Issue #116 (Web Version)
SILENT MODULATORS
It is with deep regret that we report the passing of two prominent members of the AM community.
Roger Frith, N4IBF, co-founder and former co-publisher of The AM Press/Exchange died on July 29, after suffering a long illness. He will be missed on 75 and 160 metre AM, as well as on most of the HF CW bands.
Roger's collaboration made publication of the early issues of AM P/X possible. Although he was no longer connected with our routine production, he retained a strong interest in amateur radio AM, and generously offered his assistance and advice whenever possible. In recent years, his amateur radio activities were eclipsed by his interest in early printing technology, but he remained active on AM, using what was often described as the best sounding BC-610 ever heard on the air.
Roger was working in a print shop when we first contemplated starting this newsletter. He was strongly convinced that the AM community needed its own dedicated publication. He agreed to print the first few issues using shop facilities after hours, donating his own time and materials until we had enough paying subscribers to cover the costs of production and printing. He continued to help with the collating, folding and stamping for several years, and was instrumental in acquiring the collating and stapling machines we still use today.
Roger became quite an expert in early printing processes and worked as curator of science and technology at the Tennessee State Museum, where he extended its collection of artifacts relating to the early printing industry. He taught himself printing techniques through study and experimentation, and his nationally recognized expertise in the field led to his being a visiting fellow twice at the Smithsonian Institution. He built up a collection of materials that gave the museum a specialty in the printing industry that is probably the best in the Southeast.
Vestal "Les" Lester, K6HQI, known throughout North America as net control on 14.286, became a silent modulator this summer. He used a homebrew transmitter with 833-A's modulated by 833-A's. His gentlemanly mannerisms and consistently loud, high quality signal will be missed by all that frequented the 20-metre AM group. In his memory, the 14.286 net has been named "The K6HQI Memorial 20 Meter AM Net."
CIRCULAR POLARIZATION ON HF
Bob "Bacon" Bruhns, WA3WDR, 8 Paulding St, Huntington, NY 11743-1960
A few years back, I read an interesting article, "The Enhancement of HF Signals by Polarization Control" by B. Sykes, G2HCG, Communications Quarterly, November, 1990 (reprinted from Practical Wireless).
Sykes found that most of the short-term fading seen on 10-meter beacons is actually due to polarization rotation of the incoming signal causing occasional cross-polarization nulls with ordinary receive antennas, and not due to real fading of the signal itself.
Sykes's findings agreed with some lower-band antenna stories I had heard over the years, so I set up a circularly-polarized antenna on 75 meters to see how it would work. I found that some big improvements were possible from some relatively simple antenna hardware.
A CIRCULARLY POLARIZED ANTENNA FOR 75 METERS
Basically, what I have is a 75-meter turnstile antenna, but because of space restrictions it actually is two vertical half-wave loops, top-fed, at right angles to each other. In fact, space is so restricted that the ends of the dipoles don't just meet, they overlap. This antenna configuration may be a factor in the results I got.
Both antennas are coax-fed with W2DU ferrite-sleeve baluns. The coax lines (RG-8X) are both 100 feet long, just about 1/2 wave electrically. The feed ends of these feedlines are connected together with about 40 feet of RG-8 (1/4 wave electrical). By connecting the transmitter or the receiver to the junction of the l/4 wave phasing section and one of the antenna feedlines, I am able to feed the two antennas 90 degrees out of phase. The circular "sense' is determined by which junction I select.
I built a special T-R switchbox which allowed me to select which junction to use as the feedpoint. After a little use, I modified the switchbox to allow independent TX and RX sense selection, because I found that optimum TX was usually the opposite connection from optimum RX, at least for local TX. (I haven't tried DX yet.)
THE RESULTS
In RX tests, circular polarization was noticeably better than either linear polarization most of the time, and one circular sense was clearly better than the other. The optimum circular sense always reversed at sunrise and again a couple of hours after sunset, and while it was reversing, there was no clear advantage to either sense. On nights when 75 meters "went long" (when the high-angle MUF dipped below 3.9 MHz), the optimum sense for local signals would reverse again as the signals began to fade out.
When the optimum circular sense was selected, receive signals were significantly stronger and suffered less fading and less selective fading. Receive signals often had substantially better audio quality, as if a "phasing" sound had been removed. Unfortunately, severely disturbed ionospheric conditions would still mangle receive signals, regardless of the antenna sense.
The optimum receive sense for European DX was the reverse of the optimum receive sense for signals within about a 500 mile radius, except when the band went long, and the optimum local sense reversed.
Click for Figure 1 (Two loops, concentric and perpendicular)
In transmit tests, I was received much better on one circular sense than the other, the difference being about 10 dB. During the day, optimum TX was always on the other end of the 1/4 wave phasing line from optimum RX. At night, this was usually true, but there were occasions when optimum TX was obtained with the same connection which produced optimum RX.
While other stations were exhibiting short-term fading of 20 dB, my TX signal had almost no short-term fading. I was fading slowly, with little selective fade effect. Reports indicated better quality audio in the optimum TX sense; evidently, the reduced "phasing" effect I heard on RX was evident on TX as well. Signal reports were good, too, especially considering my 80 watts output and tiny antenna.
It would be interesting to run tests with another station using circular polarization. As far as I know, mine was the only station using circular polarization in these tests. All of the other stations used linear polarization.
CIRCULAR POLARIZATION AND SENSE
Circular polarization consists of an equal amount of electromagnetic field polarized in two directions separated by 90 degrees, such as vertical and horizontal, with a 1/4 cycle time difference between them.
The concept of polarization becomes more complex when a signal is coming from somewhere in the sky instead of somewhere on the horizon. Polarization will still be at right angle to the direction of propagation, but it is no longer a simple matter of vertical versus horizontal. For example, the polarization of a very high angle skywave signal is essentially horizontal - but the wave might be north-south polarized, or east-west, or something in between; it could also be circularly or elliptically polarized, in either sense.
The electromagnetic field of a circularly polarized wave rotates once per cycle. A 75 meter signal might rotate 3,885,000 revolutions per second! As the cycle progresses, positive would be up, then around to one side, then down, then around to the other side, then up, etc. This rotation can be either clockwise or counter-clockwise, and this is what is meant by the term "sense".
When a circularly polarized antenna is used to receive a circularly polarized signal, the antenna sense must match the sense of the incoming signal, or very poor reception results, very much like the situation of a horizontal antenna receiving a vertically polarized signal. Such cross-polarization results in large signal losses, theoretically infinite, but typically around 20dB in practice.
ELLIPTICAL POLARIZATION
Circular polarization is a special case of elliptical, in which the fields are exactly equal and exactly 90 degrees out of phase.
If the horizontal field and the vertical field are unequal in strength, or if the phase difference between the horizontal and the vertical field is not exactly 90 degrees, the resulting polarization is called elliptical. Elliptical polarization lies between linear polarization and circular polarization, something like an oval. Elliptical polarization also has a sense, but it tends to favor one angle of linear polarization.
Linear polarization, such as vertical or horizontal, is a special case of elliptical, in which all the power is in one polarization direction.
SELECTIVE FADING
Although groundwave can contribute to selective fading on amateur frequencies, usually it is too weak to make much difference unless the path is very short, or over salt water. Selective fading in the amateur bands is primarily ionospheric.
There are usually several ionospheric paths a signal can take which will deliver usable power to a receive location. Portions of the signal which follow these various paths arrive with different delays. Selective fading results when these variously delayed signals recombine.
Generally, the signal from one of these paths will be the strongest, and normally it will dominate. As long as this signal does not fade, reception is good. But when this strongest signal fades, it no longer dominates. Weaker signals from one or more secondary paths interfere significantly, and selective fading results.
Under normal ionospheric conditions, most of this apparent short-term fading of the strongest signal is actually due to random polarization changes. Except for the hours around sunrise and a little after sunset, this predominant signal seems to have a generally consistent sense, varying in ellipticity from linear to circular. The use of circular polarization of the correct sense will result in optimum transmission and reception, with minimum selective fading.
MAGNETIC EFFECTS IN THE IONOSPHERE
It has long been known that the earth's magnetic field causes the ionosphere to twist the polarization of radio waves it returns to the earth. This can have strange effects.
Imagine that the polarization twist on some signal path holds steady for a while at 45 degrees clockwise, and that the antennas of two stations also differ in direction by 45 degrees. Both stations are readable to an observer with circular polarization. Station one will hear station two just fine (perfect polarization alignment), yet station two will hardly hear station one at all (crossed polarization). This is an example of non-reciprocal propagation caused by magnetic effects in the ionosphere.
There is a great deal of theory about exactly how the ionosphere acts in the earth's magnetic field, but for most of us poor hams, the most valuable information is that ordinary fading can be reduced, and improved communication can be achieved, by using a circularly polarized antenna with switchable sense.
ABOUT THAT ANTENNA...
My crossed loops don't produce perfect circular polarization at all angles of radiation. Offhand, the only type of antenna I know of which comes close is the quadrifilar helix. I considered building one of these for 75 meters, but I can't install structural supports for one here, and reversing its sense would be tricky. This is a field for further study.
If I ever get enough room, I would really like to try a pair of full-wave "CCD" antennas in a 90-degree-apex inverted-vee turnstile configuration. "CCD" antennas are full-wave antennas with approximately constant current distribution along their length, due to numerous series "unloading" capacitors along their length. A pair of half-wave inverted-vee dipoles in turnstile configuration may be a good solution for those with smaller lots.
But since my yard is small and uncooperative, I pretty much have to settle for bending my turnstile antenna into the dual half-wave folded loop form. The impedance of each folded loop appears to be around 17 ohms, which results in mismatches, odd TX loading, and unequal power fed to the two loops. Even so, this antenna provided improved reception and produced a strong and steady signal on the air.
I have considered using 6:1 step-up baluns at the antennas instead of 1:1 baluns; this would step the feed impedance up to around 100 ohms per loop, and then I could use 95 ohm cable for the feed and delay lines. The impedances would match, TX power would split evenly, and the TX load would be right around 50 ohms.
CONCLUSIONS
I always thought fading was an overall ionospheric effect, independent of the antenna; I thought the only possible improvement would come from complex, multi-receiver diversity systems. Not anymore!
It seems that circular polarization is a simple and natural method of improving signal propagation on HF. This is not new knowledge, but I sure didn't know about it, and I don't see it in the popular books on amateur antennas for HF.
CREDITS
I've already noted the work of B. Sykes, G2HCG. There is coverage of ionospheric polarization twisting in Terman's Radio Engineer's Handbook (McGraw-Hill, 1943) and Radio Engineering (McGraw-Hill, 1947). Several people told me about improved HF communications when two antennas of different polarization were used together. More than twenty years ago, I heard about one ham with a 75 meter turnstile antenna, who noticed reduced selective fading. While I was testing my antenna in early 1991, people broke in and told me details about circular polarization and the ionosphere. Thanks to all who participated and offered reports, suggestions and ideas.
 | Click here for a table showing terminal arrangements for UTC CVM-0 to CVM-5 Inclusive! |
High Quality MOSFET Audio Driver for Class B Tube Modulators
By Steve Cloutier, KA1SI (formerly WA1QIX)
Copyright 1995 S. R. Cloutier
This document outlines the design and implementation of an output transformerless audio driver for class B modulator grids. The exact circuit shown is used to drive a pair of 811-A triodes, however the same circuit can be used to drive class B modulator tubes with drive requirements of up to 500 V or more peak to peak such as a pair of 833-A triodes by simply increasing the power supply voltage.
Design Considerations
The major problem in driving class B grids, where the grid is driven positive with respect to the cathode, is the widely varying and nonlinear load which the grids present to the driving circuit. During the periods of the audio cycle where the grid is negative with respect to the cathode, the grid impedance approaches infinity. However, as the audio cycle continues, the grid voltage rises and eventually goes positive with respect to the cathode causing a sudden change in the grid impedance as the grid begins to draw current. Furthermore, the current change is nonlinear, rising rapidly as the voltage is increased.
A driver circuit can be thought of as a perfect voltage source with a series output resistor. The output resistor represents the driver resistance (or impedance). Due to the non-linear nature of the grid impedance, this resistance which exists in the driving circuit will cause distortion of the grid waveform since the voltage drop across the driver resistance varies with the grid current, which is non-linear.
The best way to eliminate this problem is to reduce the driver's internal resistance to as low a value as possible.
The circuit shown accomplishes this by using a Source Follower as the output stage to drive the grid load. The output resistance of a source follower is very low, and the source follower used in this circuit can deliver over 1 Ampere of peak current to the grid load. The output resistance is in the order of a few ohms.
Circuit Description
The implementation shown uses a plus and minus 160 volt supply, and will deliver almost 300 V peak-to-peak. The voltage can be increased to plus and minus 300 volts, which will allow the driver to deliver more than 500 V peak-to-peak. Changing the supply voltage is discussed later.
The input transformer, T1 is used simply to isolate the minus 160 V supply voltage from the input. A UTC A-series transformer will work quite well in this application, however any good quality low level transformer will suffice. An input level of 2 V peak to peak should be sufficient.
Starting at T1, the signal flows to Q1 which acts as a phase splitter. From there, the signal is amplified by Q2 and Q3, which deliver around 15 V peak-to-peak to the gates of Q4 and Q5. Q4 and Q5 further amplify the signal to a voltage approaching the plus and minus power supply voltage. The 12 volt zener diode connected from the gate circuit of Q4 (and Q5) back to the source protects the MOSFETs from the possibility of damage due to high voltage at the gate, since the MOSFET gate may be destroyed if the voltage from gate to source exceeds 20 V.
Click here to see a schematic of the transformerless Audio Driver
From Q4 and Q5, the signal flows to source follower Q6 and Q7, which provide the high current necessary to drive the grid load.
The grid bias of the modulator tubes is derived by establishing a DC offset at the gates of Q6 and Q7, controlled by the bias pot R3.
The voltage at the drain of Q4 (and Q5) is controlled by the 250K resistor, R1 (and R2). This voltage should be set to be equal to the output voltage. The two 12 V zener diodes in series, shunted by the .15 uF capacitor allow some variation (approximately 12 V) between the voltage at the drain of Q4 and Q5, and the voltage at the gates of Q6 and Q7. This will prevent small circuit drift caused by temperature or minor power supply output changes from changing the bias applied to the modulator grids. If the difference between the voltage at the drains of Q4 and Q5 and the gates of Q6 and Q7 is greater than 12 V, then R1 and/or R2 should re-adjusted.
The 12 V zener diode connected from the gate of Q6 (and Q7) to the output protects Q6 (and Q7) from the possibility of damage due to excessive voltage from gate to source. The 1/2 Ohm resistor between the source of Q8 (and Q7) and the output, and transistor Q8 (and Q9) form a current limiting circuit. If the current across the 1/2 Ohm resistor exceeds 1.5 amperes or thereabouts, the voltage drop across the resistor will be sufficient (more than .65 volts) to turn on transistor Q8 (or Q9). At this point, the transistor will clamp the input voltage, preventing any further increase in output current. If the overload or short circuit is sustained, the output fuses will blow, protecting the MOSFETs from overheating. Without this circuit, a failure in a tube, or short circuit could destroy the output MOSFET Q6 or Q7 when the current rating of the device is exceeded.
The negative feedback circuit (outlined in broken lines in the circuit diagram) is optional. The value of the 2 meg Ohm resistor should be selected to suit the particular modulator circuit which you are using. In the circuit shown, the current flowing through the 2 meg Ohm feedback resistor is around .5 mA. The amount of negative feedback is controlled by adjusting R4. If you do not choose to use a negative feedback circuit, do not include any of the components shown within the broken lines in the circuit diagram.
There are a number of reference voltages shown in the circuit diagram. These reference voltages are useful for verifying that the low level circuitry is assembled properly and operating correctly. All of the voltages shown which are enclosed in rectangular boxes are with respect to the "0V Ref." point shown on the diagram. This reference point is connected to the negative power supply voltage.
This circuit uses the Motorola MTM-4N85 850 volt, 4 ampere MOSFETs. Other MOSFETs may be used, provided they will handle the combined plus and minus power supply voltage. Other transistors may also be substituted for the 2N2219-A.
Changing the Supply Voltages to Drive Larger Tubes
The power supply voltages used in this circuit may be increased to allow the driver to deliver a greater peak-to-peak grid voltage and/or to supply a greater negative bias voltage. If the bias requirements are greater than minus 20 volts, you will need to change the zener diode D1 to a larger voltage, or put additional zener diodes in series with D1. You should also increase the value of the negative power supply voltage if the peak grid voltage plus the bias voltage exceeds the negative supply voltage.
As an example, if you wanted to drive the grids of class B 833A triodes, you would need a minimum of 400 Volts peak-to-peak, and a bias voltage of up to minus 100 volts.
The power supply voltage in this case should be increased to plus and minus 300 volts, and the zener diode D1 should be increased to a 100 Volt zener diode, or several lower voltage diodes should be used in series. This should allow the driver to deliver around 500 Volts peak-to-peak, while supplying a negative bias voltage of up to minus 100 Volts.
If the voltage is increased above plus and minus 200 V, you should increase the power rating and the resistance of the 20K and 10K power resistors used in the MOSFET output circuit. The zero-signal current flowing through the 10K resistors connected to the drains of Q4 and Q5 should be no more than around 15 or 20 mA.
With a plus and minus 300 volt supply, the value of the 10K resistors should be increased to around 20K Ohms. The value of the 20K resistors connected between the sources of Q6 and Q7 and the negative power supply should be increased to around 35 or 40 K Ohms, keeping the current flowing through Q6 and Q7 at around 10 mA.
When using a much higher supply voltage, you may not be able to obtain the correct drain voltage at Q4 (and Q5). If this turns out to be the case, adjust the size of either R1 and R2 to a larger value, or increase the size of the 330K resistors connected to R1 and R2.
Practical Considerations
The plus and minus power supply must be capable of supplying the idle current of the circuit, and the plus power supply must also supply the grid current to the modulator tubes. In this circuit, the plus and minus power supply is capable of supplying 80 mA continuous current.
The voltage stability of the positive side of the plus and minus power supply is very important. If the supply voltage "sags" during periods of high modulator grid current, the voltage change will produce distortion of the modulating voltage. If the power supply voltage sags, you should use a simple regulator in the positive power supply.
The circuit shown here is an example of a simple regulator which you can use to keep the positive power supply voltage from fluctuating under load. The MOSFETs in this circuit, although not dissipating large amounts of power, will need a proper heat sink. The power dissipation can be calculated by multiplying the voltage drop across the device by the current flowing in the circuit. In this circuit shown, the power dissipated in Q4 and Q5 will be around 4 watts for each device. The power dissipated by Q6 and Q7 will be greater due to the grid current which must flow through Q6 and Q7. I used a 50 watt heat sink for each device Q6 and Q7, and put Q4 and Q5 together on a single 50 watt heat sink.
If the power supply voltages are increased, the device dissipation should be re-calculated and the appropriate heat sinks should be used to keep the MOSFETs from overheating.
Setup and Adjustment
Warning: This driver is capable of supplying a considerable amount of current to the grids of the modulator tubes. Improper adjustment of the driver when connected to the modulator grids may result in damage to the modulator tubes if the misadjustment results in the driver supplying a high continuous positive voltage to the tube grids.
When you operate the driver for the first time, do not connect the grids of the modulator tubes to the driver. It is also highly recommended that you bring up the voltages slowly, using an adjustable autotransformer (or Variac). As you bring up the line voltage, check some of the voltages within the driver circuit to make sure that there are no wiring errors or defective components in the driver.
The first thing to set is the drain voltage of Q4 and Q5. Using a volt meter connected between the drain of Q4 and chassis, adjust R1 so that the voltage at the drain of Q4 is equal to the steady state voltage you want to apply to the modulator grids (the output voltage). Do the same for Q5 by adjusting R2. Note that the bias control R3 will not have much effect on the output voltage until the voltage at Q4 and Q5 is adjusted properly.
Once the voltage at Q4 and Q5 is set up, you should be able to make fine adjustments to the output voltage using the bias control R3. After adjusting R3 to obtain the desired output voltage, go back and check the voltage at the drain of Q4 and Q5, and re-adjust R1 and R2 if necessary to set the drain voltage equal to the output voltage.
NOTE: if the voltage at the drain of Q4 or Q5 differs from the output voltage by more than 12 V, the bias control will have very little effect on the output voltage, and the output voltage regulation will degrade significantly.
Once set up, the driver should be quite stable, since MOSFETs have very low temperature drift.
After the DC adjustments have been completed, the driver is ready to be connected to the modulator grids. If you are using the optional negative feedback circuit, you should first operate the driver and modulator without the feedback connected, to ensure that the driver and modulator are working properly. After this, you should connect the feedback with R4 set to maximum resistance (minimum feedback). Adjust R4 to obtain the desired feedback level. To reverse the feedback polarity, simply reverse the output connections between the modulator grids, in case you obtain positive feedback, and the modulator oscillates!
More Information and Feedback to the Author (over the Internet)
If there is sufficient interest, I can create a database of articles and information about AM in general, and establish a network based conference (forum) on AM. If anyone is interested in this, or has questions or comments about the circuit presented in this article or would just like to say hello, I can be reached over the Internet via E-mail at: cloutier@ostech.com. Also, using your Web browser, feel free to check out www.ostech.com:8080 and webnotes.ostech.com.