuit was overcoupled, so it showed two peaks within the desired channel and a small dip between them. The single-tuned circuit ideally flattened the frequency response within the desired channel while the double-tuned circuit rejected undesired signals on other channels.

THE Q FACTOR

As tuners were miniaturized, the size of their components was reduced. Today we have extremely small resistors and capacitors and of course the transistors are now on an IC substrate. These components probably perform better than earlier and larger parts, but I believe that the same cannot be said for the inductors of the tracking filter. Sure enough, there are inductors available so small that you can only move them with tweezers in hand and only with a magnifier to see what you are doing. They cover the range of inductances needed for TV tuners; their downfall is their low Q factor. A low Q factor for the inductor lowers the frequency filtering properties of the tracking filter (RF selectivity).


(click thumbnail)Fig. 2: A triplet of contiguous channels carrying DTV signals of equal power.Linley Gumm has provided me with a plot of the selectivity of a single tuned circuit; one L and one C (a Varactor to tune it). This is shown in Fig. 2. I asked him to plot several curves in which he varied the flatness across the desired channel (44) from 0.5 dB to 3 dB. One of these curves should correspond to the passband flatness of a single tuned tracking filter tuned to Channel 44, so it may be typical of the RF selectivity where the tuner has for a tracking filter one tuned circuit. I do not mean that all tuners being produced have but one tuned circuit, but this is what one tuned circuit can provide. Tuners with more than one tuned circuit in their tracking filter should have much greater RF selectivity. From Fig. 2, one sees that undesired signals on channels N+/–2 are attenuated by about 5 dB; and signals on channels N+/–4 by about 10 dB and so on. However, this is a theoretical calculation. It assumes ideal reactances—an ideal inductor and an ideal capacitor, neither of which is commercially available. The capacitor is in fact a silicon diode, which has significant ohmic resistance, lowering the circuit Q. We’ve already noted that tiny inductors have tiny Q factors, too. So these curves represent an idealized tracking filter. Actual RF selectivity will be lower.

What this signifies is that we cannot expect much attenuation of undesired signals in tuners of DTV receivers. This is why the Frequency Agile Active Filter I described in the June 25, 2008, issue of TV Technology would not use the state-of-the-art in miniaturized inductors. Freed of space considerations, perhaps what are known as helical resonators would be used. Note I said resonators (plural); perhaps we can get back to the classical structure of three tuned circuits to comprise our tracking filter. It is also possible to improve the tuning capacitors (varactor diodes) by fabricating them of gallium arsenide rather that silicon. These would cost more, but perhaps the performance gain is worth it.

Helical filters are usually fixed tuned. Whether they can be tuned over the frequency ranges needed for a Frequency Agile Active Filter, I cannot say at this time.

Stay tuned and with a high Q.