Amplitude Modulation Using Op Amp Inputs

[epsolutions]

The circuit below, once completed, is intended to modulate amplified noise from a 470uF cap with a CMOS generated ELF sine wave.

As you can see, I am looking for a place to feed the ELF into the TLC272 stage to perform the AM function. Is this possible? If not, what is the simplest way of achieving the AM?

 

[Harald Kapp]

We do have an old discussion on AM with op amps here. You'll have to throw in a FET to modulate the signal. The opamp per se cannot perform the required multiplication.

Alternatively: AM is effectively multiplying the carrier signal by the modulating signal. Multiplication can be replaced by addition if you first apply a log() function to both signals, then add the log'd signals and apply exp() to the sum:
a * b = exp(log(a)+log(b))

 

[epsolutions]

I did not see a resolution or working schematic in that discussion.

Is the solution as simple as feeding the ELF through a JFET into pin 3 of the TLC272 with no other added parts? I would try now, but do not have a JFET on hand.

But my intent is to modulate the ELF with the capacitive noise. If I feed into pin 3 will that not cause the reverse, i.e. the noise modulating the ELF? See diagram below. Do I need an alternative circuit topology to achieve this?

 

[Harald Kapp]

In the schematic "carrier" would be your ELF, "V_i" would be the noise.

What exactly are you trying to achieve= Usually the carrier frequency is much higher than the modulating signal. In your scheme it is the other way, carrier being ELF and noise having a wide band spectrum. You can't, as far as I know, transmit a higher frequency signal faithfully by a lower frequency carrier.

Another point: a capacitor is not a good source of noise. At least to my knowledge. Use a resistor instead.

 

[bertus]

Hello,

A transistor or zener will be a better source for noise:
https://kb8ojh.net/projects/noisegen/

Bertus

 

[epsolutions]

Yes. But junction noise is not what I want. Anyway, that part of the circuit is fine.

To use the ELF as the carrier, it seems I will need an additional op amp, with a JFET, after the noise stage to function as the modulator. The noise will be low pass filtered (not shown on my original diagram) because the desired result is two inter-modulating ELF's one being fixed and the other random. Hence, which is the "carrier" may not matter in practice. I will get a JFET and try it both ways.

I also found this configuration using diodes.

 

[epsolutions]

I refer to the diagrams in my first and second posts above. The FET I ordered finally arrived by post. It is a Fairchild 2N3819. The pinout is given as D G S when viewed from the front.

As shown in second post diagram, I fed the ELF sine wave 2Vpp into the gate and connected the drain to pin 3 of the TLC272 in first post diagram. The source was connected to earth. This destroyed the op amp.

I imagine there needs to be resistors added but I am unfamiliar with this type of design. The objective is to modulate the TLC with the ELF. Can anyone please point me in the right direction?

 

[Harald Kapp]

The pinout is given as D G S when viewed from the front.

Depends on what you decide is the front. For me the front would be the flat side with the labeling on it. With the pins facing down (so you can read the print), the correct sequence is S G D (as per datasheet):

But reversing D and S would not destroy the opamp.

I imagine there needs to be resistors added

I wouldn't know where and why. You can short the opamp's input pin to gnd and this should not destroy the opamp. I guess something else happened.

The real issue is that you selected an unsuitable FET. Your choice is a depletion JFET, not a MOSFET. A JFET (Junction FET) has no isolating oxide layer between gate and channel. Isolation is provided by a pc-junction which has to be reverse biased (negative V_GS). When you applied a positive gate voltage, the pn-junction was forward biased and the modulating signal (f_carrier in post #3) was directly connected to the input of the opamp via the gate-drain diode of the JFET. This may have caused the defect in the opamp if the input voltage from the carrier signal was too high.
Apart from that the 2N3819 is designe fore VHF/UHF applications and not well suited for use with comparatively low frequency signals.

Your image in post #3 shows an enhancement MOSFET (2n7002) which has an isolated gate and requires a positive V_GS to be turned on. It is designed for switching applications. Controlling it with a sine on the gate will not produce the expected result: the threshold voltage is somewhere between + 1 V and + 2.5 V (note the wide range). Simply stated: below that the transistor is off, above that the transistor is on. This is why the image in post #3 shows a square wave as the carrier signal.


You can use the 2N3819 as a voltage controlled resistor with a sinusoidal input, see e.g. this app note.
Or, if you can accept a square wave carrier, use the 2N7002 as per the original schematic.

 

[epsolutions]

Yes, you are correct about the 2N3819 pinout.

I misunderstood from an earlier reply that a JFET was the way to go. At the time, I also ordered an AD633 four quad mixer IC. They are not cheap, but I am now thinking that is the least challenging option for the type of output I am seeking.

Here is an LTSpice sim of a similar circuit someone was kind enough to have sent me previously.