The January 1954 issue of "Proceedings of the Institute of Electrical and Electronic Engineers" was a special issue devoted entirely to the NTSC compatible color television system adopted by the FCC. At that time, it was considered a marvel of technology to pack red, green and blue video signals into a 6 MHz channel in such a way that the millions of TV receivers already in homes could continue to operate without any modification, rendering good picture quality while permitting broadcasting in glorious color to new receivers in the hands of "working millionaires" as Bob Hope once quipped.
I immediately purchased a copy for $10 and began to study this marvelous new technology. At that time, the price of this special issue was a strain on my budget so that buying one of these early NTSC receivers was out of the question. I wasn't a working millionaire, but I was working, so I decided that my career would benefit from designing and constructing my own NTSC receiver, which I actually did using the information in that issue.
I modified my Emerson rear-projection TV console to accommodate two Norelco video projectors. There wasn't space for a third, so I settled on displaying the luminance signal on both projection CRTs while feeding the chrominance signal to a push-pull chroma demodulator and recovering the I component. The +I chroma went to the grid of one CRT, the -I signal to the other. In the common light path I mounted a dichroic mirror which reflected the orange colored Y+I modulated light, while it transmitted the Y-I blue modulated light.
Both images arrived at the screen and were laboriously registered. I couldn't reproduce pure red, green or blue, but on the other hand, flesh tones looked natural, and there were no green or purple people on my screen. This project may have lead to my being hired by Tektronix and what followed was a wonderful career building instrumentation for NTSC, PAL, PAL-M (Brazil) and SECAM applications.
SEQUEL PUBLISHED
Now, 52 years later, the IEEE has published its sequel in the same publication. This is the January, 2006 Special Issue on "Global Digital Television: Technology & Emerging Services." This issue covers not only the ATSC DTV system but the European and Japanese DTV systems, DVB-T and ISDB-T, as well. I urge my readers to order a copy of " Proceedings of the IEEE," volume 94, number 1, for January 2006. If you don't have the January 1954 special issue, you can probably obtain one over the Internet. I still have mine. It jump-started my career.
I immediately turned to the paper on DTV receiver implementation beginning on page 119 by several well-known authors. I was pleased to see DTV receiver performance discussed in detail. Such information is generally a trade secret.
For example, I learned that a double-balanced mixer whose third order intercept power is +15 dBm has been used in some consumer DTV receivers. I first became aware of third order intercept power (IP3) and third order intermodulation product power (IM3) in the literature of wireless telephone and microwave engineers. Both Dr. Oded Bendov and I have tried to familiarize broadcast engineers with these parameters. I found the authors of this IEEE paper using them too.
(click thumbnail)Fig. 1 depicts the relationship of IP3 and IM3 to the signal power that overloads the active device being discussed. By definition, at IP3, the third order intermodulation product power equals the signal power overloading the device. You might wonder: what is the use in knowing the value of IP3 when it results in extreme distortion?
Knowing this parameter value for a given active device such as the mixer of a DTV receiver, you can determine the maximum signal power input to the device that can be usefully applied. It indirectly defines the upper end of the device's dynamic range. Fig. 1 shows that IM3 decreases by 3 dB for each 1 dB reduction in input signal power; so avoiding signal overload is the strategy to follow.
TOO LITTLE OR TOO MUCH
DTV reception fails at both ends of the dynamic range of the tuner. If the received signal is not more than 15.2 dB above the total noise, reception fails--the so-called "cliff effect." When the received signal overloads the tuner, IM3 is generated within the channel, thereby increasing the total noise power as these IM products look like, and act like noise. The result is that the SNR may fall below the threshold value, preventing reception. This can easily be proven by placing a 75-ohm attenuator at the antenna connector on the receiver.
In fact, I have used a switched 75-ohm attenuator to measure the signal level reserve. It is really nice to know that with up to perhaps 8 dB of attenuation, you still have the SNR above threshold on all local channels. It is even more helpful to find that you need 5 dB of attenuation for one or more channels to be received. It may not be generally known that the desired signal alone can, if it overloads the receiver, cause reception to fail, leaving no clue as to why this station cannot be received.
Interference from other TV signals can overload the receiver with the same result. Attenuating the RF power to the receiver will quickly prove this. This is likely when there is a nearby station on either adjacent channel, especially in the UHF band where receiver selectivity is relatively poor or even non-existent, meaning there may be no RF selectivity in the tuner ahead of the mixer. It is quite difficult to put a tracking filter on an IC chip and putting the entire tuner circuitry on a chip is in vogue today.
Let's see where knowing the mixer IM3 takes us. First, we know from Fig. 1 that the mixer will be useless when signals approaching its IP3 reach its input port. As Fig. 1 also shows, the onset of signal compression begins below IP3.
Manufacturers usually specify this as the input power at which the signal suffers 1 dB of compression, as this is easy to measure. Some signals with 1 or more dB of compression can be correctly demodulated, but not 8-VSB. Remember that we always measure the average power of a DTV signal.
There are eight symbol levels in 8-VSB, four are below the average power, and four are above it. It will be the highest power symbols (+7) that get "crushed." Those may be decoded as +5 if compressed. The 1 dB compression point of a device is 10 to 15 dB below its IP3 power as shown in Fig. 1. So now we know that the strongest signal, desired or undesired, at the mixer input must be, say, 15 dB below the mixer IP3.
The authors of the IEEE paper confirm that RF selectivity in the UHF band does little to avoid interference by first adjacent channel signals (ACI). We can infer that the maximum signal power at the input to this and similar DTV receivers must be below 0 dBm.
This is the function of the RF automatic gain control loop which was ably described by these authors. Ideally, the RF AGC loop starts with the mixer output, before the IF filter where adjacent channel signals can be present. This wideband IF signal is rectified and used to control either the gain of the RF amplifier or to control a PIN diode attenuator before the RF amplifiers.
I suspect that a number of receivers out there don't have this kind of RF AGC. They may simply sample the signal power after the IF filter where all traces of adjacent channel signals have been removed.
This permits strong adjacent channel signals to be amplified up to 15 dB in the RF amplifier when the receiver is tuned to a weak signal. In this case, even 3 dB of RF attenuation at the input of the receiver may allow it to receive the desired (weak) DTV signal. Oh, the power of attenuators.
Soon, all DTV stations may be operating at their maximum authorized power. I predict we will be hearing an outcry when folks discover they can't receive all local stations in digital.
Further, channel changes will only make things worse when we have lost channels 52-69, and that is coming too! You need to understand how interference to your signal can be overcome.
History does repeat itself. The early years of NTSC were very frustrating. Critics of NTSC called it "Never Twice the Same Color." One network proposed to use a version of SECAM in studios to solve the color problems with videotape recording. The DTV reception problems (interference) can and will be solved by improved tuner designs once tuner manufacturers sense the business need to improve their product. My next column will review the current state-of-the-art inactive devices for tuners and demonstrate what can be done with available technology. Stay tuned.
And please advance your own career by obtaining the invaluable "Proceedings of the IEEE," volume 94, number 1, January 2006.
