Dick Cappels' project pages

return to HOME

Driver and 1 Sq. Meter Loop for 187 KHz
A pair of bipolar transistors makes a 1 watt output stage that drives a 1 square meter loop antenna.

Notice: Before operating a radio transmitter, find out what kind of transmitter operation, if any, is permitted in your locality. Radio transmitter operation is a serious legal matter. In the United States, operation of unlicensed intentional radiators is covered by Part 15 of Title 47 of the Code of Federal Regulations. This design can be readily adapted to different frequencies and different power levels. If you choose to build and operate the transmitter described here, you do so at your own risk. I'm only publishing this as an example of what can be done.


This page describes an antenna, a 1 watt final output stage, and a transformation network for radio experiments at 187 kHz.

The Antenna


The Bangkok antenna was made with a frame of 1/2" PVC pipe and running a length of 0.5 mm square stranded copper wire (1 conductor of zip cord) through it. 90 Degree elbows are at three corners and a "T" is at the fourth corner (lower right-hand corner in this photograph), providing an exit for the 50 cm wire leads. A second antenna made in Arizona looked the same, including the slightly un-square look, but is white.

I built a loop antenna that is 1 meter on each side. The reason for building this first antenna was to see if I could successfully load the transmitter into it. Why one meter on each side?) PVC pipe comes in 2 meter lengths (in Bangkok, anyway. In Arizona the lengths of 1/2 inch schedule F pipe were a little more than  3 meters in length). My interest in loop antennae comes from the fact that they are much simpler and more compact than capacitive antennas. No capacitive top hat, no loading coils, and no ground radial system to worry about.

In the U.S., the Federal Communications Commission rules limit unlicensed transmitters in the 160 to 190 kHz band to a maximum input power to 1 watt and maximum antenna length of 15 meters. With such a  small antenna compared to wavelength, the radiation resistance would be very low, and real (Ohmic) resistance in the antenna would tend to be a lot larger than the radiation resistance, making the antenna even less efficient. This puts a premium on getting the real resistance of the antenna as low as one practically can. Some have made antennae  from copper water pipe and RG 8/U coax shield, and they would be  more efficient for their size than this one, which used 0.5 square mm copper wire.

 This 1 square meter antenna has radiation resistance of about 10 nanohoms, making the 0.4 Ohm resistance (including skin effect, which about doubled the resistance from the DC value) the dominant loss in the antenna. The lower the resistance, the more efficient it would be - and from these numbers, efficiency is nearly directly proportional to 1/R; cut the resistance in half and double the efficiency.

The antenna, which is a square loop of 0.5 square mm copper wire 1 meter on each side, measured 6.2 uH and under 0.2 ohms DC.

The Output Stage

The output stage can be recognized as basically the same as published by Murry Greeman, Lyle Kehler, and Bill Ashlock (possibly among many others).

This output stage has a rich lineage. Murray Greeman, ZLBPU is credited as the originator of the circuit as an LF output stage.  Bill Ashlock found that the output signal was cleaner when the transistors has separate base driver resistors. I chose to move the clamping diodes to the end of the base resistors connected to the transistor bases and eliminate the third resistor in the base circuit that was in common with both inputs.

I encourage you to do an internet search on
Murray Greeman ZLBPU, Bill Ashlock, and Lyle Kehler) for more information on Lowfer transmitter and antennas. Here are some URL's to get you started. Though these links worked on the day I created this page, I don't make any promises to keep them up-to-date.

I added a 2N7000 to gate the NPN output transistor as a means to modulate the stage with either CW (A1) or tone bursts (quasi-A2). I damaged several 2N7000's with ESD in earlier experiments and have started adding gate protection to some circuits. Here, an MPSH34 (chosen for no other reason than because I have a lot of them) is used. Signals below ground are clamped to one diode drop below ground. Signals more positive than about +6 volts are clamped by avalanching emitter-base junction. On the breadboard, the avalanche voltage was measured at 5.8 volts when the emitter was biased with 1.8 milliamps.

Raising the Antenna's Impedance so it can be driven

The reactance of the antenna at 190 kHz is = 2 x Pi x 190 kHz x 6.2 uH = 7.4 Ohms.  Counting the DC resistance of less than 0.2 Ohms and the skin effect about doubles the resistance. Since the resistive losses only increases the total impedance by 0.2% (Z = Sqrt(L^2+R^2), so I will keep it simple by ignoring it.  The resistive component largely determines antenna efficency as it is very large with respect to the antenna's radiation efficiency.

The the transistors in the output stage are only capable of driving up to a few hundred millliamps into the antenna, and with a load of 7.4 Ohms, saturation losses in the output stage would eat up most of the 1 watt input power permitted under FCC rules. The challenge became one of how to transform the antenna's impedance to one high enough for the output stage to drive. This challenge is the main reason for this experiment.

Note that there is no need to "match" the antenna to any particular impedance, only to raise its impedance to that which the small output transistors can drive comfortably. The overall objectives are to get as many amps of RF flowing in the antenna as possible given the constraint of 1 watt maximum into the output stage

I spent a lot of time on step-down transformers, but they all had high leakage inductance (that's uncoupled inductance between the primary and secondary) and they took a substantial amount of wire, so the Ohmic losses in wire were significant. Even exotic multifilar windings (up to 9 parallel windings to reduce leakage inducatance), as recommended by a knowledgeable friend, did not improve the leakage inductance enough.  

Finally, I decided to try a capacitive impedance transformation circuit as shown in the schematic. While I would like to say that I solved a system of equations and then ordered capacitors with the needed values,. In reality, I had to make do with parts in my junk box, and this driver produces square waves, not sine waves, so the formulae I worked out for only useful for the fundamental, and would completely miss the effects of the higher harmonics. For example, during the switching transitions, the output stage attempt to change the capacitance of the matching network instantaneously, which would result in very high charging current. That's what the 10 Ohm resistor is for, but the way, to limit current on those edges.

The method was to pick a pair of capacitances which, when put in parallel resonates with the antenna's inducance, resonates at 189 kHz. The larger the ratio of lower capacitor's value to that of the upper capacitor, the larger the ratio of amplifier current to the, and then try each combination out with SPICE until I saw something I liked.

I finally settled on the combination of mylar capacitors shown in the schematic - .092 uf directly across the antenna and .0165 uf and a 10 Ohm resistor in series with the driver. The .092 uf capacitance was made by placing two .047 uf capacitors in parallel and the .0165 uf capacitance was made by putting two .033 uf capacitors in series.

The results

 I checked the circuits frequency response with a tunable oscillator and confirmed that the peak in voltage across the coil is at 187 kHz Then I connected the output stage to a 187.5 kHz signal source (See "187 KHz RF Source" also on this site.). The voltage at the decoupling capacitor on the output stage was 9.01 VDC and the input current, as measured across the 10 ohm resistor in series with the power supply, was 82.6 milliamps, for an input power of 744 milliwatts. 

Top trace: Voltage out of amplifier -5 volts per division,
Lower trace: current out of amplifier - 200 ma per division.
Timebase: 1 us per division.
Note  the current spike that corresponds to the transitions of the output voltage.

The waveforms above show the signal coming out of the amplifier. Looking at the edge of the voltage pulse at the top, you can see that the transistor was having a hard time charging the capacitance of the matching network during the transitions. this is apparent by the low slew rate. The lower trace shows peak currents of about 325 milliamps peak during the transitions. The output signal, terms of both voltage and current, are rich in harmonics. The Q of the antenna circuit is about 30, giving it a 6 kHz 3db bandwidth. Look at the antenna signals below and see how well the network cleaned up the signals.

Top trace: Voltage across antenna inductance -10 volts per division,
Lower trace: through antenna inductance- 1 amp per division.
Timebase: 1 us per division.
Note that the voltage (upper trace) leads the current (lower trace) byt 90 degrees.
This is good because it shows that the coil's impedance is mostly inductive.
It also confirms that the currentand voltage probes were  in the right branch of the circuit.

The current in the antenna inductance (lower trace, above) was 2.5 amps and the voltage across the coil (upper trace, above) was about 19 volts. Ignoring the edges of the square wave driving signal, this makes the ratio of amplifier output current to antenna current approx = 2.5A/250 ma = 10:1.  This allows the transistors to safely stay in saturation while driving 1.25 amps peak into the antenna.

Overall, this system leaves a lot to be desired in terms of efficiency.

Just looking at the output stage and impedance transformation network for a moment, the real power delivered to the  antenna (the resistive part) is  that power that is actually delivered to the resistance - wire losses are about 0.4 ohms. The 10 nano-ohms radiation resistance (the transmitted power) will be ignored for the moment. Power is calculated using RMS current.

real power = (((2.5 amps peak-to-peak)/ 2.82)^2)/(0.4 ohms) = 314 mw

The formula above calculates the real power in the antenna. Here,
Power is the power dissipated in the antenna's resistance of about 0.4 ohms and
IPP is the peak-to-peak sine wave current in the antenna, which is about 2.4 amps.

The power into the output stage, not counting the power lost in the decoupling resistor in series with the power supply, is 9.01 volts x 82.6 milliamps = 744 milliwatts. This allowed me to calculate the approximate efficiency of the output stage and impedance transformation circuit. See the calculation below.

efficency = 100% (power out/power in) = 42%

Where is the other 430 milliwatts?

Take a look at the dissipation of the 10 ohm resistor in series with the output of the output stage.

Voltage across 10 ohm resistor that is in series with the output stage's output.
Vertical: 5 volts per division.
Timebase: 1 us per division.

This is a good application for a wide band true RMS voltmeter, but I don't have one, so I look at this as one 800 nanosecond per cycle sine wave that is 6 volts peak-to-peak and another 4 microsecond sine wave that is 4 volts peak-to-peak, that repeats every 5 microsecconds. The positive half cycle of the 800 ns signal is followed by the positive half cycle of the 4 microsecond signal, to you have to stare at the picture for a movement to get the idea. I can figure the RMS power dissipated by the resistor for each little sine wave and then add them together in proportion to their total time. Total power in the resistor is about 230 milliwatts.

From the input voltage measurement and output voltage wave form, the transistors are loosing about 1 volt out of 9, or about 11% of the input power in saturation losses. This would amount to 74.4 milliwatts.

A reasonable allocation of the input power would be as follows:

                Transistor saturation losses             11%        74  milliwatts
                Losses in 10 Ohm resistor              31%        230 milliwatts
                Power to the antenna resistance      42%        314 milliwatts
                Other losses                                    16%        119 milliwatts

The total power doesn't add up to 744 milliwatts because of rounding.

So, why is there so much power being sucked up by the 10 Ohm resistor and in the "Other losses" category? Its because only about 81% of the power in the square wave is at the fundamental frequency and the impedance transformation network, including the antenna inductance, is a bandpass filter. If the network were driven with a sine wave, more of the power would end up in the antenna, and indeed the 10 Ohm resistor would not be needed. Click here for  a brief explaination of where the 81% figure came from.

Compare this with the radiated power. If the antenna's radiation resistance is 10 nano-ohms, then the radiated power is 7.9 nanowatts, or 100% x 7..9 nanowatts / 744 milliwatts = 0.000001% overall efficiency.

The output stage's efficiency can be improved by increasing the power supply  voltage and also increasing the ratio of capacitances in the capacitive impedance transformation section. The net effect of these changes will be to reduce the collector currents and also reduce transistors' saturation voltages and to reduce the fraction of the total power supply voltage represented by the transistors' saturation voltages.  

I suspect that a large part of the "Other losses" is dissipation losses in the mylar capacitors. Changing to a high efficiency ceramic capacitor would likely help.

One thing this experiment did was reinforce the belief that its worth experimenting with other topologies to improve efficiency. Of course, the largest possible jump in overall efficiency would come from moving up to a full-sized 15 meter antenna. What topologies would be good ones to investigate? A class B linear amplifier driving a sine wave into the antenna would give about the same efficiency as seen here, but with more power being dissipated in the transistors. A class  D switching amplifier (basically pulse width modulation fed into a low pass filter) could give substantially better performance.

I believe the approach that is the subject of this page - driving a square wave into a filter is quite well suited to small, low-power short range applications, where low cost components, low voltage operation, and small size are important factors.

Epilog: A cleaner, kinder modulator


 The lower 2N2222, connected as a slew rate limited switch reduces key clicks.

I was uncomfortable with the keying method used in the first experiments -that of gating the drive to the NPN output transistor because I am sure that "keyclicks" extended beyond the 160 kHz to 190 kHz band. My objective is a beacon sending Morse code by cw keying, rather than tone modulated Morse code. Toward this end. Even though the duty cycle would be very low and probably would comply with the requirement that out-of-band emissions be 20 db below in-band emissions, I modified the output stage to eliminate these emissions. The modificaiton was to add  a second 2N2222 as a switch to modulate the power to the output stage. Notice that I reduced the 100 uf decoupling capacitor for the output stage to 1 uf to enable the transmitter to switch on and off quickly enough, and that the antenna returns to the negative end of the decoupling capacitor rather than to "ground".

The lower 2N2222 is connected as an integrator, which is a type of low-pass filter.. The closed loop gain is set by the input resistor and the feedback capacitor so that the output will be reduced 3db at frequencies above 1/(2Pi R C) = 338 Hz, with the output reduced by 6 db every time the frequency doubles. It doesn't take too many octaves to be 20 db down, as the FCC requires out-of-band emissions to be, so this can operate safely within about 2 kHz of the band edge.  

The saturation voltage of the 2N2222 was about 80 millivolts, substantially less than the 300 millivolts given in the data sheet. This puts the modulator's losses at 100 x 80 millivolts / 11.5 volts = 0.7% of the input power. I can live with that.

Remember, read that Part 15!  Read the Liability Disclaimer at the bottom of this page, too.


Contents © 2004 Richard Cappels All Rights Reserved. http://www.projects.cappels.org/
 First posted on the World Wide Web in April, 2004. (04/04/04). Updateds 6 April, 2004., May 2004, 16 June 2004.

You can send  email to me at projects(at symbol)cappels.org. Replace "(at symbol)" with "@" before mailing.

 Liability Disclaimer and intellectual property notice
(Summary: No warranties, use these pages at your own risk. You may use the information provided here for personal and educational purposes but you may not republish or use this information for any commercial purpose without explicit permission.)
I neither express nor imply any warranty for the quality, fitness for any particular purpose or  user, or freedom from patents or other restrictions on the rights of use of any software, firmware, hardware, design, service,information, or advice provided, mentioned,or made reference to in these pages. By utilizing or relying on software, firmware, hardware, design, service,information, or advice provided, mentioned, or made reference to in these pages, the user takes responsibility to assume all risk and associated with said activity and hold Richard Cappels harmless in the event of any loss or expense associated with said activity. The contents of this web site, unless otherwise noted, is copyrighted by Richard Cappels. Use of information presented on this site for personal, nonprofit educational and noncommercial use is encouraged, but unless explicitly stated with respect to particular material, the material itself may not be republished or used directly for commercial purposes. For the purposes of this notice, copying binary data resulting from program files, including assembly source code and object (hex) files into semiconductor memories for personal, nonprofit educational or other noncommercial use is not considered republishing. Entities desiring to use any material published in this pages for commercial purposes should contact the respective copyright holder(s).