Has Anyone Tried To Do A SS Version Of The Magnatone *True* Vibrato?

[R.G.]

First off, I have no idea what's in that pedal. It would be interesting to see what use was made of zeners to get variable resistance.

But let's think about resistance for a minute. That which we call a resistor would smell as sweet by any other name - oh, wait! That's roses, and Shakespeare. A resistor is a thing which is defined by its voltage/current relationship. For any resistor, V= I * R. If we plot that as the value of current against voltage, we get a straight line. Which is good, because if the line was not straight, the current not a linear function of the voltage, there would be distortion of any signal passing through it. As a side light, the slight change of resistance with higher voltages in carbon composition resistors is what gives them a little distortion is likely to be the reason for the "carbon comp is better" myth.

The slope of the I versus V line  IS the resistance. For high resistances, the current increases as voltage increases, but very little. The I vs. V line for high resistance is nearly flat. For low resistances, the current increases a lot for only a little voltage, so the I vs V line is much more upright.

What is the resistance of a diode? Its I vs V curve starts flat-ish until we hit about 0.5V, then it suddenly increases, bending upwards until it's almost but not quite vertical. So a diode has a resistance which is very high for low voltages across it, very low for voltages more than about 0.8V. In between, the actual resistance (the SLOPE of the I vs V line) changes smoothly between very high to very low, megohms to tens of ohms.

IT'S A VOLTAGE CONTROLLED RESISTOR!!

Mostly. If you use it for a voltage controlled resistor, you must restrict the signal voltage wobbles to less than about 25mV or distortion gets very big. We know that because we feed diodes a voltage or two to MAKE them distort. But if you will keep the signal levels low, diodes make fine voltage controlled resistors. And they have been used that way. The Thomas Organ Vox amps used four diodes in a "bridge" to capitalize on their variable resistance in exactly this way for a tremolo. They lived with the noise problems from tiny signals.

Magnatone varistors were the opposite. They too vary from a high(ish) resistance at low voltages across them to a low(er) resistance with more voltage across them. In the original varistors' case, the change in resistance was gradual over many volts and you could put a bigger signal through them without distortion.

Modern varistors have been improved by industry to be more abrupt in resistance change, making them nearly useless for the Magnatone application, but more useful for protective devices, which is how they're used.

Zeners, too have a change from essentially non-conducting to low resistance, as they pass their zener threshold and start conducting. This "knee" where the resistance changes is where you'd have to use zeners as variable resistors, and the sharpness of the knee, or lack of it, would in turn affect how big a signal you could use in the "variable resistance" region of their action. Low voltage zeners are known for having sloppy, round conduction knees, so maybe a string of low voltage zeners could be put together to use as a variable resistor.

I've used diodes for variable resistors. It's a tough thing to keep the signal levels small, keep the resulting noise from amplifying the signal back up to an acceptable level, and remove or cancel the necessary much higher voltage (compared to the signal) that you have to use as a "control voltage" to make the resistance change.  They're a bit of a pain.

[bool]

...
Low voltage zeners are known for having sloppy, round conduction knees, so maybe a string of low voltage zeners could be put together to use as a variable resistor.
...

3-volt-ish zeners are usually rounded-knee enough, but have sloppy tolerances so averaging these in a string could be the ticket to having both smooth-enough response curve and repeatability.

[R.G.]

Could be. Back when I was designing power supplies, we were flatly told by the Old Grizzled Engineer that reviewed our designs to not use zeners under 4.7V for just the sloppiness/round-knee reasons. Even with a constant current driver, the raw tolerances were horrible, and even with part selection for voltage at a given current, you were probably just picking a point on the knee, not the better-fixed full conduction part of the characteristic.

Using a high-variance part like this probably works for low volume stuff, like boutique pedals. The builder gets to sift through the bin of zeners, stacking up ones to get the best sounding "roundness" and at the same time about the right voltage drop. That would be death on a high volume production line, but works for low volumes and high labor costs.

It would be interesting to see a well done study of knees in 3V zeners versus the roundness of four 1N4148s.

[jubal81]

Here's an Ampgarage thread talking about replacing varistors with both MOV & resistor and zener and resistor configurations (with diagrams)

https://ampgarage.com/forum/viewtopic.php?t=23839&postdays=0&postorder=asc&start=0

[anotherjim]

You want a sloppy zener? If you could pull the one out of a CD4046 pll package. I've never worked out what condition might get it anywhere near the promised 5.1v. I think there's a reason the pin is officially labelled as "ZEN".

[Paul Marossy]

For now, I'm happy to have a working Wobbletron with an LED/LDR instead of the FET. I would like to be able to have a little slower speed capability but overall it's a pretty cool sound, like it much better than tremolo which doesn't sound as musical. This video is just testing it before I put my self designed PCB based on  this schematic into a box this morning. I changed a few part values but nothing is substantially different.

I have a question though. When I first put this circuit together I tried using a VTL5C1 and I got nothing out the other side (in other words the LDR section had no action at all). I got nothing with a VTL5C2 either. But I cut a VTL5C7 in half and used the LDR part with a bright amber colored LED and then it worked great. Why is that? I built a Demeter tremolo some years ago and it used a VTL5C1, so I assumed it would work in this scenario as well but that was a total no go.

[Mark Hammer]

That's the circuit I used, too.

I have a small handful of sealed optos in the parts drawer, but often prefer to use discrete LEDs and LDRs unless the circuit explicitly calls for a given vactrol, simply because I can actually see if the LED is illuminated, fried, or accidentally installed backwards (remember, it's ME building  ).  I usually don't know, in advance, what sort of LED efficiency the circuit used, so the LED may be too dim, or too bright.

For my stereo unit, I used a dual-ganged 100k, simply because I had more of them.  I reduced the effective value with a parallel fixed resistor, in the manner shown in Briggs' schematic.

One of the nice things about this particular LFO design is that one could implement envelope-control of speed fairly easily, via a second LDR/LED combo in parallel with the speed control.  Pick harder and the LDR resistance in parallel with the speed pot goes lower, yielding faster speed.

[Paul Marossy]

That's the circuit I used, too.

I have a small handful of sealed optos in the parts drawer, but often prefer to use discrete LEDs and LDRs unless the circuit explicitly calls for a given vactrol, simply because I can actually see if the LED is illuminated, fried, or accidentally installed backwards (remember, it's ME building  ).  I usually don't know, in advance, what sort of LED efficiency the circuit used, so the LED may be too dim, or too bright.

I used the other half of my cut up VTL5C7 to see the LED and it was not very intense so I surmise that the LED wasn't bright enough to cause a change in the LDR. Apparently none of the three different ones I have are bright enough.

For my stereo unit, I used a dual-ganged 100k, simply because I had more of them.  I reduced the effective value with a parallel fixed resistor, in the manner shown in Briggs' schematic.

That's exactly how I would do it, with a dual ganged pot. I bet that sounds nice.

One of the nice things about this particular LFO design is that one could implement envelope-control of speed fairly easily, via a second LDR/LED combo in parallel with the speed control.  Pick harder and the LDR resistance in parallel with the speed pot goes lower, yielding faster speed.

There's an effect like that on a Kurt Rosenwinkel CD I have where it speeds and slows down according to how he is playing. I was always intrigued by that and still am. Not sure what he was using to get that effect but it's cool.

[Mark Hammer]

It's a feature I enjoy.  However, it will take some experimentation to find a set of time constants that give the right feel.  It's one thing to make an envelope follower that tracks an individual picked note or chord, and quite another to have an envelope follower track the "arc" of one's picking, such that it feels like you're gradually turning the modulation rate up and then down yourself.  And if the envelope-control is implemented by placing an LDR in parallel with the Speed pot, you have to figure out how sensitive you need to make the envelope follower, and what value of LDR resistance you need to use, to have usable changes in speed in response to your picking/strumming.  It's not impossible, but compared to the envelope-followers in autowahs, compressors, and gates, that are already worked out for us, it will involve a bit more labor and tinkering on your part.  But well worth the effort.  You could always do the same thing with a footpedal, I suppose, but then what would you operate your wah with, right? 

[Eb7+9]

Surfy Bear used a string of zener diodes in place of the varistors in his new vibrato pedal and results sound pretty good.

Sorry, sounds very limited (at full depth) to my ear ... and proves my point

NSL32-SR3 optos (available at Smallbear) can easily get controlled variations spanning 100r to 10meg using proper drive ckt / that's five decades of large signal resistance variation ... what I used in my opto-vibe pedal to see what it could yield ... way too much for practical use in fact, but at least useful in knowing how much is enough / unlike idle nay-saying

Using diodes as VCRs will give you even less linear range than FETs ... there's only one other technology that can match both range and headroom of optos

[Mark Hammer]

Because of how human hearing - and indeed all of human perception in any modality - works, vibrato requires either fast changes, or big changes in pitch, to be detectable.  The ideal vibrato sound is really one that just makes the instrument sound more "nervous" rather than boing-ey.  I found that, as much as I like nice slow modulation rates, the pitch wobble at slow rates was very hard to detect, even with the intensity turned up full.  HOWEVER, with a stereo unit, the pitch wobble was much easier to perceive, even at very slow modulation rates, and even with only a single stage of phase shift.  Why?  Because one is perceiving the pitch of the guitar not in relation to itself a moment ago, but in relation to the other channel.  That difference can be more easily heard, even when the modulation is slow.  That's part of why I recommend a stereo build.

[Eb7+9]

Actually it's the other way around ...

Standard dry/wet or wet only chorus/pitch vibrato gives more of a pitch bendy effect
eg., Roland Chorus pedals, JC120 amps,  Rocktron X100 etc.

While perfectly balanced anti-phase (wet-wet) vibrato gives more of an animated effect

I provided examples of this in my second stereo-vibe build page

http://www.lynx.net/~jc/stereoVibe2.html

To really pull this off you need two matched audio channels ...

... a stereo pa set exactly the same in two channels panned hard right/left
... two matched guitar/bass amps and cabs

on a pair of JCM800's it's delicious
Whereas in mono mode (both sides in sync) kinda bland

Not your typical pitch bending effect at all ...

[Mark Hammer]

YES!

A tip of the hat (and a jar of salsa if you drop by again) for recognizing what it took me an additional 13 years to stumble onto.  The stereo implementation IS more animated.  And whether one uses a phase-shift oscillator, like I did, or a more sophisticated LFO, like you did, I think the use of independent LFOs to reduce the periodicity is key to making it more animated than dizzying.

[Paul Marossy]

Because of how human hearing - and indeed all of human perception in any modality - works, vibrato requires either fast changes, or big changes in pitch, to be detectable.  The ideal vibrato sound is really one that just makes the instrument sound more "nervous" rather than boing-ey.

A few things I noticed after playing this thing for a few hours:

1. The effect is not so noticeable when doing single note lines unless the speed and/or intensity is at maximum
2. The effect is much more pronounced when playing two or more notes simultaneously
3. It seems that the lower frequencies are easier to hear the wiggling on than the higher frequencies (or it registers in my brain faster)
4. Delay after the vibrato helps to accentuate it

This all works perfectly for my style of playing so I don't see any of those things as a limitation.

I found that, as much as I like nice slow modulation rates, the pitch wobble at slow rates was very hard to detect, even with the intensity turned up full.  HOWEVER, with a stereo unit, the pitch wobble was much easier to perceive, even at very slow modulation rates, and even with only a single stage of phase shift.  Why?  Because one is perceiving the pitch of the guitar not in relation to itself a moment ago, but in relation to the other channel.  That difference can be more easily heard, even when the modulation is slow.  That's part of why I recommend a stereo build.

I use primarily a stereo headphone set up, so it sounds kind of like you describe I think. I need to try it with a guitar amp and see how that sounds in comparison.

[Danich_ivanov]

A few things I noticed after playing this thing for a few hours:

1. The effect is not so noticeable when doing single note lines unless the speed and/or intensity is at maximum
2. The effect is much more pronounced when playing two or more notes simultaneously
3. It seems that the lower frequencies are easier to hear the wiggling on than the higher frequencies (or it registers in my brain faster)

To make it better suited for higher frequencies try lower cap values for caps that get modulated.

If you have 100nf just like in a wobbletron, try 47nf, it should be right up the alley, or even 22nf.
I guess this is why all those old school designs used buffer before each stage.

[Paul Marossy]

A few things I noticed after playing this thing for a few hours:

1. The effect is not so noticeable when doing single note lines unless the speed and/or intensity is at maximum
2. The effect is much more pronounced when playing two or more notes simultaneously
3. It seems that the lower frequencies are easier to hear the wiggling on than the higher frequencies (or it registers in my brain faster)

To make it more suited for higher frequencies try lower cap values for caps that get modulated.

If you have 100nf just like in a wobbletron, try 47nf, it should be right up the alley, or even 22nf.
I guess this is why all those old school designs used buffer before each stage.

OK, I'll try the .047uF instead and see what that does.

[Danich_ivanov]

To correct myself a little bit, i said "capS", thinking about two of these stages, but given that there's only one, i meant cap that's just after collector, and just before "varistor" thingy. I like to think of that circuit as a filter, and that cap is what sets the range of frequencies you're modulating, and given that 100nf covers most of the spectrum guitar sits in, and you mixing that cap with out of phase signal, you're just a hint from phase cancelation, "tremolo", which is why you have to go lower in value than 100nf, untill you will really start loosing low frequencies, below 22nf (47nf seems about right), to get pronounced effect.

[Paul Marossy]

To correct myself a little bit, i said "capS", thinking about two of these stages, but given that there's only one, i meant cap that's just after collector, and just before "varistor" thingy. I like to think of that circuit as a filter, and that cap is what sets the range of frequencies you're modulating, and given that 100nf covers most of the spectrum guitar sits in, and you mixing that cap with out of phase signal, you're just a hint from phase cancelation, "tremolo", which is why you have to go lower in value than 100nf, untill you will really start loosing low frequencies, below 22nf (47nf seems about right), to get pronounced effect.

Yeah, I know which one you meant. But good to clarify that for anyone else following this thread.

[Prehistoricman]

It would be interesting to see a well done study of knees in 3V zeners versus the roundness of four 1N4148s.

I hear you. I don't have any zeners on me right now (I'll get a couple tomorrow), but here are some V-I plots for 1-4x silicon and 1-4x germanium D9E diodes.
X: 0.5V per divison, right-most border is zero volts
Y: ~2mA per division

Here's a closeup of the two diodes:



The diodes are vintage, I think. They came from a pack that looks like this:



germ x1:



germ x2:



germ x3:



germ x4:



silicon x1:



silicon x2:



silicon x3:



silicon x4:



Conclusion:
Germanium has a 2x softer transition.

[Paul Marossy]

So it turns out that if I use this vibrato with speed and intensity set to the lowest setting along with my Boss Tone "Trombetta" it sounds really cool. It's kind of like adding some synth LFO modulation to it, livens up the sound a bit.