Dec. 12, I was walking to my office when Paul Ollinger, Iowa Public TV's assistant director of engineering for RF systems stopped me to talk about a problem with our DTV installation in the Cedar Rapids/Waterloo market. Those of you who read my articles regularly know that this has been one of the smoother, more cost-effective projects that we have undertaken.
Paul had just gotten off the phone with the antenna manufacturer. It seems that when they swept the antenna after the installation, the contract engineers for the manufacturer measured a return loss of 21.2 dB on our DTV channel. For those of you more comfortable with the VSWR figure, that represents a 1:19.1 ratio. Or, put even more simply, without any ice on the antenna, about 1 percent of the power at the antenna input would be reflected back down the line. The antenna engineer then told Paul that he felt he could get the VSWR down to about 1:15.1. He then asked: What would we be satisfied with?
What kind of question is this to ask a customer? Well, I can tell him that I am not going to be satisfied with a 23.1 dB return loss and .5 percent reflected power. Reading from the manufacturer's data sheet on this particular antenna, they specify "VSWR under 1:05.1 per channel and under 1:1.1 across 20-channel bandwidth." The 7-inch transmission line has a maximum VSWR specification of 1:1.08 for 1,500 feet. I couldn't find a specification for a 2,000-foot run; however; the contract engineer did sweep this line into a termination and measured a worst-case VSWR of 1:1.0271.
This reading indicates to me that the line is in pretty good shape. So how does an antenna with the performance specification listed in the sales brochure and a transmission line that is performing exceptionally well end up with such lousy performance in the field? More importantly, what does it all mean and how important is it really?
REFLECTING ON POWER
I'll take a stab at the second question first. Remember that the other name for return loss is reflected power. Also think back over the last few years of development in DTV receivers, and what the one critical performance area is that everyone has focused on: multipath performance. Remember now that multipath is another name for ghosting; and what is ghosting but reflections of the signal? So what we are talking about here is a ghost approximately 21 or 23 dB below the primary RF signal traveling down the transmission line while the primary RF is traveling up the line. The majority of this reflection will be absorbed by the hybrid at the input to the horizontal run but some of it will mix with the primary signal. The end result is that we'll end up with some increased inter-symbol interference that will degrade the capability of the receivers in the field to decode our DTV signal.
We contacted the transmitter manufacturer and asked what kind of return loss they typically want to see at the output of the transmitter. The manufacturer would like to see a 32 dB return loss that translates to a 1:1.05 VSWR. They stated that they could tolerate as much as a 1:1.10 (26.4 dB), but anything beyond that will make them concerned about the performance of the RF system. I have to admit that I too am very concerned about the poor performance of this system right out of the box.
If I accept a 1:1.15 VSWR that was offered, what happens to the performance of this system in the winter as we start to see icing on the antenna? What about as this system ages and the transmission line goes through a decade or so of seasonal changes with contraction and expansion? We could easily lose another two or three dB in the return loss column, which could easily result in loss of headroom at the receivers before we ever see enough VSWR to trip an alarm or start a power reduction.
WORRYING ABOUT HEADROOM
To solve this problem, we could put a sample loop on the output of the hybrid combiner and use that sample to drive the adaptive equalizer and compensate for the problems created by the high VSWR. The problem I have with that is, how hard should an adaptive equalizer be working under normal conditions? How much headroom is there in the equalizer? At the BTS Symposium a couple of years ago, Bob Plonka of Harris Corp. presented a paper that made it appear that adaptive equalization could take care of virtually any problem in a transmission system. At this year's symposium, Kerry Cozad of Dielectric presented a paper that seemed to indicate that we shouldn't worry too much about high VSWR numbers in a system.
These are a couple of very smart engineers that I respect and maybe I am making too much out of this. I'm going to have to research both of these papers to try and figure out what the appropriate amount of worry is for this system. It just seems to me that we are potentially compromising sound engineering because we can fix the problems with external correctors. However, there are no free rides. So what am I compromising now and in the future? IPTV plans to make this DTV assignment and antenna our permanent home when analog is shut down. That means our expectation is that the system will perform for the next 25 to 30 years as the existing analog system has. Shouldn't it be as technically sound as possible so that the system has room to age without system failures with minor changes that are inevitable on a 2,000-foot tall tower in Iowa?
The question that I have to get answered in short order is, given the factory specifications of the antenna and the measured performance of the transmission line, how do we end up with a VSWR that is significantly worse than either of the two figures would indicate? The answer to that question is critical because at this point I figure either the antenna or the transmission line has a problem, and since they were both supplied by the same manufacturer, they should play well together. So, back to the vendor's question: What will I be satisfied with? That's easy: deliver a system that meets the specifications that were the basis for our purchase. Give us what we paid for!
