Dick Cappels' project pages http://www.projects.cappels.org
return to HOME
A Superhet/Direct Conversion AM receiver
for 181.818 kHz A simple portable receiver with the antenna being the only LC circuit.Photo of completed receiver. Its pocket sized, but not intended to be used that way because the antenna is a highly directional ferrite loopstick. Its only a matter of luck that there was room for all the components on the board, which is wired point-to-point without a ground plane. A 100 uf power supply decoupling capacitor was added just above the quad op-amp package after this picture was taken.
This is a description of an experimental AM receiver for VLF. It is crystal controlled to receive 181.818 kHz (more or less) and operates as either a single conversion superhetrodyne or a direct conversion receiver. The bandwidths are expected to be 15 Hz to 3.5 kHz in superhet mode and 15 Hz to 10 kHz in the direct conversion mode. Operation is from a single +5 volt regulator and it can be powered from a 9 volt radio battery or other source of +8V or higher. The reason I built this was to verify some of my assumptions about these components would work together, having had numerous discussions of the SA-612/MK-484 and SA-612 direct conversion receiver concepts over the last year or so with my friend Jeff, besides that, I want to have a direct conversion receiver on hand for some planned experiments.
So, why 181.818 kHz? Its near the middle of a band of frequencies in which some countries, including the United States, allows unlicensed operation. For the U.S., see Code of Federal Regulations, Title 47, Volume 1, Section 15.217 Operation in the band 160 - 190 kHz. U.S. Government Printing Office, or look for it on the web. This particular frequency is easily derived from some common microcontroller crystal crystals, in this case a 6 Mhz crystal. Of course with the current generation of on-chip clocks, starting with the ATtiny series, crystals aren't really necessary for some applications. Other projects using this particular frequency are in the works.
Schematic. Depending up the state of the mode switch on the left, either pin 8 or pin 9 of the AT90S2313 goes high. When pin 8 goes high, it supplies power to the MK-484 (ZN-414 replacement - an integrated IF/AM detector). When pin 9 goes high, it stops the 2N2907 from shunting the direct conversion signal at the output of the SA6-12.
In the superhetrodyne mode, the receiver consists of a ferrite loopstick antenna, a local oscillator, mixer an IF filter and an IF amplifier and AM detector, followed by an audio amplifier.
ConstructionThe loop antenna is a resonant circuit made of a 760 microhenry coil wound on a 5.5 cm long ferrite rod covered with one thickness of 80 grams per square meter printer paper, in parallel with a 1000 pf capacitor.. The ferrite rod was picked up at a surplus store, so I had to experiment a little to find the correct number of turns. I started by winding 100 turn closely spaced in the middle of the ferrite rod, with a tap at 50 turns. I used #30 heavy polythermaleze magnet wire. I measured the inductance between the 50 turn tap and each end of the coil, and measured it a third time for the entire 100 turns, then I calculated an inductance factor for the rod for each of the windings, and then found the average inductance factor. The inductance factor, also referred to as AL (A-sub-L) in some literature is the factor that, when multiplied by the number of turns squared, is equal to the inductance. Here, I will use the symbol K.
Photo of Loopstick. The winding is near the center of the rod. A thickness of paper reduces parasitic capacitance and cushions the wire against abrasion of its enamel coating. The two inner-most plastic wire wraps are used to space the coil slightly off the board while the two on the end secure the loopstick to the board.
L = K N^2 (inductance equals K times N squared). I'll leave the algebra to you, but the average K for this ferrite rod came out to 9.405E-8 Henries per turn squared. To resonate at 181.818 kHz with a 100 pf capacitor, the inductance would have to be (1/(2 * Pi * 181E3)^2)/1000E-12 = 766.2 microhenries. The number of turns, N = Sqrt(L/K) = Sqrt (766E-6/94.E-8) = 90 turns. I wound 90 turns and measured the inductance to be 781 microhenries. I took one turn off, and measured again, and the inductance came out to 763 microhenries, and so used that.
181.818 kHz Magnetic fields circulating through the ferrite core induce current in the winding and that generates a voltage across the 1000 pf capacitor. Since the inductance is resonant at 181.818 kHz with the parallel capacitance, the voltage across the resonant circuit is larger at 181.8 kHz that at other frequencies, thus the antenna serves to convert time-varying magnetic fields at the desired frequency to voltage while filtering out other frequencies. The higher the inductive reactance of the inductor to the resistance in series with it and the capacitor, the higher the "Q", or quality factor, and the higher the voltage generated across the resonant circuit. A large part of the total resistance is skin effect and eddy current effects, both of which could be minimized by winding several parallel conductors of a smaller diameter, but that is a refinement we can leave until later.
The voltage across the resonant circuit is connected to input pins 1 and 2 on the SA-612 mixer. The incoming signal is mixed with the output of the local oscillator, which is an AT90S2313 microcontroller. The microcontroller sits in a loop and generates one square wave cycle on its output for every 22 clock cycles. Its clock frequency is set at 6 MHz by the 6 MHz crystal connected to its on-chip clock oscillator (AT90S2313 pins 4 and 5). When operating in the direct conversion mode, the AT90S2313 produces 181.818 kHz. The 272.727 kHz output pulses on pin 11 of the AT90S2313 are 5V peak-to-peak and are divided down to 240 mv P-P, the level necessary to drive the local oscillator input of the SA-612, by the 20k and 1 k resistive divider.
The 330 pf capacitor across the output of the voltage divider combines with the approximately 1K output resistance of the divider to make a single pole low pass filter at about 500 kHz. The intent here was to knock the corners off the square wave to reduce the amplitude of the higher level harmonics of the local oscillator, so as to reduce the receiver's sensitivity to frequencies related to those harmonics.
The SA-612 provides some gain to the input signal and the 181.818 kHz input mixes with the 272.727 kHz local oscillator to create an image of the input signal at 454.45 kHz. Close enough to the ceramic filter's 455 kHz center frequency.
The output of the SA-612 is 1.5 k Ohms, which, based on some web surfing, I suspected to also be the termination impedance of the surplus store ceramic filter, so pin 4 of the SA-612 is connected directly to the filter's input. The output of the filter is terminated in 1.5 k Ohms and connected to the input of the MK-484.
The MK-848 is a "remake" of the Ferranti Semiconductor ZN-414 TRF receiver chip. It serves as an intermediate frequency amplifier and AM detector. The connection of the MK-484 is straight from the manufacturer's data sheet, except I substituted a 1.5 k resistor for a loopstick and resonating capacitor. The 1k output resistor and .047 uf capacitor to ground roll off the high frequency response of the MK-484 at 3.4 kHz. The power for the MK-484 is supplied through the 1 k resistor on its output. Per the manufacturer's specifications, the maximum voltage at this point is 1.8 volts, so the voltage is regulated by the forward drop of a light emitting diode. This diode has a 120 Ohm internal resistance, so it might have a slightly higher drop than some others. The current to this shunt regulator is supplied through the 3.3 k Ohm resistor, and it is only supplied when pin 8 of the AT90S2313 is high, which corresponds to operation in the superhetrodyne mode. When in the direct conversion mode, pin 8 of the AT90S2313 is low, so there is no output for the MK-484.
The output of the MK-484 is baseband (audio) and it is coupled to the input of an op-amp connected with a voltage gain of 100X (40 db), with a low frequency cutoff at 15 Hz, and the upper frequency limit being determined by the op-amp's gain vs frequency plot. The op-amp following this drives the headphone connector and it is set up to have an adjustable gain, according to the setting of the 50 k Ohm volume control, from X0 to X50. When the wiper of the pot shorts the inverting input to the op-amp's output, there is virtually no gain from the amplifier. When the wiper is at the end of the pot connected to the 1 k Ohm input resistor, the gain is approximately 50X.
As might be expected, with a total possible gain of 5,000 (68 db) with an LM324, there is a lot of hissing and popping. I also tried a TL034 J-Fet op-amp with the same pinouts, and found, unsurprisingly, that a lot of the lower frequency noise went away, especially the "popcorn" type noise. There are other quad op-amps out there, some with even better noise figures. I guess the other key parameters, besides noise performance, to keep an eye on when considering other op-amps is power supply requirements, dynamic range of inputs and the output, and stability with unity gain (as in the case of the audio output stage with minimum gain).
The low pass filter made of the 20k resistor and the 0.47 uf capacitor on the pin 10 of the LM324 serves to couple the DC bias established at the first Op-Amp, and it also serves to roll off the amplifier's gain at frequencies below 15 Hz.
The signal coupling into the 1k resistor, then to the inverting input is coupled through a 10 uf capacitor. At the mamimum gain setting, this network's low ferquency corner is 15 Hz, but decends to lower frequencies as the gain is increased.
When in the superhetrodyne mode, the current signal from pin 5 of the SA-612 is shunted to the +5 volt power supply by the 2N2907, which is held on by pin 9 of the AT90S2313 being low. When in the direct conversion mode, pin 9 of the AT90S2313 goes high, turning off the 2N2907, allowing the signal current to create a voltage drop across the SA-612's internal 1.5 k Ohm resistance.
When in the direct conversion mode, as set by the Superhet/DC switch, the AT90S2313 outputs one square wave cycle for every 33 of its 6 Mhz clock cycles. The square wave is slightly asymmetric, but that should have little effect on the efficiency of the detector. In this case, when the 1818.818 kHz local oscillator mixes with the 1818.818 kHz signal and/or its sideband(s), the result on pin 5 of the SA-612 is a baseband (audio) signal. This signal is couple do an auido amplifier with a gain of 20, and this signal is then applied to another input of the X100 amplifier, and then to the audio output amplifier with the 50 k Ohm volume control.
Power is supplied from the 7805 voltage regulator. There is minimal decoupling capacitance on the board. The 7805 that I use, and I have used dozens of this particular SGS regulator, is stable without an input capacitor across it, and does not seem to require one on the output either. The 100 uf across the +5V output was added to reduce cross-coupling through the power supply of signals between stages.
Note the way the signals from the output of the MK484 and the direct conversion signal from op-amp pin 1. Op-amp pin 1 is a low impedance, and thus a virtual ground for the voltage divider that sets the gain for the amplifier following the output of the MK-484. The gain of that stage (op-amp pins 5,6, and 7) is eaual to (100k + 1K)/1k = X101. The 10 uf capacitor, in conjunction with the 1k resistor, rolls of the low frequncy gain t 15 Hz. When the output of the direct conversion amplifier (op-amp poins 1,2, and 3) drives the negative end of the 10 uf capacitor, the signal gain from that point through the upper op-amp (pins 5,6, and 7) is equal to 100k/1k = 100, with the same 10 uf capacitor causing a 15 Hz lower corner frequency. Thus, the two signals are summed together.
To switch between modes of operation, one detector or the other is disabled (see text above) and the local oscillator frequency is changed.
Photo of circuit side of board. Its ugly, but works just fine at these frequencies. Just make sure the solder flux is cleaned away from the high impedance areas.
Thank goodness, I bought a bunch of those plastic pencil boxes before they went out of style, and its even better that the ferrite rod fit inside. I cut a piece of hole-per-pad prototyping board to fit the case and made a cut-out around where the 9 volt battery would go. The ferrite rod is held in place with plastic wire ties. All of the connections are ponit-to-point with no ground plane. In most cases, I was able to place components in position so I could just bend their leads over to adjacent components and solder them down and clip them off. A few kynar insulated wires are used were necessary. The most difficult part of the entire project was filing holes for the switches.
Upon power-on reset, the process goes through its houskeeping tasks, mainly setting up the I/O data and data direction registers. It then enters one of two similar looking timing loops, one of which is shown in the listing below. One of the loops outputs a complete square wave on the local oscillator output pin every 33 machine cycles, and the other loop outputs a complete square wave on the local oscillator output pin every 22 machine cycles. Control passes from one loop to another, depending upon the state of the mode control pin (AT90S2313 pin 6), which determines whether the receiver is working as a superhet or direct conversion receiver.
In pseudo code, here is what happens:
set local oscillator output pin high
set output switching pins high and low to agree with current mode (superher or direct conversion)
delay for the rest of half a cycle
set local oscillator output pin low
jump to top of this loop or to other local oscillator loop depending on the state of the mode control pin
sbi RFSigPort,RFSigPin ;Output High sbi and cbi are 2 clock instructions.
cbi Swout0PORT,Swout1Pin ;Output switch bit 1 low
sbi Swout1PORT,Swout0Pin ;Output switch bit 0 high
sbi RFSigPort,RFSigPin ;Spend 11 cycles each in high and low.
cbi RFSigPort,RFSigPin ;Output Low
cbi RFSigPort,RFSigPin ;Spend either 11 clocks sending low.
rjmp Loop33 ;Do this forever.
Listing. This routine sets the local oscillator output pin high with the sbi command. Subsequent sbi commands before repetition of the loop are merely used as two clock cycle delays. Similarly the cbi command is used to set the local oscillator pin low and subsequent use of the cbi command before repetition of the loop is merely to serve as two clock cycle delays. The loop above divides the clock by 22. A similar loop divides the clock by 33.
This didn't have to be done with an AT90S2313, any AVR that can run from 6 MHz crystal would work. I suspect a good many other controllers such as PICs would also work well for this task, but I used an AT90S2313 because that's what I had on hand. Note -there are programming services that can program a chip for you.
I did some quick listening tests using an on-off keyed data signal source at 181.818 kHz that uses a resonat loop antenna. Subjectively, in superhet mode, the receiver seems to give a little worse signal-to-noise ratio than my RadioShack DX399 receiver when using the TL034, which is a lower noise op-amp than the LM324. I suspect that a little additional RF gain with a low noise preamp would be helpful. In direct conversion mode, it seems to be as "sensitive" as the trusty old RadioShack, but then its operating in a different mode.
HOME (more projects)
Contents ©2004, 2005, 2008 Richard Cappels All Rights Reserved. http://www.projects.cappels.org/
First posted in November, 2004. Revised November 13, 2005.
You can send email to me at projects(at)cappels.org. Replace "(at)" with "@" before mailing.
Use of information presented on this page is for personal, nonprofit educational and noncommercial use only. This material (including object files) is copyrighted by Richard Cappels and may not be republished or used directly for commercial purposes. For commercial license, click here.
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).