"Leslie" rotating speaker LFO

[ElectricDruid]

Hi All,

The last couple of days I've been working on a new LFO design, thinking that it might be fun to have something inspired by the Leslie rotating speaker.

What I've done is basically a variant of my StompLFO chip (same eight waveforms), but with two frequency controls. One control for the "Fast" speed, and the other for the "Slow" speed, although you can set either anywhere between 0.1Hz and 15Hz. Then there's a button which toggles between the two speeds, and a couple of extra controls that alter the "ramping speed" - e.g. how quickly the LFO speeds up or slows down. You can set the "ramp up time" independently of the "ramp down time", and both are variable between 0.2 secs and 6 seconds.

There's one other change from the StompLFO, which is that I tweaked the lowest values in the frequency table so that the LFO will stop completely at the very lowest settings. This allows you to do dramatic "slow down and stop" effects, which is quite cool.

Here's the prototype on my workbench:



And here's a sample of how it sounds in action:
https://electricdruid.net/sounds/ExampleLeslie.mp3

The sound file goes through a few of the LFO waveforms, ramps up and down, square wave, triangle and sine. I haven't got the random waveforms debugged with the new code yet, but they should be fun too. The sound is an Erica Synths wavetable oscillator, Polivoks filter and VCA, with the Filter and VCA controlled by the Leslie LFO - so both the timbre and the volume are being modulated.

[antonis]

What..??!!!~!

7 pots, ONLY...?? 

[ElectricDruid]

7 pots, ONLY...?? 

Only 7 ADC inputs on that 14-pin chip, Antonis! I used everything I could get!

To be honest, I could really use one more (and there is a pin free, but it's a digital-only pin) since then that might allow me to have separate Depth controls for the two modes too. So you could go from slow and deep to fast and shallow. That might be good. But there's no pin, so I either have to drop that idea, drop the Offset control (which is very useful on the original StompLFO, I reckon), or move up to a 20-pin chip (and I don't really need *six* more pins - a couple would be fine).

[garcho]

Brilliant, bravo!

I have a TC Viscous Vibe with one of those programmable "mash" foot-switches, goes a long way for some leslie type effects, like switching speeds or ramping. It also "takes a while" to stop after turning it off, nice touch. Ramping speed, (or inertia?), is an important part of the sound, such a cool effect.
I have a scavenged Leslie from an old organ laying around I've been meaning to rebuild, with extra controls and MIDI of course, we'll see if I ever get around to it. A heavy thing to schlep to gigs, good for the studio though, but then why use a real one, oh lazy conundrum...

I'm very much looking forward to seeing your LeslieLFO chips available in the store!

[Mark Hammer]

I have two pedals that implement ramp-up/down with separate Fast and Slow speeds:  a Pearl PH-44 phaser and a Line 6 Roto-Machine.  In each case, there is only modest overlap in the range of the Fast and Slow speeds.  One of the benefits is that less change in speed is crammed into each degree of pot rotation, allowing for easier dialability of the two speeds.

In the case of the Pearl, the ramping time is a function of the difference in the two speeds.  Smaller differences between fast and slow speeds yields faster ramping.  The Line 6 pedal has a 3-position slide switch for fast-medium-slow ramping, irrespective of speed difference.

How much variability in ramping time is actually useful?  Hard to say.  One possibility, that I'm quite confident you could code for, is 4 different ramping speeds in each direction, but linear or logarithmic for each speed.  That might actually be useful for capturing the differential momentum of something speeding up vs slowing down; especially when trying to capture rotors of different masses.  So linear (spd 1-2-3-4) then log (spd 1-2-3-4).

While I doubt the utility of different waveforms for the deliberate emulation of rotating speakers, I would imagine the goal is to have something that is capable of being applied to more than that specific scenario, making other waveforms useful.

Now, here's the $64k question:  Larger true Leslie cabs, rather than the smaller Vibratones found in small parlour console organs, have separate rotors for woofer and mid-tweeter horns.  Is it technically possible to have two analog outs for their respective speeds, and use one of the analog ins to set "stagger"?

[iainpunk]

is there a way to control at which DC level it stops in stop mode? i really like phasers with fixed lfo/notch option next to the of moving notches.

cheers

[ElectricDruid]

Hi Mark,

Thanks for the thoughtful feedback and the selection of good ideas!

You might be right about the speed ranges. It would be possible to limit the controls to a smaller range, which would improve the control feel since you'd get more "detail". It would also serve to differentiate the two controls, since currently they are both identical, and it doesn't matter which you use as which. I think you might have convinced me - I'll put that one on the list!

The ramping time is a different story. Since the controls are just a pot acting as a voltage divider, it's simple to replace that with a resistor string and a switch (3 position slide switch, 4 or 6 position rotary, whatever) to select the junctions on the resistor chain if you don't want a fully-variable control. So the pot doesn't rule out the switch, but going for a switch would rule out a pot. So I'm leaving that one as it is.

Currently the ramping from one speed to the other is done linearly. This is because of the way I've done it - an interpolation between frequency increments for the LFO. Since I have a frequency table which is log-scaled (I usually do frequency controls on my chips so each octave covers the same amount of pot rotation) I could instead do an interpolation between the "fast" knob position and the "slow" knob position and then look-up the result in the table. That would work. Offering a choice between those two wouldn't be impossible, but since they're completely different methods and would require entirely different routines, it wouldn't be simple either.

I wasn't especially trying to go for a accurate Leslie clone, or to provide an LFO for such an endeavour. Rather, the thought was to add some of the drama of the Leslie speed-switching effect that organists use to other modulation effects and to get it into a format that guitarists could easily use. That said, you could fairly simply use a couple of chips to model the Treble horn and Bass rotor of a proper Leslie cabinet, and you'd be able to set the frequencies and the ramp-up and ramp-down times of each bit individually to get it correct. The heavier Bass rotor famously has more inertia and slower ramp times than the treble horn, and I think I remember reading that speed-up is different to slow-down too, by at least a little bit.

To answer the $64K question (the invoice is in the mail, btw ) yes, it's definitely technically possible. There's a bigger 28-pin chip with *three* 10-bit DAC outputs (the 16F1778) that I've looked at for tri-phase chorus LFOs and similar (coming to a website near you one year soon!)  and that could produce two analog outputs for both rotors and potentially offer variable "stagger". But on the original Leslie cabinet, there's no linkage between the rotors, is there? They just work independently, I thought, so a two-chip solution is in some ways simpler and just as authentic.

[ElectricDruid]

is there a way to control at which DC level it stops in stop mode? i really like phasers with fixed lfo/notch option next to the of moving notches.

Currently the only way is to choose when you stop it! That's fairly doable if you just twist the relevant frequency knob to zero when it gets close to where you want. If you stop it by having it ramp down all the way to zero, where it stops is basically random, since you don't know what part of the LFO cycle it will be in.

I've worked on this problem before for a client who wanted an LFO that did such a thing (slowed down to stop at a particular point) and it's actually a right PITA to do, since it involves working out ahead of time what the effect of the frequency changes will be and therefore how many further cycles/much further phase shift you'll get before stopping. It's awkward, or at least it is with my maths skills!

[Mark Hammer]

Hi Mark,

Thanks for the thoughtful feedback and the selection of good ideas!

Always my pleasure and honour, Tom.  My thanks in return for being so open to ideas.

You might be right about the speed ranges. It would be possible to limit the controls to a smaller range, which would improve the control feel since you'd get more "detail". It would also serve to differentiate the two controls, since currently they are both identical, and it doesn't matter which you use as which. I think you might have convinced me - I'll put that one on the list!

Thanks.

The ramping time is a different story. Since the controls are just a pot acting as a voltage divider, it's simple to replace that with a resistor string and a switch (3 position slide switch, 4 or 6 position rotary, whatever) to select the junctions on the resistor chain if you don't want a fully-variable control. So the pot doesn't rule out the switch, but going for a switch would rule out a pot. So I'm leaving that one as it is.

Fair enough.

Currently the ramping from one speed to the other is done linearly. This is because of the way I've done it - an interpolation between frequency increments for the LFO. Since I have a frequency table which is log-scaled (I usually do frequency controls on my chips so each octave covers the same amount of pot rotation) I could instead do an interpolation between the "fast" knob position and the "slow" knob position and then look-up the result in the table. That would work. Offering a choice between those two wouldn't be impossible, but since they're completely different methods and would require entirely different routines, it wouldn't be simple either.

I think we're willing to wait.  That's who good things come to, from what I'm told.

I wasn't especially trying to go for a accurate Leslie clone, or to provide an LFO for such an endeavour. Rather, the thought was to add some of the drama of the Leslie speed-switching effect that organists use to other modulation effects and to get it into a format that guitarists could easily use. That said, you could fairly simply use a couple of chips to model the Treble horn and Bass rotor of a proper Leslie cabinet, and you'd be able to set the frequencies and the ramp-up and ramp-down times of each bit individually to get it correct. The heavier Bass rotor famously has more inertia and slower ramp times than the treble horn, and I think I remember reading that speed-up is different to slow-down too, by at least a little bit.

That was sort of my reasoning behind suggesting linear AND log ramping.  Although perhaps only one of those is really needed for ramping in each direction; i.e., linear for acceleration, and log for decceleration.

To answer the $64K question (the invoice is in the mail, btw ) yes, it's definitely technically possible. There's a bigger 28-pin chip with *three* 10-bit DAC outputs (the 16F1778) that I've looked at for tri-phase chorus LFOs and similar (coming to a website near you one year soon!)  and that could produce two analog outputs for both rotors and potentially offer variable "stagger". But on the original Leslie cabinet, there's no linkage between the rotors, is there? They just work independently, I thought, so a two-chip solution is in some ways simpler and just as authentic.

I honestly don't know the answer to your question.  But a buddy in town, not far from me, services Hammond and Leslie products, with an entire basement full of cabs, and would probably know.  I'll have to ask him.

[vigilante397]

And here's a sample of how it sounds in action:
https://electricdruid.net/sounds/ExampleLeslie.mp3

I have absolutely no idea when I would ever use that sound on my pedalboard.

I absolutely need that sound on my pedalboard. NEED.

[Mark Hammer]

Well, here's what my Leslie-servicing buddy said:

"Here's the theory of operation regarding Leslie rotors. There are two
induction motors per rotor. The fast motor is directly belt driven to both
the horn and the bass rotor. The slow motor is linked to the fast motor via
a clutch wheel which the slow motor engages by way of the rotor being
thrust upon the clutch wheel when energised. As you can imagine with any
analog device, there's a considerable amount of drag when the speeds are
switched....particularly on the bass rotor. It's so massive that it takes a
fairly long period of time to get to full speed when fast and also when
slow. The horn on the other hand is much lighter and due to the thrust
bearing mount is much more responsive to speed changes. Also, since these
motors are not RPM controlled like steppers for instance, they all rotate
at slightly different speeds. So looking at the influences governing the
speed variations, there are a few factors: Belt slippage, inherent
resistance to speed change due to mass, drag, air and bearing resistance,
Also internal motor resistance from bushings  and bearings.....you get the
picture. Each Leslie is unique."

What this tells me is that, if one is aiming for a more authentic rotating speaker sound, differential ramping is in order.

[anotherjim]

I'm not sure it's the case that Leslie motor RPM is uncontrolled. Aren't they AC synchronous motors? So speed would be proportional to your supply frequency. Don't they have different diameter pulleys for 50Hz countries?
It would be true to say the two rotors are never synchronised due to tolerances in pulley diameter matching.

For multiphase, it would be worth having only 2 if there are pinout limitations. Isn't it fairly trivial to produce a 3rd phase by externally mixing 2 phases that are spaced 120deg?

[Unlikekurt]

Jim is pretty spot on.  Speed is proportional to mains frequency and thus there are adapter pulleys for when you take the band "across the pond".  Dismissing wear of bearings, belts and rubber grommets and also setting aside how much tension the owner or technician has set for the bottom belt (he belts are more like ropes than belts as well) and which pulley position they've selected for the upper rotor belt, the horn changes speed much quicker than the rotor and it can also achieve greater speed than the bottom rotor overall.
That being said, to truly mimic it, you'd need to account for those differentials and have a crossover as well (like a Leslie) so that the two elements are controlled separately.
That said, doesn't seem like that's what Tom is looking to do anyway haha

[PRR]

I'm not sure it's the case that Leslie motor RPM is uncontrolled. Aren't they AC synchronous motors? ...

I would be pretty sure they are wound as Induction motors. Line Frequency is a goal they never reach. They must "slip" to suck current to carry load and friction.

https://www.google.com/search?q=Induction+motor.+Line+Frequency&newwindow=1&source=lnms&tbm=isch&sa=X

Phonograph induction motors are sized and wound for low and predictable slip. Power-saw induction motors may slip 20% at maximum power and more when over-loaded. Room fans may slip a large percent of final speed for many seconds at start-up. Designing for low stall torque allows other design optimums (especially price...).

I suspect that a sudden change in bass rotor speed will slip at least 20%, and may not reach 2% slip ever.

The 50Hz/60Hz (20%) thing affects all speeds about the same and is worth a pulley change.

[box]

Hi All,
I made this circuit and replaced tda1022 with mn3007.  Maybe this solution from the 70's will give you some idea for modulation control.
Unfortunately as usual I can't ulpoad file.More please on PM.

box







Link to rotor demo:
https://filebin.net/te7oc9zjk4wsgsnm

[Rob Strand]

I'm not sure it's the case that Leslie motor RPM is uncontrolled. Aren't they AC synchronous motors? So speed would be proportional to your supply frequency. Don't they have different diameter pulleys for 50Hz countries?
It would be true to say the two rotors are never synchronised due to tolerances in pulley diameter matching.

Most small single-phase applications use shaded-pole motors.  Exactly the same ones you see on old turntables.    While the speed is roughly proportional to the mains frequency the exact final speed will depend of friction, the specifics of the motor, and pulleys.  A whole lot of details.   Putting that altogether they are unlikely to be synchronized.   If they were nearly synchronized it would probably end-up with an annoying slowly changing pattern.

No idea how they dealt with the different mains frequencies.   (Maybe gear ratios, or maybe they did nothing.)

[ElectricDruid]

Jim is pretty spot on.  Speed is proportional to mains frequency and thus there are adapter pulleys for when you take the band "across the pond".  Dismissing wear of bearings, belts and rubber grommets and also setting aside how much tension the owner or technician has set for the bottom belt (he belts are more like ropes than belts as well) and which pulley position they've selected for the upper rotor belt, the horn changes speed much quicker than the rotor and it can also achieve greater speed than the bottom rotor overall.
That being said, to truly mimic it, you'd need to account for those differentials and have a crossover as well (like a Leslie) so that the two elements are controlled separately.
That said, doesn't seem like that's what Tom is looking to do anyway haha

I did a bit of research into Leslie cabinets at one point because I was working on an FV-1 program to do a rotary speaker effect.
What I found was:
Drum slow 0.67Hz
Drum fast 5.7Hz
Rotor slow 0.8Hz
Rotor fast 6.67Hz
The drum has more inertia than the rotor, so speeds up and slows down more slowly than the rotor. I can't remember what I learned about the crossover. I used 800Hz, but I don't remember if I found that number or did it by ear.

While I wasn't intending going and building an analogue Leslie emulation, I *was* intending to make it easier for people to do so. This chip does the speeds you need and does the ramping for you, and with a couple of chips for the drum and the rotor, you could do something pretty authentic.
But I think the key point for me is that the Leslie speaker shows not only that we like modulation effects, but that there's something very dramatic about *changing* modulation effects, and that's the bit I wanted to get. Slow phaser is great. Fast phaser is good too. But a phaser that slowly speeds up during the height of your solo when you stamp on the button - that adds some real excitement!
It's this use of the *speed change as an effect in it's own right* that for me is the crucial bit of the Leslie speaker and the way organists have used it, and that's what I'm after with this.

[Rob Strand]

Crossover schematic,

Lows work out as f0 = 624Hz, Highs work out as f0 = 1007Hz, so a rough estimate for the crossover would be sqrt(624 * 1007) = 793Hz.    Close enough to 800Hz.   From what i can work out both drivers are supposed to be 16 ohms so reason for the crossover asymmetry isn't clear.  (Off-hand, some possibilities: it relates to the woofer impedance rising, it's a tweak to make the crossover work acoustically, helps reduce power to the highs, the 16 ohm info is wrong.)

[garcho]

I concur Tom, a Leslie sim is great and everything but only one route to take. Dynamic LFO speed and depth and that kind of thing really goes a long way in making LFO modulated sound seem "organic" and less hamfisted. That's the beauty of lots of LFOs in a modular system, or software like Renoise.

[anotherjim]

I love the slow down of a Leslie. Pity it can't be made to do it continually without speeding up in between. Or can it? Is there an equivalent of a Barberpole/Shepard tone in a modulation effect?