Maker Pro
Maker Pro

Analog Circuit to Randomly Modulate Amplitude and Frequency

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
I have a 20Hz 2Vpp sine wave and would like to simultaneously randomize both its amplitude and frequency (or phase) within a manually adjustable range of 10-50% deviation.

For example, the original signal would be made to randomly swing between 2Vpp and 0.2 to 1Vpp (adjustable), with a frequency of between 20Hz and 2 to 10Hz (adjustable). By "random" I mean in a non-repeating pattern.

The intended use is for a specialized visual effect.

I would prefer the simplest possible stand-alone circuit, constructed of commonly available analog and digital IC's. In other words, not a programmed device.

Can anyone please suggest what are the design options, and how best to proceed?
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
Amplitude could be modulated by a voltage controlled amplifier.
Frequency could be controlled by a voltage controlled oscillator.
The control voltage would be the random input signal e.g from another set of low frequency oscillators.
Both types of circuit are know from e.g.analog synthesizers. Look up schematic designs using your favorite search engine.

The circuit can become rather complex. A programmable solution where the "oscillators" and modulators are done in software would be less effort in hardware (one microcontroller plus an amplifier) - to be replaced by effort in programming.
 

Alec_t

Jul 7, 2015
3,587
Joined
Jul 7, 2015
Messages
3,587
What visual effect requires a sine-wave? It might be easier to come up with a solution if the wave-form isn't critical.
Be aware that light flashes/pulses in the ~ 5Hz-10Hz range can trigger epileptic seizures in susceptible people.
 

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
How about a random noise source fed into a quadrature mixer along with the sine wave? If so, could you provide any further information on that type of circuit.
 

Ylli

Jun 19, 2018
401
Joined
Jun 19, 2018
Messages
401
You are asking for output variations that an engineer usually works hard to avoid. An oscillator is normally designed so that the output frequency and amplitude are as insensitive as possible to the supply voltage.

So I guess if you want the frequency and amplitude to vary with supply voltage, then you could just use a very poorly designed oscillator. :)
 

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
Sinusoidal?
Just an observation based on the good advice above.
"I have a 20Hz 2Vpp sine wave"
Sometimes it is easier to make changes at the source. That is, create what you wanted rather than bend the sine wave you have already.
.
As also mentioned above, keeping a sine wave (if that has meaning here) may be demanding. Depending on how you get your visual effect, sometimes not doing this might be just as interesting.
.
Bandwidth?
I presume that you want it to vary slowly?
"random" means a lot of things. You mention ranges of amplitude and frequency. If I modulate to get this and the input to these modulators is smooth, slowly varying then the result can be viewed as a meandering "sine" wave that smoothly varies in amplitude and frequency/phase.
Another approach would be random step changes in amplitude and or frequency at regular or random times.
If instead I let the modulating inputs jump round their ranges rapidly (compared to 20 Hz) and chaotically, then the modulating circuit notionally stays the same, but the output will no longer really look sinusoidal. It will have significant modulation sidebands.
So, if you generate your modulation input from an analogue source or two, you'll need to limit its bandwidth: filter it.
.
A pair of noise sources, with some filteringing, as inputs to I & Q modulators sounds as if it will do it. It also sounds complicated. Probably also have to make a phase shifted version of your original sine wave. At a fixed frequency, that may be the easiest part.

As Harald Kapp said, software sounds as if it could be easier.
.
 

hevans1944

Hop - AC8NS
Jun 21, 2012
4,878
Joined
Jun 21, 2012
Messages
4,878
You haven't told us yet what you are going to DO with your randomized signals. Like, what "specialized visual effect" do you intend to create, and more specifically, how do you intend to create the effect? What hardware does the sine wave of randomized amplitude and frequency drive, control, or actuate in order to produce the "specialized visual effect" you seek to obtain? How much power will be required? What electronic design, do-it-yourself (DIY) construction experience, and/or electronics training do you have? Are you a drive-by poster seeking a quick answer, or are you willing to invest some time and effort in a back-and-forth dialog hopefully leading to a working, practical, solution?
 

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
Sinusoidal?
Just an observation based on the good advice above.
"I have a 20Hz 2Vpp sine wave"
Sometimes it is easier to make changes at the source. That is, create what you wanted rather than bend the sine wave you have already.
.
As also mentioned above, keeping a sine wave (if that has meaning here) may be demanding. Depending on how you get your visual effect, sometimes not doing this might be just as interesting.
.
Bandwidth?
I presume that you want it to vary slowly?
"random" means a lot of things. You mention ranges of amplitude and frequency. If I modulate to get this and the input to these modulators is smooth, slowly varying then the result can be viewed as a meandering "sine" wave that smoothly varies in amplitude and frequency/phase.
Another approach would be random step changes in amplitude and or frequency at regular or random times.
If instead I let the modulating inputs jump round their ranges rapidly (compared to 20 Hz) and chaotically, then the modulating circuit notionally stays the same, but the output will no longer really look sinusoidal. It will have significant modulation sidebands.
So, if you generate your modulation input from an analogue source or two, you'll need to limit its bandwidth: filter it.
.
A pair of noise sources, with some filteringing, as inputs to I & Q modulators sounds as if it will do it. It also sounds complicated. Probably also have to make a phase shifted version of your original sine wave. At a fixed frequency, that may be the easiest part.

As Harald Kapp said, software sounds as if it could be easier.
.

I was using a 20Hz sine wave as an example. I would like the circuit described to perform the same operation upon any signal fed into it. These would all be in the sub to low audio range.

The "meandering sine wave" you refer to above is exactly what I am looking for. Think sunlight through trees blowing in the wind. Random amplitude. Random "frequency" or rate of motion.

I understand about filtering, and also that it would be difficult to LPF white noise, as a random signal source, to such low frequencies. Perhaps I need a pseudo-random digital generator, CMOS or similar approach.

You mention dual noise sources, I & Q modulators and phase shifting the original signal.
Can you please describe in more detail how that type of circuit would work?

As for the purpose of desired visual effect, it is for laboratory experiments relating to plant growth that require non-deterministic lighting as would be found in a natural setting. A hardware solution is preferred for replication purposes.

Any additional help would be appreciated.
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
A hardware solution is preferred for replication purposes.
A software solution can be even more reproducible.
Random hardware effects are not reproducible whereas a pseudo random generator in software can easily recreate the same pattern over and over.
 

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
A software solution can be even more reproducible.
Random hardware effects are not reproducible whereas a pseudo random generator in software can easily recreate the same pattern over and over.

Software is frequently promoted as the "easy" answer. But I would need to write a program with variable parameters to accommodate a range of experimental conditions. Even then, there would be no ability to make gradual adjustments in real time while monitoring for effects. Hardware replication is accurate enough for my purpose.
 

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
Thanks for the clarification.
It appears that you want to modulate a light source. (incoherent, normal light source)
It appears that you want to have that modulation noise-like with mostly low frequency components.
I assume that your example of the 20Hz was an example modulation signal.
.
20Hz seems a little fast for the rest of your description. Sine seems an unlikely distribution for the resulting light amplitude, but I understand, just an example, for us.
.
If you have a way to modulate the light intensity from an analogue waveform, then it seems you only need to generate the noise-like waveform for the input. It seems you need an idea of the distribution of amplitudes required and the spectrum of the noise.
The suggestions above of filtered noise sources seem appropriate.
Not sure that I agree that either white(ish) noise sources are difficult nor filtering them is difficult. If you are aiming at a noise waveform with significant low frequency content/energy then the whiteness of the starting waveform may no matter that much.
If you want control to change the spectrum and effectively correlation over time, seems you want to be able to change the filter or effectively have a low frequency version of a graphic equaliser.
You might want to search for controllable/programmable noise sources. Might find an instrument that does what you want. It might be digital inside.
 

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
I quote Alec_t's warning about modulated light.

What visual effect requires a sine-wave? It might be easier to come up with a solution if the wave-form isn't critical.
Be aware that light flashes/pulses in the ~ 5Hz-10Hz range can trigger epileptic seizures in susceptible people.

Not software? Sure, ease relates to skillset. This forum has some very broadly capable people.

It may be that the easiest way to make the initial noise source, before filtering, is digital, such as a pseudo-random noise generator, even if implemented in digital gates rather than implemented in software.
I just found this one
https://www.digikey.co.nz/en/articl...ponents-generate-binary-sequences-white-noise
Perhaps it is of interest.
Filtering to the spectrum, correlation you want, can be analogue.
 
Last edited:

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
Thanks for the clarification.
It appears that you want to modulate a light source. (incoherent, normal light source)
It appears that you want to have that modulation noise-like with mostly low frequency components.
I assume that your example of the 20Hz was an example modulation signal.
.
20Hz seems a little fast for the rest of your description. Sine seems an unlikely distribution for the resulting light amplitude, but I understand, just an example, for us.
.
If you have a way to modulate the light intensity from an analogue waveform, then it seems you only need to generate the noise-like waveform for the input. It seems you need an idea of the distribution of amplitudes required and the spectrum of the noise.
The suggestions above of filtered noise sources seem appropriate.
Not sure that I agree that either white(ish) noise sources are difficult nor filtering them is difficult. If you are aiming at a noise waveform with significant low frequency content/energy then the whiteness of the starting waveform may no matter that much.
If you want control to change the spectrum and effectively correlation over time, seems you want to be able to change the filter or effectively have a low frequency version of a graphic equaliser.
You might want to search for controllable/programmable noise sources. Might find an instrument that does what you want. It might be digital inside.

The dominant "frequency" of the varying light needs to have a specifiable ELF component. As stated in my OP, all I want to do is randomise this in amplitude and frequency. For example, within a 10-50% deviation.

For this reason, I would start with an ELF sine wave. No rectilinear functions in nature. There is no problem performing AM or FM upon it. I can do that with an XR2206, even simultaneously. But making these modulate according to a slow-moving random waveform is the key.

To address the problem in stages, how would I generate in hardware a random waveform with exclusively ELF content? I have tried LP filtering white noise without success. The required bandwidth is relatively narrow requiring very high Q.
 

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
I quote Alec_t's warning about modulated light.



Not software? Sure, ease relates to skillset. This forum has some very broadly capable people.

It may be that the easiest way to make the initial noise source, before filtering, is digital, such as a pseudo-random noise generator, even if implemented in digital gates rather than implemented in software.
I just found this one
https://www.digikey.co.nz/en/articl...ponents-generate-binary-sequences-white-noise
Perhaps it is of interest.
Filtering to the spectrum, correlation you want, can be analogue.

The circuit in the article that uses a CD4015 and CD4030 looks very promising. Would it be possible to clock it slow enough so that the pulses occur exclusively within the ELF domain? I have used analog filters on PWM signals before in audio applications.
 

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
Ok, I just googled "ELF" and got 3Hz - 30Hz. Sure, you can slow clocks on digital circuits if that's what's needed. The clock scales the spectrum. Just scale the spectrum nto the range you want.
.
>>No rectilinear functions in nature.
No idea what you mean byu that one, sorry.
You get amplitude modulation from your example of the trees and leaves. The amplitude is one-sided, zero or above. Seems very unlikely that an amplitude distribution will be sinusoidal.
I guess from comms fading, if I mix I & Q inputs, each complex Gaussian, I get a Rayleigh amplitude distribution. As you have a direct path, I guess you get a Ricean amplitude distribution or something similar, which gets us back to making this with I & Q modulators, or just creating the noise-like waveform with correct-enough properties.
 

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
In dealing with natural phenomena, we very rarely see rectilinear wave forms.

Assuming the pseudo random generator works well at ELF, then an I Q modulator does seem like the most viable option. In that case, phase modulation would be acceptable in lieu of FM.

Can anyone recommend an IC that would be best suited for this purpose? I can these check the application notes.
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
Even then, there would be no ability to make gradual adjustments in real time while monitoring for effects.
I understand that you are reluctant to use a software solution. But these reasons given are invalid. You could use potentiometers and a built-in ADC to change parameters on the fly.
 

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
Yes, but I would have to outsource the programming for anything like that. I would like to proceed more along the lines described by Nanren888. Might be of interest to some other analog fans here as well.
 

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
How about a random noise source fed into a quadrature mixer along with the sine wave? If so, could you provide any further information on that type of circuit.

Just for the record, I think you mentioned the quadrature mixer option.
It's more the starting with a sine wave that bothers me.
Google "quadrature mixer"?
 

epsolutions

Sep 7, 2019
116
Joined
Sep 7, 2019
Messages
116
Not sure why that would be bothering. The mixer is intended to "distort" the sine wave. It would still look sinusoidal in part, but with amplitude and phase changes as determined by the modulating random wave form.

Can anyone recommend a mixer IC part that would be most suitable for this purpose?
 
Top