the tremolo effect: origins, mechanisms and improvements!

In this article we will explore the universe of the tremolo. Presenting the origin of this effect, the difference between a harmonic tremolo and a tremolo. How they work and finally, how to improve this effect which is more than half a century old!
In short, a maximum of content to master the subject, hang on!

The harmonic tremolo is back

The signal’s gone to the chopper!

From the lonely desert plains of spaghetti western to the stuttering chopped-up notes of the most epileptic EDM, tremolo is everywhere.
Tremolo was the first effect designed specifically for electric guitar, even before reverb, when the DeArmond pedal was released in 1948.

Tremolo was then built-in many vintage amps (with wildly different circuits depending on the brand), before it was defeated by phaser and chorus in the 70s and 80s.
Eventually, it came back with a vengeance by the end of the nineties, at which point it became a precious tool for guitarists trying to sound rootsy and authentic.
Since then, every builder has introduced their version with many variations in how much you can actually do with them. 

Split it up and pan!

At the opposite, harmonic tremolo is one of the rarest and most interesting effects, allowing you to make your sound highly personal.
It brings a much livelier motion than a tremolo, it doesn’t have the outdated stigma of the chorus, and it takes up less space than a delay or reverb.

Harmonic tremolo lives under many names: vibrato, harmonic vibrato, harmonic tremolo, vibe, univibe, and so on.
In every instance, the goal remains the same, i.e. to change the pitch of your notes in a cyclic way (rather than changing the volume like a regular tremolo) while putting the lows and highs out of phase.
That separation between frequencies evokes the Leslie cabinet, and it’s not just by coincidence:
The first harmonic tremolos, including the Univibe and the built-in vibratos from old Magnatone amps, have been designed to emulate the Leslie.

But those newer references have of course become sought-after effects in their own right. And a few makers have designed their own version since then.

The tremolo effect

How does it work in theory?

The radio!

It all began with the invention of the radio and, more specifically, AM transmission during the First World War.

Our dear engineers/researchers were looking for a way to send an audio signal into the air. So a signal with frequencies between 20Hz and 20kHz. In the world of signal processing it is a very low frequency signal, we regularly work in the field of MHz in radio!

To broadcast this signal in the air, the ideal antenna size is calculated by this formula:
Antenna length = c / min. frequency.
Length = 3*10^8 / 20 = 1.5*10^7 m
But not even the Eiffel Tower could do it!

So our physicists didn’t let it happen and took the problem in reverse, they decided to send high frequencies in the air, on the MHz scale (yes yes that’s what you select on your radio when you change stations) and then they managed to insert our voice in it.

And yes, if we go back to the formula with a 433MHz carrier, we only need an antenna of about 69cm to be in optimal conditions!

Amplitude modulation

The creators of this method called it, amplitude modulation or AM. FM stands for frequency modulation rather than amplitude modulation. It makes the signal much more robust against interference, but they invented it a bit later!
The audio becomes the modulated signal and the high frequency the carrier.

Once your radio picks up the high-frequency signal, it applies a simple low-pass filter to recover the audio and remove the high frequency. (Disclaimer: I didn’t look at the schematic but I guess it can’t be much more complicated than that on the early analog models and on the first processing stages).

Consequently, this operation inspired the first tremolo designers.
Their purpose was to vary the volume of an audio signal very slowly and cyclically.

The LFO

The first tremolo creators therefore kept the principle of amplitude modulation (AM) but the modulating and modulated signal are now inverted.
And yes! The modulators in this case are very low frequency, usually from 0.1Hz to 20Hz. If it was bigger, we may hear them, they would create pitches like what you get with a synthesizer. And that’s not the purpose of a tremolo!

Then the modulated signal is simply the incoming guitar (or other instrument) signal.

This very low frequency is called a LFO for Low Frequency Oscillator. It exists in all shapes, here I used a sinusoid. But a triangle, a square or anything you can imagine work too!

That’s why you often find a whole bunch of “waveforms” when you’re looking for a new tremolo. Because unfortunately it’s one of the only things that manufacturers have had the chance to play on in the last few decades.

So it’s time to bring a little more originality to this more than exciting effect.

Optical tremolo main schematic

The way a tremolo works is no longer a mystery for you. So I’m going to present you a “simple” schematic that creates this beautiful analog amplitude modulation.

The optical tremolo uses a vactrol

In this schematic, the signal enters by the left. Then it meets from the beginning a rather mysterious component, the famous vactrol! It is an optical component (hence the name of optical tremolo that we often find). On one side on its legs 1 and 2, we have a LED. And, on the other side on 3 and 4 we have a photoresistor. The 2 components in this version are encapsulated in the same housing. But there is a DIY solution where you buy a simple LED and a simple photoresistor. It is much cheaper but the performances are not the same level!

To come back to the vactrol, a photoresistance is really magical for our assembly. If the LED is off, the resistance will be very important, around MOhms. Then, when the LED is ON, the brighter it is, the lower the resistance will be. We can go down to 40 Ohms with our vactrol.

How’s it going so far? Let me get this straight:

The LFO is at 1 => The vactrol LED lights up => The photoresistor has low impedance
The LFO is at 0 => Vactrol’s LED turns off => Photoresistor’s has a very high impedance

It is said that a vactrol is of good quality when you can go down as low as 40 Ohms and quickly! Without latency.
The datasheet of our vactrol, the NSL-32, we use it on the Lazy Comp. It also has already been seen in several quality effects : NSL-32

Inverter amplifier

The inverter amplifier is one of the most widely used. It has advantages and disadvantages. It manages the volume of our tremolo with the possibility to go to a zero volume or to a maximum volume. On the other hand, it will cause a 180° phase shift in the signal chain.
The setup that does not invert the phase (we call it non-inverting) is not able to have a zero gain. It will be at the minimum a gain of 1, which is not practical to mute the sound!

In audio we don’t only talk about resistance but more about impedance. It takes into account the resistors but also the capacitors, which is more precise on the whole spectrum. So the gain = -Z2/Z1.

The inverter amplifier and its vactrol

Let’s now identify Z1 and Z2 in our assembly and understand how our gain evolves over time!

Z2 = R12 in parallel with F5 and with the potentiometer connected between G’ and F
Z1 = Rvactrol + Rtrimpot

Regarding Z2; We have chosen F5 in such a way that all frequencies from 20Hz to 15kHz pass without being altered. Above that it is very slightly attenuated. So let’s assume that in our calculation, Z2 = R12 // 5kOhms pot.

Z2 max = 5.1k // 5k = 2.5kOhms.
Z2 min = 5.1k // 0 = 0 Ohms.

How do I do the maths? The formula is simple, when 2 resistors are in parallel, the equivalent resistance follows this formula:

Regarding Z1, Z1 = photoresistor + trimpot mounted as variable resistor
At the minimum, Z1 min = 40 (photo resistor fully illuminated) + 0 (minimum setting of trimpot) = 40 Ohms.

With Z2 max and Z1 min, the gain Av = -2500 / 40 = -62.5 = +18dB
The amp will saturate like hell!
The trimpot is therefore used to set a maximum gain to never saturate by turning the volume knob.

By setting the trimpot to 50%, Z1 min = 40 + 500 = 540 Ohms, a gain of 5 now.
When the Trémolo pedal is delivered, we set the Gain trimpot to 60%. Then Z1 min is therefore equal to 440 Ohms. This is a gain of 5.7 with the volume at full volume. That is +7.5dB. With the volume potentiometer at maximum.
On the other hand, for those who like the wild tones, don’t hesitate to put it back to maximum. It’s going to crunch seriously!

Let’s come back to our gain. Av = -Z2/Z1.
When, Av = 0 no sound passes, then if Av != 0, the sound goes through and the volume will be proportional to Av.

Assuming the volume pot is at maximum, Z2 = 2500 Ohms.
If LFO at 1 => LED lights up => Z1 = 440 Ohms => Av = -Z2/Z1 = -5.7 => We increase the volume.
If LFO at 0 => LED turned off => Z1 = 5 MOhms => Av = – 0.0005 = -33dB => We mute the sound.
If LFO at 50% => LED on at 50% => Z1 = 5 kOhms approx. => Av = -0.5 => We get half volume.

Wonderful, isn’t it? We’ve just created an amplitude modulation, and thus a tremolo, with a few components!

Want to go further? Start making an FX Teacher kit!

The harmonic tremolo

In audio

This time, the operation is a little more complex but completely affordable for beginners in electronics!

The objective is to separate the signal into 2 branches. Then, we keep on one side only the treble and on the other side only the bass. Hence on the previous diagram, the LPF (Low Pass Filter) and the HPF (High Pass Filter).

Then, each branch enters an optical tremolo cell similar to the one studied before. An amplitude modulation (AM) !
And finally these 2 branches will add up to make one! (Tube Summing Amp in this diagram)
That way, nothing crazy, we separate and then put the frequency channels together!

Except that the trick is to create a 2nd LFO which is always in phase opposition with the 1st one.
So when the 1st LFO is at its maximum, the 2nd is at its minimum.
So, when the bass is off, the treble takes over! And, vice versa.

All the quality of a harmonic tremolo will therefore be in its signal processing, which cut-off frequencies should I choose to create my bandwidths? Which gains to apply? How to avoid to generate noise?
This is what we worked on to develop the Ages. We also took advantage of the fact that we have embedded an intelligence to propose a more lively interaction with the pedal! The attack detector!

In electronic

Here, the analog signal processing can be split into 3 steps.
The first one consists in buffering the signal in order to preserve its quality and integrity in the sound chain (see Ego Driver), the second one consists in separating the bass and treble channels and then modulating their amplitude. And finally, the 3rd and last one sums these bands and controls the output volume.

Bandpass filter

Let’s focus first on what interests us the most, the second part of the circuit. To separate bass and treble it’s rather simple, C4 and R7 form with ICI1B a low pass filter, so the bass is in the upper part of the schematic. Then, C3 and R6 form a high pass filter before entering IC1C, so the trebles are in the lower part of the schematic.

The filters are deliberately badly chosen to destroy the treble in the first case and the bass in the second. We cut at 660Hz for the low-pass filter, then at 1.5kHz for the high-pass filter. As a result, the filters are well enough studied to cross each other on the mids.

As a reminder, I get these cut-off frequencies using the formula:
fC = 1/(2PIR*C).

This circuitry allows both channels to be separated with an identical signal without degrading it since we have an active circuit with a high impedance input. Then, to filter on the desired frequency ranges. Ingenious!

Tone’s pot allows you to create a pan between bass and treble. If you set the pot to 0, the cursor will be at the bottom and therefore the bottom channel will be grounded, there will no longer be any treble. On the contrary, if you set the tone to maximum, the bass will be grounded and there will only be treble.

Output stage

Finally, the output stage is what we call a summing amplifier!
The total gain is no longer Vs = -VeZ2/Z1 but Vs = -VeZ2/Z1 – Ve’*Z2/Z1′.
Here Z2 is the pot of volume which is 50kOhms. Z1 is the variable resistance of the upper channel and Z1′ is the variable resistance of the lower channel.
Vs is the output signal of the effect.
Ve is the signal that comes out of the AOP IC1B and Ve’ comes out of IC1C.

Therefore, if our LFO only illuminates the 1st channel, Z1 will be very weak while Z1′ will be very important. We can say in this case that Vs = -Ve*Z2/Z1 with Z2/Z1 being quite strong.
So, an output signal full of bass.
If, on the other hand, we light only the 2nd branch with the LFO, we will have only treble.
Then, if the 2 LFOs are at the same level (it happens at half period), we will have a balanced signal because the 2 branches will have the same level.
But what is nice is that this signal will not have the same spectral shape as at the beginning, let’s not forget that we have cut a part of the mids!

Now you know everything! Well, almost… I’ve been working on the subject for 7 years, I’m sure I still have a lot of things to tell you. By the way… We organize masterclasses all over France and soon in the world, it would be great to meet there to exchange on the subject and make a pedal together!

The advantages of a digital brain

What if digital is far from being our worst enemy?

From analog to digital LFO

The tremolo has always been divided into 3 parts, the power supply, the amplitude modulation stage and finally the LFO generation.

The traditional analog LFO

For a good sound quality, it is essential to keep the analog amplitude modulation unit because it avoids unnecessary sampling of the signal in any processor and thus unnecessary loss of headroom and resolution.
Well… Excuse me, it costs less in production to put everything in one chip and to keep only about twenty components! In short, you’ll never see that with us 😉

On the other hand, regarding the power supply and the LFO, everything is possible! And there are obviously things to improve, since time!
On this part we will focus on LFO generation.

To summarize (this is not the subject of going into details at this point), we have an oscillator circuit on which we can vary the frequency and make it go from a sine to a square.
It’s quite simple, it doesn’t take a lot of development time and it has been done and redone dozens of times. That’s why some manufacturers choose this solution.
But then, how do you do it if you want a tap tempo? If we want craziest waveforms? Tons of presets?

For once, a part of the pedal doesn’t necessarily need to be analog since it doesn’t process the sound, but it creates control signals.
So what we have chosen to do, and others have understood it very well too, is “simply” to generate all these waveforms from a microcontroller!

The microcontroller

This simple chip integrates billions of transistors as well as groups of transistors, functions etc… A true independent system, able to generate an infinite number of LFOs, calculate a tap tempo and then apply it, load presets, subdivisions, secondary settings … In short, anything you want! And we’ve gone through ALL your requests.
On the other hand… it needs to be programmed 🙂

It was my apprentice Damien who was in charge of this project, he goes to the same engineering school as me and I thought it was great to entrust him with such a project.
For 1 year I was doing firmware development at NXP Semiconductors, it’s a job that can drive you crazy but at the same time it’s so rewarding when the product comes to life!
The difference is that in Arduino we use a rather common programming language, a kind of simplified C. At NXP we did everything in machine language, it was quite a headache!
And we spent months developing every single function… in short!

To simply summarize the operation of the system, the pots and switches are now scanned by the microcontroller. By turning the knobs, you give instructions to the microcontroller! You never touch the sound directly.
Behind, the firmware is studied in such a way that all possible combinations and interactions are imagined, anticipated and lead to a relevant action… or a big system crash! ^^
No, we spent weeks looking for the slightest vulnerabilities, all the big problems are handled, the rest of our research will focus on updates to go even further!
Moreover our chip is on socket to easily replace it and take advantage of an updated code lately !

Finally, the micro creates an LFO permanently and sends it to the vactrol, so your choices will only change the shape of the LFO by following our instructions in the manual.

Digital LFO

I’ll confess something to you, in reality, the LFO you create from this micro controller is not an analog signal! There is no digital to analog converter in the Arduino!
So you have to go through what is called a PWM.
This is still a problem, the following schematic only speaks analog! So we’re going to convert all this… by ourselves!

This diagram shows on the one hand the desired analog signal and on the other hand what the microcontroller generates at its PWM output!

It’s quite close to the frequency modulation seen above, except that here we’re playing with the duty-cycle ratio of the square signal. That is to say that when the signal level is important, we have a low duty cycle, the signal is often at 0 and goes up to 1 from time to time!
The opposite is the case with a low signal level, which results in a high duty cycle and the signal often remains at 1.
This is the only solution to get an LFO from an Arduino!

The purpose of the LFO filtering circuit is therefore to remove the square wave signal that is full of high frequencies and to smooth it until we get back only our analog signal.

This is a simple low pass at a low cutoff frequency of ~34Hz here! R4 and F2 constitute the low-pass filter on the AOP IC1A. The same filter setup as the previous stages!

Then, the output stage (low impedance) drives the vactrol’s LED and we limit the current with R2 not to make it burn !

The rest of the story you know it, we periodically turn the LED on and off to increase and decrease the gain of the output stages.

That’s it! You know all about tremolo and harmonic tremolo, now you just have to play through it!
Or… to create your own tremolo 😉

What about the Spinner?

Express yourself!

Until now, you could get more from your pedals by adding a tap tempo footswitch or an expression pedal.
Truthfully, what was called an expression pedal was more of a control pedal that would have the exact same effect as turning a knob with your foot.

Very handy obviously, but not too creative.

With the introduction of the Spinner, here comes the first expression pedal worthy of the name.

How the Spinner works

Come on, no secrets between us!

The magnificent boomerang shaped hand spinner is custom made by a friend in Nice (France), Joffrey Legouet, who also makes beautiful aluminium guitars. For the moment he only makes them for his friends, but I’ll tell you more about it as soon as I can!
With Joffrey, we’ve managed to encrust magnets inside each blade of the boomerang.
So we have three magnets with identical magnetic fields.

Every time a magnet passes in front of the “sensor” area, it will disturb the magnetic field of a hall effect sensor!
This is then processed by a micro controller and will give information to the pedal connected to the Spinner.

For example, if the Spinner is connected to the Trémolo, in accelerator mode, if the Spinner is spinning fast, the Tremolo Rate will accelerate.
Magic, isn’t it?

Once again, everything happens in the code developed by our team, the microcontroller will calculate the time between each moment the magnet passes in front of the sensor to measure its speed. Then according to the desired mode it will perform various actions.
Then the 2 modules communicate thanks to a mini jack and exchange data according to a format we have developed.

This format is very easy to use and seems to appeal to you! We will be adapting it to several of our products, and hope to share it with other manufacturers!

There you go, you know the basics of how the Spinner works and I’ll be happy to tell you more in a masterclass!
Thanks to everyone who have been hanging on and, see you soon for the next article 🙂
Especially if you have any questions, just leave a comment, it’s made for it ! I’ll answer them as soon as possible.

For the less do-it-yourselfers among you, the next part is this way: