AM Radio Circuit

[Iawia]

Hi Guys,

Just meddling with AM radio construction and got the resonator circuit to provide the frequencies of approx. 540 - 1610 kHz, my next task is to put it through the audio amplifier. The radio circuit attached is based on an AM Power amplifier circuit within the LM386 datasheet.

Got a few questions:

1. How to expand the database for NI's Multisim? I am trying to use component LM386 (audio amp), and it is not included in the student edition of the software. Do I go to Texas Instrument's website to obtain the file so that I can import this as a component in Multisim?

2. I did not know how to simulate the 'speaker', so I just made it a 8 ohm resistor. Should this be sufficient?

3. The LM386 manual places a 'ferrite bead' in parallel with the 47 ohm resistor. What is this purpose of the ferrite bead? I could not find ferrite bead in multisim so I used a 5 mH inductor. Is this ok to do?

Any help with be appreciated. Thanks for much in advance!

-t

encl LM386 datasheet
encl. my radio circuit

 

[JimB]

Any help with be appreciated.

1 Abandon your radio "design", it will not work*.
2 Do a "google" for "crystal set", to find the well established form of the circuit which has worked quite well for about 100 years.
3 Turn off the simulator and get out the soldering iron. You will not learn anything about radio receivers by playing a computer game.

* It will not work because:
The tuning inductor has far too high a value
The tuning capacitor has far too low a value
What on earth is the 1k resistor doing in there?
Your simulation of the tuned circuit is nothing like what will happen in the real world.
There is no direct current path for the rectified signal through the diode.

A ferrite bead on a wire will not have an inductance of 5mH, 5nH maybe.
You dont need the ferrite bead anyway. Just use the Zobel network on the output of the amplifier like the many other versions of that circuit except the one in that application note.

JimB

 

[Iawia]

Hi JimB,

Thanks for your reply. The crystal set certainly seems more viable and cheaper, and I will likely build this design. But before I abandon the old design, I would still like to know more details of why it will not work. I will likely redo such a dumb design in the future if I do not understand the 'wrongs'.

Using 1/{2*pi*sqrt{LC}} with the values that I have chosen in the resonator circuit picks up frequencies between 540 - 1610 kHz (roughly). These values I selected because it is what is available from digikey type websites. What range would work (the cap range I have selected is 2-17 pF) and inductor?

Will the amplifier portion of the circuit work? I mean without the ferrite bead and 47 ohm resistor? I put a 1k resistor in the RLC circuit bc I was not sure how much current would be passing through the LC components. From your response, pretty low I guess?

Once again, I am not going to physically build this circuit, I will likely just simulate it just for teaching purposes, and then build the crystal set.

-t

 

[Nigel Goodwin]

Fairly obviously, just check what value tuning capacitor a normal AM radio uses - this is required to give the tuning range needed, and a 2-17 pf one is no where near suitable, giving only a tiny tuning range.

I agree with JimB, get the soldering iron out and build it, computer 'games' aren't doing electronics.

 

[JimB]

Using 1/{2*pi*sqrt{LC}} with the values that I have chosen in the resonator circuit picks up frequencies between 540 - 1610 kHz

That is quite correct, and mathematically we can use any value of capacitance and the appropriate value of inductance between +/- infinity, to give any frequency we want.

In the real world there are practical limits.

If you examine the circuit of a medium wave radio, you will see that the tuning capacitor had a maximum value of 350 to 400pF.

Try calculating the the frequencies you get with a capacitance range of 50 to 400pF and an inductance of 220uH.

As for the amplifier, yes it will work ok, but you do not need that choke.
Do a search of this site for LM386 amplifiers, the circuit has been discussed many times.

JimB

 

[Willen]

* Duplicated post removed!

 

[Willen]

There is no direct current path for the rectified signal through the diode.
JimB

Hi JimB,

I found something interesting on your this sentence. I used to thought that after the rectifier the signal is still AC audio signal. Sorry for my basic poor curiosity. Is it a DC and won't pass through capacitor? I am being confusing.
I guessed that the rectifier without DC current path is like an common emitter single transistor amplifier with 'emitter disconnected' from Gnd, isn't it?

Once I made such radio and without knowing I used after the rectifier to ground- 1nF cap and 82k resistor. It is working till now. I think 82k is working as DC current path.

 

[Nigel Goodwin]

Hi JimB,

I found something interesting on your this sentence. I used to thought that after the rectifier the signal is still AC audio signal. Sorry for my basic poor curiosity.

The rectifier is for rectifying the carrier, NOT the modulation - this strips off one sideband (otherwise the two sidebands would cancel each other).

 

[audioguru]

The datasheet for the LM386 power amplifier IC shows an AM radio with the ferrite choke feeding the speaker, the speaker wires tightly twisted together and an RC lowpass filter at the input to avoid feedback causing oscillation at the radio frequencies because the LM386 has plenty of voltage gain at the radio frequencies.

The extremely simple and wrong circuit shown here is missing the coupling capacitor needed to feed the speaker so it is frying the amplifier, the speaker and the battery.

 

[Willen]

The rectifier is for rectifying the carrier, NOT the modulation - this strips off one sideband (otherwise the two sidebands would cancel each other).

oh I understood it strips off the sideband but how the diode strips off the carrier waves so we just hear audio, instead of carrier?

 

[ramuna]

Hi Willen,
As Jim said in post #2: "There is no direct current path for the rectified signal through the diode." What he means is that, for the the diode to conduct direct current, the diode has to be forward biased, ie the voltage at the anode of the diode must exceed the voltage at its cathode by at least the forward voltage drop of the diode (which is typically 0.6V for silicon signal diodes and 0.33V for germanium ones). In the radio schematic, there is no dc voltage return path to ground from the diode's cathode. Hence the diode can never be forward biased, and can never conduct dc. All that it can conduct in the configuration shown, is a tiny amount of AM RF ac, due to the diode's intrinsic capacitance, typically < 20pF in signal diodes. AM demodulation via diodes works like this: the am rf signal goes through a forward biased diode, which only passes that part of the waveform greater than the forward voltage drop. With the diode positioned as shown in the schematic, but having a dc return path to ground, eg a 10k resistor to ground from the diode's cathode (Ignore the schematic from this point on and instead concentrate on the text of this current post), we end up with a amplitude modulated train of purely positive voltage rectangular waves, which is the upper, positive half of the original AM waveform, less the forward voltage drop. This is then passed to the top connection of a resistor and capacitor in parallel with each other. The other ends of the resistor and capacitor connect to ground. The top connection is also connected, via a dc blocking capacitor, to the input of the audio amp. The resistor/capacitor combo act as an integrator, squeezing together the modulated rectangular waves into a positive dc voltage with no gaps & with the same modulation as the original AM signal. But because it is a varying dc voltage, it can pass through the blocking capacitor. This passage will convert the varying modulation voltage, which was always positive, into an ac voltage (imagine the dc voltage being pushed down through the x-axis, so that half of it is above and half below this axis). We have achieved demodulation of the am signal, the ac voltage at the input to the audio amp, is the audio signal which was modulated onto the carrier at the transmitter. Whew!!

 

[Willen]

Hi Willen,
As Jim said in post #2: "There is no direct current path for the rectified signal through the diode." What he means is that.......................................................... Whew!!

Wow! While reading your all text I was feeling like I am watching video tutorial in YouTube!

Again something is hidden from me- I know if there are two peak to peak cycle of modulated audio, inside of this two peak to peak cycle there are 100s of carrier peak to peak.

In the process you described, audio amplitude is rectified (selected positive wave above than 0.33V in germenium) and converted this varying DC into AC, which AC is audio. Now where carrier gone or muted? We never notice 500KHz carrier in our radio! I think our amplifier or speaker is not able to play it or our hearing is limited on 20KHz? Or really striped off the carrier (500Khz) during demodulation?

Even in some books I am not being able to see actually where 500KHz (lets say) gone! Thank you!

 

[Nigel Goodwin]

oh I understood it strips off the sideband but how the diode strips off the carrier waves so we just hear audio, instead of carrier?

The remaining part of the carrier is removed by the low-pass filter.

 

[ramuna]

Hi Willen,
You said:"In the process you described, audio amplitude is rectified (selected positive wave above than 0.33V in germenium) and converted this varying DC into AC"
Nearly right ! The radio-frequency (ie carrier) waveform is rectified, not the audio. The audio modulation can be regarded as a squishing in and a stretching out of the carrier, from above and below, ie along the y-axis. Now go back and read what I wrote in post #11 and you will see that the demodulation process I've described first rectifies the RF voltage into a modulated positive dc rectangular waveform (this is the rectification step), then integrates this into the dc audio waveform, then converts this into an ac audio waveform, then passes this to the audio amp, and finally from the audio amp output to the loudspeaker.

 

[JimB]

I feel a practical demonstration coming on!

So I lashed up the circuit shown in the schematic, fed it with 1 volt of RF at 1MHz, amplitude modulated 50% with a 1kHz tone.
I then ran through the circuit with the scope, taking pictures as I went, and the results are as you see here.

JimB

 

[audioguru]

Jim, why does your circuit have the 22nF/22k ohms highpass filter that cuts all adult male voice frequencies (330Hz cutoff)?
I would use a 220nF capacitor for a 33Hz cutoff.

Oh yeah. Bruno Mars is a soprano whose voice sings very high.

 

[JimB]

Jim, why does your circuit have the 22nF/22k ohms highpass filter

I could go into a very technical explanation about optimising... yada yada yada.

The plain simple truth is that apart from the lowpass filter at the output, it was simply lashed together using bits that were laying about on the bench as leftovers from something else which I was doing earlier this week.

If I get another fit of enthusiasm in the next day or so, I may plot the frequency response.

Thank you for your critique.

JimB

 

[JimB]

Well, it is a dull day here in the Aberdeenshire countryside, so I spent an hour or so plotting the frequency response of my hastily constructed demonstration diode detector.

The results are as predicted by AG.
With the original 22nF coupling capacitor, the low frequency response drops off by 3dB at around 300 -400 Hz.
Increasing the coupling capacitor to 220nF extends the frequency response down to the 30 - 40 Hz region.

The 0dB reference level is 1000Hz with the 22nF coupling capacitor.
I used this reference for the 220nF results in order to show the 1dB or so increase in output at 1000Hz with the larger capacitor.

JimB

I should also point out that the high frequency response shown here also includes a contribution from the modulation response of the signal generator.
A quick and dirty measurement shows that the generator response is about 5dB down at 100kHz.

 

[audioguru]

With the coupling capacitor value 10 times higher then I have my deep voice back, instead of sounding like a chipmunk.

In North America the AM radio stations are spaced 10kHz apart so radios cut the high audio frequencies to avoid a 10kHz beat tone. Then audio begins to be cut at about 2.5kHz or less.
AM radio stations here use pre-emphasis similar to FM radio stations with a boost at about 4khz to try to make their sound less muffled.