RF Column 32 - June 1994 Copyright (c) 1994,1995 H. Douglas Lung ALL RIGHTS RESERVED TOPICS: Report on two unpublished NAB 94 Engineering sessions Power line harmonics - ignore them at great risk! Update on a truly digital TV transmitter from Tim Hulick Tips on curing UHF tube harmonic problems from Walter Pries ------------------------------------------------------------------------------------------------------- This month I planned to bring you information on our network's installation of the Scientific Atlanta Digital Video Compression system and some information on a new, economical LPTV transmitter I saw at NAB. Well, as I approach the deadline for this column, I haven't received the test results on the LPTV transmitter and the first part of the compression gear just arrived. The software for the compression system is still being tweaked, but I'm hopeful I'll be able to give you a preliminary report next month. Last month I didn't have room to include all the NAB engineering conference items I wanted to, so this month I'll give you a synopsis of two papers which don't appear in the NAB Proceedings. One covered power line harmonics, the other was Dr. Timothy Hulick's update on his digital TV transmitter. I also have some advice on curing UHF klystron harmonic problems courtesy of Walter Pries. Hidden in the HDTV Station Issues session at NAB Sunday afternoon was an interesting presentation titled "Dealing with Power Line Harmonics in Broadcast Facility Design." Richard Stephen outlined the problems that may arise when electronic gear is hooked up to the power line. Problems? What problems? Exploding transformers and hot neutral wires are two of the problems power line harmonics can cause. Harmonics are created when the load on the power line draws power only from the peaks of the AC sine wave. A common myth is that switching power supplies in equipment is responsible for most harmonics. Actually, any electronic device using solid stuite detailed. Here's my quick interpretation of how power line harmonics are generated. As you will recall from basic electronics, Fourier analysis shows that a wave shape anything other than a perfect sine wave must contain energy at harmonic frequencies. In an AC to DC power supply, the output of the rectifier is connected to a filter capacitor. The rectifier will conduct and draw current from the AC supply only when the instantaneous voltage of the AC exceeds the voltage on the capacitor plus the voltage drop of the diode. Therefore, the entire time the AC waveforms voltage is less than the DC voltage on the capacitor no current is drawn. The current waveform won't look anything like a sine wave - it will be a series of short, steep pulses rich in harmonic energy. How does this cause problems? Lets ignore for a moment the harmonic content of the pulses. The peak current of the pulse will be significantly higher than the average current you'd read on an RMS meter. While the RMS meter gives a good reading for the average current (after all, RMS stands for "Root Mean Squared") regardless of the wave shape, it doesn't reflect the peak load placed on the transformer, wiring and other equipment on the line. You may have seen cases where a cheap peak reading analog meter showed currents so far off that you figured the meter was defective. Non-RMS meters use a conversion factor based on a pure sine wave. Feed them something different and that factor doesn't work anymore. In order to properly operate electronic equipment with DC power supplies, the generator, transformer and wiring must be able to handle these peak currents. Furthermore, non-symmetrical peak currents will induce current in the neutral wiring to the transformer, even though average curr ent readings show the load is balanced. Richard Stephen outlined a few horror stories from facilities with high harmonic currents. Some of the problems he found included a generator whose crankshaft broke, a power company transformer outside a computer manufacturer's plant that kept exploding, a TV station where DC power supplies wouldn't stay in regulation and motors that burned up. Most heavy duty motors run on three phase AC. When the AC waveform to these motors contains harmonic energy, that energy goes to trying to turn the motor in the opposite direction! The result? Heat. Harmonic currents have the same effect on generators (alternators). As harmonic energy increases on an AC waveform, the peak of the waveform becomes smaller -- not necessarily lower in voltage, just narrower. If it doesn't have the energy to charge the capacitors in a DC power supply above the minimum required by the regulator, the output will contain hum. Why hasn't this been a hot (please excuse the pun) topic in the past? Richard mentioned one reason. Older power supplies, particularly transmitter power supplies, used tube rectifiers. These rectifiers and the choke input filters that usually follow them present a much higher impedance to the power line and thus create less severe peak currents. Take out the old tube VHF transmitter and put in a nice solid state transmitter and now there is a problem. DC power supplies replace the clean loads such as filaments and big cooling motors. In studio facilities, I believe the problem hasn't shown up as much in the past because older gear was replaced with equipment that drew much less power. However. as electric typewriters are replaced with computers and laser printers the percentage of AC power being converted to DC power is increasing. Harmonic currents increase and the peak current increases. When the peak current reaches the limits of the transformers and switch gear the problems will start. What can be done to reduce power line harmonics? Richard Stephen mentioned that governments in Europe are regulating how much harmonic current a device can generate and how much harmonic energy a business can place on the power line. That hasn't happened yet in the United States. I have noticed more power supply manufacturers are including harmonic current in their specifications and advertising. Some U.P.S. (Uninterruptible Power Supply) systems use zig-zag filters to reduce harmonic energy, but Richard dismissed them as not being the answer - they can actually increase the currents in some cases. The final answer is to build electronic equipment that doesn't generate harmonic currents. In the real world, the answer is to design your electrical system to handle the harmonics. Some of Richard Stephen's suggestions included doubling neutrals (not the size of the neutral wire, rather run two wires of code size) and oversizing the generator, if one is used, by as much as 100% depending on the type of load. If a U.P.S. is used, it too should be oversized and have input filtering. Before accepting a U.P.S., take a look at the output waveform while under full load -- it should be reasonably clean. Manufacturers have started building transformers designed to cope with high harmonic currents. These transformers are called "K rated" transformers. For a broadcast facility you'll need a K rating between 14 and 15. If the load is predominantly electronic, a K rating of 20 may be required. K, by the way, is calculated using the sum of the squares of the current at each harmonic times the square of the harmonic. You'll be seeing more discussion of power line harmonics as the use of electronic equipment grows. For the last two years I've been following Dr. Timothy Hulick's work at Acrodyne on his high power digital transmitter called "ADAM". It seems that at least once a year Dr. Hulick delivers an update on how this project is progressing. As usual, Dr. Hulick's paper at NAB this year included a lot of background material his method of switching multiple amplifiers at various power levels on and off with digital data from an analog to digital (A/D) converter -- essentially building a high power D/A converter. He has written numerous articles on this and I've covered it here, so I won't repeat it. What he has found is that he can increase the efficiency of the transmitter by an extra 10 percent by pulsing the most significant bit amplifier. In a standard video transmitter no video exists above the tip of burst (looking at it from a transmitter power standpoint -- as a video baseband signal I would say "no video exists below the negative tip of burst). Therefore, most of the time the transm itter only has to generate the signal from white to the extreme tip of burst. During the limited time when more power is required for the sync pulses enough extra power can be added to make a sync pulse reach 100% peak power. Dr. Hulick also explained that in his work at Acrodyne on "ADAM" he has developed a way to generate RF HDTV signals. In addition to controlling amplitude via the video A/D converter he can control the phase of the carrier as well, making it suitable for QAM. If you think about it, there is no reason why this transmitter, which is essentially a big analog to digital converter coupled to a digital to analog converter capable of producing 1,000 watts or more, shouldn't be able to produce any waveform. One of the major problems with the existing ADAM is that high power filtering is required to reduce the lower sideband to levels in compliance with the FCC's response for TV VSB (vestigial sideband) transmitters. At the NAB Convention, Dr. Hulick said he found a way to generate a VSB signal without the filtering. He promised to release details after the patent application is filed. If Acrodyne and Dr. Hulick can build a transmitter that can efficiently produce any RF modulation scheme they will have a made a major break-through that will have applications well beyond TV transmitters. More harmonics... Walter Pries, an RF engineer with DSI Communications, called me recently to relate some experiences tracking down UHF transmitter harmonics. UHF transmitter second harmonics can fall in the same frequency range used for aircraft transponders. Third harmonic energy can interfere with broadcast microwave systems in the 2.0 Ghz. band. Measuring harmonics at from high power transmitters at UHF frequencies is difficult, if not impossible, to do by sampling signals from the transmission line (or waveguide). In large diameter transmission line or waveguide, the efficiency of the coupling probe will vary unpredictably at frequencies above 1 Ghz. At gigaHertz frequencies, 6-1/8 inch line starts to look like waveguide. The only way to get a true measurement is to do it the way the FCC does - measure the transmitter harmonics over the air, using a dipole cut for the harmonic frequency. Broadband fan dipoles will work for less critical measurements -- I show you how to build one in a future column. Once measured, what can you do if UHF transmitter harmonics are problem? Walter Pries offers this suggestion. If the transmitter uses multiple output amplifiers, see if one of the amplifiers is creating most of the harmonics. You can do this by monitoring what happens to the second harmonic as the amplifiers are switched off. If switching off one amplifier causes a dramatic drop in the harmonic, work on that one first. All UHF klystron TV transmitters I've seen have low pass filters on the output of each klystron. Walter has found that under certain situations the standing wave ratio (SWR) caused by reflections from the antenna or diplexer can effectively cancel out one or more slugs in the low pass filter. He's found that simply rotating the low pass filter -- that is, swapping the input and output -- will usually move the slugs enough to eliminate this effect. Walter Pries is one of those rare engineers that knows how RF behaves and, perhaps more important, how to control it. You can reach him through DSI RF Systems at 1-603-654-6800. If you have any tips useful to other RF engineers working with TV transmitters, microwaves or satellites, I'd like to share them. Give me a phone call at 305-884-9664 after 6 PM Eastern Time or send a fax to 305-884-9661. You can also leave notes on CompuServe -- my ID is 70255,460 or via Internet -- use 70255.460@compuserve.com. Speaking of CompuServe, you'll notice the old version of the Cheap Remote software is still posted. I found some bugs in the video switching part of the software. I won't post it until I have the bugs killed and have had a chance to fully test it. The main part of the software is working fine - it just paged me to let me know the visual power at an LPTV site dropped one percent below my alarm limit. If you'd like a copy of the software in its current form, drop me a note. Next month look for the items delayed from this month (I hope!) and a report on databases and maps available on CD-ROM. Of three mapping packages I tried, I found only one that really works for plotting coverage maps and locating transmitter sites. Copyright (c) 1994,1995 H. Douglas Lung ALL RIGHTS RESERVED