Tremolo Musings (Depth Control, LED's, VCA's, CV offset)

[Joncaster]

I'm still messing around with my harmonic trem on the breadboard, can't quite get it absolutely right,
But I've also been using this time to learn more and try things out.
Most of the time I actually just don't know what i'm doing, but improving with every session.
So forgive me if I'm completely wrong on points. I'm looking for feedback in my thinking/approach.

I've tried the transistor modulator thing (could have perfed that, was happy), but moved on to LED/LDR's to see whats what.

The things I have really battled with:
Getting LED's to light up nicely, and have a nice movement to them.
It seems I'm always fighting the forward voltages, warring with the LDR's and dividers, and the depth control isn't an elegant solution no matter what I try.
Most of that is sheer inexperience, and not having an O-scope isn't helping, i know.
Whether I take the LED straight from the LFO, AC coupled, DC coupled, through a transistor driver, high side, low side, etc...It seems to lack something.
I get a really nice movement and sound from the tremolo, and the control isn't up to it, or the control is great, but not the sound.
Speed is not a problem, its the depth, and the shades of musicality within that parameter.

One thing that has struck me:
Tremolo depth should be Unipolar out of the depth control, shouldn't it?
0-9V and the LFO waveform shifts it's Q point upwards to half supply at maximum depth. (this assumes that maximum attenuation happens at maximum voltage)
If its AC coupled, it just makes the signal louder as you increase depth, which can be drastic, and you have to compensate with limiting the gain excursion of the stage you're modulating.
The problem with a unipolar lfo signal and LED's is the forward voltage, nothing happens till 2.7V (using an NPN as high side switch, which is the nicest for an accurate pulse), and that's half the waveform.
I haven't given up the fight just yet, but man, it's a frustrating challenge getting a tremolo to sound great.

Plus, i've added a Phase Splitter to the mix, and am currently doing a bad job of biasing a transistor LTP for duty.
I found that a Cathodyne is simpler to implement, but doesn't have enough swing to really light up the LED's (works fine for the small signals a BJT needs, though)

I've also thought of going towards VCA/OTA's:
I feel like I should at least finish my LED/LDR version first, but i've been eyeing out a VCA chip I can get locally (Mitsubishi M5283P, anyone?)
Can also try the OTA thing from RG's document on them (I found an LM13700)

But that brings me back to the depth control, and Control Voltage Offset:
If I wanted to give the VCA an LFO voltage from the top down, I could use an Attenuverter? (been reading a lot of synth stuff haha)
So the CV goes from minimum depth being say, 5V and as you bring the depth up, it brings in the waveform till max depth is 5-0V

There was a handy little drawing of this offset CV in the STOMPLFO documents.

So if anyone has any insights, I'de love to hear.
Thanks

[ElectricDruid]

The LED brightness is down tot he current flowing through it, not the voltage, as you discovered. Trying to drive an LED with a voltage gives a very non-linear result, with a nasty cliff edge as you go below the forward voltage and the thing switches off altogether. I know of two solutions. One is to drive the LED with some digital signal, like PWM. This is great for PICs and so on, but not so good for analog circuits (you *could* implement a voltage-to-PWM driver like is found in PWM phasers, but it's a lot of parts - not recommended). The other way is to use the LED in the feedback loop of an op-amp so that the op-amp does the voltage-to-current conversion for you. There's a nice example on this page:

https://damienclarke.me/effects-pedals/posts/building-a-vactrol-vca/

Note he uses a rail-to-rail op-amp since his CV goes down to 0V. If you've got a op-amp LFO which wobbles about between(say) 2V and 7V, you won't have that issue.

HTH,
Tom

[Joncaster]

The LED brightness is down tot he current flowing through it, not the voltage, as you discovered. Trying to drive an LED with a voltage gives a very non-linear result, with a nasty cliff edge as you go below the forward voltage and the thing switches off altogether. I know of two solutions. One is to drive the LED with some digital signal, like PWM. This is great for PICs and so on, but not so good for analog circuits (you *could* implement a voltage-to-PWM driver like is found in PWM phasers, but it's a lot of parts - not recommended). The other way is to use the LED in the feedback loop of an op-amp so that the op-amp does the voltage-to-current conversion for you. There's a nice example on this page:

https://damienclarke.me/effects-pedals/posts/building-a-vactrol-vca/

Note he uses a rail-to-rail op-amp since his CV goes down to 0V. If you've got a op-amp LFO which wobbles about between(say) 2V and 7V, you won't have that issue.

HTH,
Tom

Thanks Tom, thats a really helpful article.
I'll order a few Opamps and go to town, expand some horizons.