Forums :: Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives :: Books
Support This Site: Contributors :: Advertise


It is currently Dec Sun 16, 2018 2:47 am


All times are UTC [ DST ]





Post New Topic Post Reply  [ 6 posts ] 
Author Message
 Post subject: Another Variation of the LM386 Based AM Transmitter
PostPosted: May Fri 25, 2018 9:12 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1336
Location: Burke, VA
In the original 35Z5 circuit, the output of the LM386 audio amplifier is used to modulate the power supply of an RF crystal oscillator. My aim here is to insert a switch-mode modulator between the LM386 and oscillator so that the oscillator operates with a fixed power supply. This also allows the use of a programmable oscillator which does not tolerate a wide power supply voltage variation.

The LM386 output consists of an alternating signal riding on a DC offset equal to half the supply voltage. If this output is chopped by a high-frequency square wave, we get a waveform that resembles an AM signal but not quite. It looks more like a half-wave rectified AM signal where the negative part is suppressed and set to zero.

Attachment:
waveform1.PNG
waveform1.PNG [ 22.13 KiB | Viewed 623 times ]


To recover the negative part of the chopped signal and create our AM signal, we need to recover its envelope by averaging and subtract it from the signal. This operation is equivalent more of less to high-pass filtering with a cutoff frequency set many times higher than the largest audio frequency to avoid distortion.

Attachment:
waveform2.PNG
waveform2.PNG [ 19.81 KiB | Viewed 623 times ]


As shown is the block diagram, the chopper is implemented as two switches driven by an RF square wave. When the square wave is high, the upper switch is on and the lower switch is off, passing the LM386 output to the midpoint of the switches. When the square wave is low, the upper switch is off and the lower switch is on, grounding the midpoint or the chopper output. The average block recovers the audio envelope with an amplitude exactly one half of the actual envelope. The recovered envelope is subtracted from the chopper output to create the AM signal. Since this AM signal is rich in odd harmonics, a high-Q lowpass filter follows with the cutoff frequency matching the square wave frequency. The lowpass filter also serves as an antenna matching network.

Attachment:
AM_Block_Diagram.PNG
AM_Block_Diagram.PNG [ 17.05 KiB | Viewed 623 times ]


In the partial schematic diagram, the chopper switches are implemented as a pair of NPN and PNP transistors (Q1 and Q2). The values of the speed-up capacitors C6 and C7 are chosen to counteract the storage time of the transistors just high enough for the chopped waveform to have 50% duty cycle. C8 is used to reduce the overshoot when Q1 turns off. The audio envelope is recovered by a low pass filter formed by C9 and the primary inductance of T2. This recovered envelope is subtracted from the chopped waveform to create our AM signal across the primary of T2. The output low pass filter and antenna matching network are formed by L2, C17 and C18. Since the bases of the transistors are not ground referenced, a pulse transformer T1 along with C10 are used to couple them to the an RF square wave oscillator. L1, C11 and R5 form an EMI filter and are not required for proper operation. R4 is used for hooking up a voltmeter during tuning. The readout in mV corresponds to the mA current at the LM386 output.

Attachment:
HBAM_schematic_part.PNG
HBAM_schematic_part.PNG [ 45.63 KiB | Viewed 623 times ]


The circuit was prototyped and built with a homemade PCB to fit a BUD plastic box. The telescopic antenna has a fully extended length of 115 cm (45"). The entire unit is mounted on a metal leg of an old torchiere floor lamp which serves as the ground part for the antenna.

Attachment:
HBAM_top_800x450.jpg
HBAM_top_800x450.jpg [ 40.55 KiB | Viewed 623 times ]

Attachment:
HBAM_bottom_800x410.jpg
HBAM_bottom_800x410.jpg [ 66.69 KiB | Viewed 623 times ]

Attachment:
20180508_045916_790x370.jpg
20180508_045916_790x370.jpg [ 39.14 KiB | Viewed 623 times ]


Some waveform captures to verify the operation of the unit.

Attachment:
Unfiltered_and_L_Match_Outputs_800x370.jpg
Unfiltered_and_L_Match_Outputs_800x370.jpg [ 40.21 KiB | Viewed 623 times ]

Attachment:
Sine_and_Sawtooth_Modulation_200Hz_50%_50V_per_div_800x370.jpg
Sine_and_Sawtooth_Modulation_200Hz_50%_50V_per_div_800x370.jpg [ 37.38 KiB | Viewed 623 times ]

Attachment:
Sine_and_Sawtooth_Modulation_1kHz_50%_50V_per_div_800x370.jpg
Sine_and_Sawtooth_Modulation_1kHz_50%_50V_per_div_800x370.jpg [ 35.39 KiB | Viewed 623 times ]

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Another Variation of the LM386 Based AM Transmitter
PostPosted: May Fri 25, 2018 9:16 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1336
Location: Burke, VA
Attachment:
20kHz_and_40kHz_Sine_Modulation_50V_per_div_800x370.jpg
20kHz_and_40kHz_Sine_Modulation_50V_per_div_800x370.jpg [ 40.81 KiB | Viewed 622 times ]

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Another Variation of the LM386 Based AM Transmitter
PostPosted: May Sun 27, 2018 8:29 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 3660
Novel idea well executed. How does it sound?


Top
 Profile  
 
 Post subject: Re: Another Variation of the LM386 Based AM Transmitter
PostPosted: May Mon 28, 2018 1:29 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1336
Location: Burke, VA
Macrohenry - Tom - thank you for the kind words. This thing sounds pretty good although it has some distortion issue at very high modulation level.

The idea used behind this implementation may look novel at first but it is really not. This modulator belongs a class of voltage-fed switch-mode converters. Examples of this class are the buck step-down switching voltage regulator and some electronic ballasts for CFL bulbs. The other class is the current-fed converter. Some examples of this second class are the boost switching regulator and the resonant Royer converter. A good place to gain understanding of these two classes is the 1959 conference paper by P.J. Baxandall, "Transistor Sine-Wave LC Oscillators".

There are plenty current-fed modulator implementations based on Sokal class-E tuned single-ended switching power amplifier. I could not find any voltage-fed switching modulator implementation besides the voltage-switching oscillator described in the Baxandall paper. Out of curiosity, I took Baxandall's oscillator and turned it into a voltage-switching modulator to see how well it performs.

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Another Variation of the LM386 Based AM Transmitter
PostPosted: May Mon 28, 2018 4:13 pm 
Member

Joined: Apr Thu 12, 2007 3:24 am
Posts: 1721
Location: Milwaukee,WI
What cores are you using for the coils that you can get such high values with so few turns?


Top
 Profile  
 
 Post subject: Re: Another Variation of the LM386 Based AM Transmitter
PostPosted: May Mon 28, 2018 6:04 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1336
Location: Burke, VA
The small ferrite toroid used for the base drive transformer was salvaged from a CFL bulb. It has 10 turns of bifilar wire giving an inductance of 242uH. Its material is probably equivalent to Amidon type F.

The bigger ferrite toroid for the output transformer was used as a bead for a computer power supply. It has 15 turns of bifilar wire giving an inductance of 135uH. Its material is probably equivalent to Amidon type 77.

The 220uH loading inductor (L2) for the antenna was salvaged from a printer switching power supply. When driven by a 2V square wave, it resonates with the antenna capacitance (plus C17 and C18) with a peak amplitude of more than 80V. So its Q factor is at least 40.

_________________
Binh


Top
 Profile  
 
Post New Topic Post Reply  [ 6 posts ] 

All times are UTC [ DST ]


Who is online

Users browsing this forum: Ed Engelken and 5 guests



Search for:
Jump to:  
























Privacy Policy :: Powered by phpBB