LFO ? What ?

[Phend]

Hi,
I've been looking for a good description of LFO, ie Low Freq Oscillation.
So instead of oscillating around thought I would ask,
What does it do ?
How is it made ?
What effects is it "in" ?
Any other educating help on the topic.
Thanks

[antonis]

I think you take it the wrong way..

First comes "oscillator" understanding (countless textbooks..) , then comes "LF" (0 - 20Hz) and then come its applications (e.g. signal pitch or amplitude)

[ElectricDruid]

Ooh! ooh! ooh! <stands on chair with arm in air> I know this one! I know this one!

What does it do?

Since an LFO is generally (but not always exclusively) below the range of audio, it doesn't make a sound. Instead, it is used to control something else, some other parameter of the sound. If it controls volume we call it "tremolo", if it controls pitch, we call it "vibrato". Other things get more complicated - an LFO controlling a delay could be either Chorus or Flange, depending on how other things are set up. Controlling allpass filter stages and mixing in some dry signal gives you a phaser. Not mixing the dry signal gives you a variation on vibrato.

I like to think of LFO's and envelope followers / envelope generators as two different ways of doing the same thing. Both provide control signals that can alter whatever parameter you can dream up for them to control. An LFO provides a repeating variation over time. An envelope (however created) provides a non-repeating variation over time.

How is it made?

However you like! Probably the most common circuit is the "schmitt trigger-integrator LFO", which generates square waves and triangle waves. Variants of that appear in loads of circuits. "Phase-shift oscillators" turn up in a few too, (univibe, I think?) especially where you might want sine waves. There's probably a "Wein-bridge oscillator" in a pedal somewhere, but I can't think of an example. There are some that use the 555 timer chip, but it's horrendously noisy, and even the CMOS 7555 version doesn't totally fix the issues. Personally, I like to use microprocessors to generate LFOs, since it makes the circuit simpler and lets you add lots of features and waveforms too.

The important thing about LFO circuits is that *as long as you make sure the output of one is the same as another* (in terms of voltage or current, whichever is appropriate) then they're interchangeable. It's just a control signal, so it doesn't matter where it comes from. You can swap out the LFO from one pedal and paste in the LFO from another. Usually this is a question of getting the output voltage into the right range, but it's not *always* that simple. In the same vein, you can often swap an LFO for an expression pedal and swap a repeating up-down-up-down sweep with your foot waggling up-down-down-up instead! Again, it's just a question of getting the output signals to match up.

What effects is it "in" ?

Well, I mentioned a lot above, but here's a list:

• Tremolo
• Vibrato
• Phase
• Variations of those above, like the univibe, or the harmonic tremolo.
• Chorus
• Flange
• Any other modulated delay, including most things that advertise "tape wow and flutter" or "record warp".
• Autowah / Filter effects (often these are envelope controlled, but not always)

Those are the common members of the group of "modulation pedals", although I'm bound to have forgotten a couple. After that, you can do weirder/more unusual stuff, like use an LFO to pan back and forth between two different effects (that would be a tremolo variation, really), or use an LFO to control a gain stage for a variable distortion, like someone mucking with the Drive knob while you play.

Hope this helps!

[antonis]

How is it made ?

I presume Tom's "However you like!" isn't very enlightening, so let's make things more plain..
(below is the EA tremolo oscillator thing explained..)

https://www.electronics-tutorials.ws/oscillator/rc_oscillator.html

[Mark Hammer]

You could do an awful lot worse than Tom's comprehensive response.  Only a couple details to add.  Think of it as a few chocolate sprinkles on top of the whipped cream and cherry that Tom provided.

There is a general distinction made between "audio" and "sub-audio" frequencies.  Audio are those that we can hear; usually construed as 20hz-20khz.  Sub-audio is that below 20hz.

But there is a bit of overlap between the two.  Keep in mind that hearing is basically a "special case" of touch; evolved to detect at distances, those small pressure waves that are of biological and adaptive significance to us.  You've likely heard tell of the "Fletcher-Munson curves".  These plot differences in spectral sensitivity, with maximum sensitivity in the range of the human voice; very helpful with honing in on mom's voice in utero, and getting used to human speech intonation. 

Many folks can't really hear down to 20hz, but could readily feel differences in speaker vibration against their skin around and above the 20hz boundary...presuming the speaker itself can reproduce that low.  The big question is "what sorts of discriminations in frequency can we make?".  When a change in some aspect of sound is being modulated. like Tom described, the changes produced have to be detectable AS changes.  So many of the circuits/FX we find LFOs (sub-audio oscillators) in generally don't sweep much slower than once every 10 seconds (0.1hz) or much faster than 10 times a second (10hz).  Any slower and the change in sound becomes somewhat undetectable and the same is true of modulation rates faster than 10hz or so.  Some things are more detectable/perceptible at slower modulation rates than others, while other things are more perceptible at faster rates than other.  For instance, you'd rarely find a tremolo or vibrato or Uni-Vibe that would sweep much slower than 1hz.  If the depth/sweep-width was substantial enough, I suppose you probably would, but nobody really likes that much modulation depth at that slow a rate.  And even though Uni-Vibes are essentially 4-stage phasers, their LFOs rarely sweep as slowly as those on phasers, and in turn phasers don't usually sweep as slowly as flangers.  In each instance, the slowest speed of sweep is selected based on what is still audible when change is slow.

SOME "low" frequency oscillators go well into the audio range, because interesting sounds are produced when a signal is modulated quickly.  The textbook case is those circuits where the modulation of an amplifier stage at sub-audio frequencies gets you tremolo, but moving into the audio range begins to produce "sideband products", which are the sum and difference of the audio signal frequency and LFO; commonly called ring modulation.  John Hollis' "Frobnicator" is a nice example of that as is the Fairfield Circuitry "Randy's Revenge".  The old blue rackmount MXR Digital Delay has a push-pull switch that hikes up the LFO by 100x for that specific purpose.

The specific waveform of an LFO can vary with respect to its suitability for a given purpose.  For example, while most folks are reasonably happy with a triangle wave modulating a phase shifter, a sine-wave sounds more "musical" to most ears when it comes to vibrato.  The old Boss CE-1 had a few weaknesses, IMHO, but one of its strengths was that it used a different LFO waveform for chorus and vibrato (vibrato was a sine-like wave).  One of the recommended LFO waveforms for flangers (and some phasers) is what is often referred to a "hypertriangular".  That is, the LFO sweeps in triangular fashion as the sweep ascends, but turns into a sinusoidal wave at the "bottom" of the sweep.  The net effect is that it accelerates as it approaches the upper "turnaround" but deccelerates as it approaches the lower part of the spectrum, where small differences in notch location are more audible.

When the modulation is predictable, we describe it as having "periodicity" - i.e., a predictable up-and-down.  Some effects become more "musical" and more listenable for longer intervals when they are "aperiodic".  That is, no immediately discernible up and down pattern.  That could be random (as in a sample-and-hold unit employing a noise source of random voltages), or something much simpler, like the blending of two independent LFOs that occasionally sync up but go their own separate ways the rest of the time.  The Marshall SV-1 chorus/vibrato has two independent LFOs that can be blended, and I think one of their solid-state amps does as well.  Injecting some aperiodity into an effect can liven it up and make it more interesting, especially in stereo.  I made myself a dual Magnavibe a while back.  One of the circuits provides a modest pitch wobble in mono, whose subtlety people appreciate.  Two independent units, in stereo, is ridiculously lush.

[antonis]

You could do an awful lot worse than Tom's comprehensive response.

Don't know if you adress to me, Mark but if you indeed do, you've missunderstood my answer diction..

Anyway, English isn't my native language so there are occasions where a decent phrase might result into a meaning totally different from the "original" intent..

[Phend]

Wow, I've been LFOing and not even knowing it !
Newbee 101.
Thanks

[Mark Hammer]

You could do an awful lot worse than Tom's comprehensive response.

Don't know if you adress to me, Mark but if you indeed do, you've missunderstood my answer diction..

Anyway, English isn't my native language so there are occasions where a decent phrase might result into a meaning totally different from the "original" intent..

No, it wasn't directed at you.  It was meant to be a compliment to Tom.  It was MY English that wasn't so good! 

[PRR]

....SOME "low" frequency oscillators go well into the audio range....

The ARP 2500 didn't particularly distinguish. One range ran 0.06Hz to 16Hz, the other 32Hz to 8kHz, plus a one octave up or down fine tune. >500,000:1 range in one oscillator. Only Bob Pease could do better. And you could pull it quite a bit more with voltage control.

[Rob Strand]

I've been looking for a good description of LFO, ie Low Freq Oscillation.

Technically it's Low Frequency Oscillator (a noun).

When you see LFO it generally means an oscillator used for modulation.   Typically less than 20Hz but it doesn't have to be.

For audio sub 20Hz frequencies are called subsonic.     While subsonic and LFO have the same upper frequency bound the reason for the names are obviously different.

The reason LFO's are often bounded to 20Hz is because modulation over 20Hz starts breaking into the audio and this is unwanted.   You might see >20Hz on a ring modulator but these are special cases.

What you will often find is circuits for low frequencies are simply called LFO's.   That's a generic term and doesn't necessarily imply modulation.

What about a LED flasher?   LFO or not?  The label is pretty arbitrary.

In calibration standards you won't directly see LFO but what you will see is VLF (Very Low Frequency, implying a band of frequencies).   That loosely implies 0.01Hz to 10Hz although the lower bound not to clearly defined.   There is no meaning of modulation here.  People measure voltage, capacitance, resistance at these low frequencies  (and do so to a large number of digits.)   A circuit generating the such an AC voltage standard might be called an LFO but most formal documents avoid it.

https://www.semanticscholar.org/paper/A-VLF-voltage-standard-from-0.01-to-50-Hz-Shinmin-Yin/2dd4bb46bc0dfc5fa36ca20bd1cf7caae5dc29e7

https://www.semanticscholar.org/paper/NBS-Provides-Voltage-Calibration-Service-in-Range-Schoenwetter/16cd094f1f7c4354f960dbfd15f8e999982614be

LF and VLF have completely different meanings for RF people!

An example where "Low Frequency Oscillation" doesn't imply modulation, just a circuit.
https://patentimages.storage.googleapis.com/03/e4/b2/d610cbbff1cba3/US3376518.pdf

[antonis]

And an all time classic reference (Ray Marston) for practical oscillator circuits..

https://www.nutsvolts.com/magazine/article/bipolar_transistor_cookbook_part_5

P.S.
IMHO, it should be crucial to distinguish "timer/phase shift" circuit (usually RC configuration) and "gain" circuit (any amplifier configuration)..
Then, it should be much more easy to understand the way of interaction (e.g. 180^o shift via NFB or 0/360^o via PFB..)

[Phend]

Lots of Good info, How is Lfo measured or verified?  With a scope ?

[GibsonGM]

A DMM can show a 'changing voltage' that may or may not make sense...kind of a go, no-go indication.  An analog meter can show the oscillation if it's not too fast.  A scope will show the wave itself.

Sometimes one can 'hack' an LFO with an LED and CLR to see it changing, depending on if the circuit is capable of driving the LED...

[Mark Hammer]

The bottom line is that LFOs are not listened to or heard themselves.  They are used to modify the sound one IS listening to and hearing.  That is, we hear their impact.

Some real-world LFO activity is unintended, like the "beats" produced when two notes are very very close but not at identical pitch.  Imperfect envelope followers, in tandem with imperfect strings, can result in unintended and unwanted low frequency modulation of effects that use a sidechain, like autowahs, compressors, limiters, or noise gates.  And because this modulation is subsonic, what we hear - or at least think we hear - gets attributed to something else, like distortion.

[ElectricDruid]

Lots of Good info, How is Lfo measured or verified?  With a scope ?

A scope is definitely favourite. Watch out with software scopes though. I once got caught out designing an LFO and checking the output on my laptop. I couldn't get the LFO waveforms to look right at all. *Eventually* I figured out that the laptop's input had a highpass filter at about 20Hz (since it was intended for recording audio, not doing LFO measurements!), and that's why all my LFO waveforms were wrong. Doh.

Never happens with a proper benchtop scope.

Another way to test an LFO is to feed it to something you *can* hear, so get it modulating an oscillator, for example. Like that you can hear if it's working, and waveshape is pretty obvious too (square wave is BEEBARBEEBARBEEBAR, ramp wave is WEEWEEWEEE etcetc). Very tiring to listen to though.

[Phend]

U Know now I Know LFO.
Thanks to all
Hope this thread will help others with a similar question.
(Thanks to the search box)

[Rob Strand]

I throw out a challenge: what synth was the first to have "LFO" marked on the front ? Arp Odyssey?

[ElectricDruid]

I throw out a challenge: what synth was the first to have "LFO" marked on the front ? Arp Odyssey?

Ooh, good question! That's going to take a bit of serious research!!

[ElectricDruid]

I throw out a challenge: what synth was the first to have "LFO" marked on the front ? Arp Odyssey?

Does the Moog Modular's 901B Oscillator's "Lo" mode count?

Otherwise the whiteface Arp Odyssey from 1972 is the earliest I can find, so I think you might be right. The Minimoog famously *didn't* have an LFO. It copied the "Lo" mode of the 901B on the third oscillator, an oversight that even Behringer put right with their recent clone (although they still didn't fix the damn envelopes!!).

[GibsonGM]

Since LFOs and Arduinos have been mentioned in a couple of posts recently, does anyone (Tom?) have any script info for using an Arduino for generating basic waves like triangle?  I've found a few on the net (none here), and they involved other 'things' like DACs...I mean, just the arduino itself...I've had this Uno sitting here for a long time and no real use for it, it could make a neat function generator for the bench.