Designing Broadband Autotransformers for HF
WA2WVL Email: wa2wvlfk@hotmail.com 11 Sept 2006 rev 5
Introduction:
Since 1959 * many articles have appeared in technical journals,ham radio magazines and other publications describing the design and construction of broadband ferrite power transformers. The majority of these articles are predicated on the use of “Transmission Line” techniques to attain maximum bandwidth. A second technique, about which very little has been written, is the Broadband Autotransformer. With this approach high power transformers can be built covering 1.8 to 30 MHz and beyond.
Transmission line transformers have several problems not present in autotransformers.
Very few impedance ratios are available. Ratios of 1:1, 2:1, and 4:1 are common but greatly limit their application.
Other ratios can be done but are overly complex.
Insertion loss cannot be designed and usually is too high for a high power (1500watt) transformer.
An autotransformer can be designed for a 50 ohm input impedance to match any load from 2 ohms to greater than 100 ohms. Also it can be designed to have an insertion loss lower than 1% (.05db) of the power being handled.
The key to achieving this performance is maximum mutual coupling between all turns of the transformer. One would think that toriod construction would have tight mutual coupling but in actual fact the low permeability of common materials used at HF frequencies allows a great deal of leakage flux between turns and results are very poor.
The most successful construction technique has been found to be the ferrite loaded (copper or brass) tube with the additional turns wound thru the hole.
In this article I will deal with transformers designed to handle 1500 watts although the technique should work even better with smaller cores, tubes, and shorter wire lengths. The steps to design this transformer are as follows.
Step #1 Decide on the frequency range to be covered and the average power level, in this case 1500 watts. For some applications a single frequency or ham band may be all that is required, such as feeding a short vertical antenna.
Step #2 A design example will be done in the following steps for a 50 to 75 ohm transformer to be used at both ends of a CATV line to get a 50 ohm system for 1.8 to 30 MHz.
Step #3 Find required turns ratio. Figure 1 is a schematic of the autotransformer that is being designed. The one turn tube is at the top of the transformer. To transform 50 ohms to 75 ohms the required turns ratio is square root of 75/50 or 1.225:1
Step #4 Find integer turns giving a close match to desired ratio. Looking at the possible choices 11 to 9 is very close (ratio=1.222).
Step #5 Compute how many cores are required. For KW class transformers I have settled on a core (Fair-Rite Products 2643101902) which is 1.122 inches in diameter, 1.125 inches long with a .543 inch hole (a good fit for .5 inch tube). In Table 1 is shown the core loss of this core on 160, 80, 40, and 20 meters. You will see that the highest core loss occurs at 1.8MHz. A conservative design would be to allow 2-3 watts per core or about 54.8 volts across 4 cores or 13.7v/core. At 1500 watts the 50 ohm tap will have 273.8 volts or with 9 turns, 30.4 volts per turn. With 4 cores, only 7.6 volts per core will result (well under the 13.7v/core for 2 watts loss per core).
Step #6 Determine what maximum wire size can be used. In this design 10 wires must go thru the tube and should not be a jam fit. Teflon wire is preferred since it can operate at high temperature and is fairly tough mechanically. The 600 volt type is adequate for this design. About 7 feet of #16 was used.
Step #7 Build the core assembly. Two lengths of .5 inch thinwall brass tubing were cut to 2.520 inches. A pair of #12 buss rings were made to go on the front ends of the brass tubes. The tubes were loaded with the cores and a #12 buss wire was wound between the back ends of the tubes to connect them together.
The cores stacks were taped together. A flaring tool was used to roll over the ends of the tubes and then all the buss wire was soldered to the tubes.
Picture 1 shows the finished core assembly. Be sure that the ends of the tubes are smooth and flared if possible so that the Teflon wire is not cut.
Step #8 Determine how to mount the core assembly and connectors so that lead length is minimized. I mounted everything on the lid of a box with a wide strap between the connector grounds. A plastic box (such as home depot sells) is adequate. If a metal box is used a .25 inch insulator should be between the cores and the metal.
Step #9 Wind the transformer starting at the top of the winding. First one turn, and then the remaining 9 turns. Be sure that all turns are wound in the same direction.
Step #10 Now that the transformer construction is complete, terminate the 75 ohm side with small carbon film resistors.Measure input VSWR on the 50 ohm connector at all of the frequencies of interest using an MFJ-259B bridge or other equipment.
Step #11 VSWR was low over most of the frequency range but if desired an "L" network (highpass) can be used to improve the match at the low end. This is shown in Table 2.
Step #12 Last but not least, Test under power, especially at the end frequencies. Loss at 1.8 MHz is due to higher magnetizing current (which decreases with increasing frequency) while loss at 30 MHz is more complex and is affected by capacity to ground and dielectric loss of the cores.
Picture 2 is the finished assembly using “N” connectors.
The measured results are shown in table 2.
Some Broad-Band Transformers by C. L. Ruthroff
Proceedings of IRE, Vol. 47, August 1959 Pages 1337-1342