Tremolo LDR, FET Or OTA?

[POTL]

Hello everyone I wanted to build Tap Tempo tremolo, Electric Druid project is powered by Vactrol or LED + LDR. I'm interested in working in two modes, normal and harmonic tremolo. You can rarely buy Vactrol in stores, but they can be replaced with LED + LDR, I have questions.
1) Is it possible to replace LDR with Jfet (J113 as an example)? What are the pros and cons will be? If I use Jfets in harmonic mode, do I need to support them?
2) OTA What are the pros and cons. I am starting from the BOSS TR-2 circuit, but adapted for the LM13700
3) Do I need to match LDR. In general, about optics and comparison is a common question for many circuits, Envelope, Phaser.
4) I'm interested in Tap Tempo and different waveforms - square, saw, random, etc. Which method will work best?

[ElectricDruid]

In general for a tremolo I'd say the answer to the question
"LDR, FET or OTA?"
is
"All of the above!"

However, in the case of the Druid Tap Tempo Tremolo, the LED/LDR was a choice made for very specific reasons. Firstly, the TapLFO has a PWM digital output. To be turned into an analogue signal, normally this PWM output needs filtering. But a PWM signal is the best way to control the brightness of an LED, and that can be done directly with no need for filtering. Instead of trying to control the actual brightness of the LED, instead you just turn it on and off really fast, and the ratio of the "on" time to the "off" time sets the brightness. At every point in time the LED is either on or off. This improves the linearity of the response a lot. The slow response of the LDR smooths the output a bit more, but honestly even a "super quick" LDR would be more than enough filtering for the PWM.

You could adapt the circuit for other methods. For FETs you'd need to filter the PWM output and add the appropriate bias. For an OTA, you'd need to filter the PWM output and turn the result into a control current. There are commercial tremolo pedals out there that use the chip and use these approaches, and they sound great, so it's certainly possible and worth doing.

[POTL]

Thank you Do I understand correctly that there will be no difference in the sound? Will all 3 ways handle any waveform? How about noise? Clipping? or does it depend on the design? What about juxtaposition? Will there be a pair of matching jfets or ldrs to match?

[garcho]

I suggest going the LED/LDR route. The design itself deals with some issues that can make the other strategies frustrating. Like any effect with LFO, you'll have to be a little careful about power supply filtering and grounds and that sort of stuff.

there will be no difference in the sound

A textbook tremolo is volume change only, why would it change the sound? Can you hear a difference in a clean buffer between transistor, OTA, and op amp? I would go the op amp route but I doubt you'd hear much of difference in anything, except perhaps a little OTA hiss.

Will all 3 ways handle any waveform?

Why wouldn't they? I assume you mean LFO waveforms? In LFO the LF is low-frequency. If they can handle audio (higher frequency than LFOs, unless you're in some weird alias faux-ring mod territory), they can handle LFO waveforms.

How about noise?

Noise itself has nothing to do with the tremolo effect. You'll probably get most noise from an OTA but many commercial products use them with little trouble. You'll get noise from a FET, you'll get noise from an op amp. Unless it's a noiseless mechanical passive volume adjustment, there will be noise. But tremolo specific noise? No. The problem is more about the LFO introducing noise and thumps, which is common in any design with an LFO.

Clipping?

Again, this has nothing to do with tremolo. Read up on OTA vs FET vs op amp clipping if you plan on overdriving it or are trying to avoid it completely. Again, I'd go with an op amp and plan on not overdriving it since it's a amplitude dynamic effect and not a dirt box.

What about juxtaposition

Wha?

Will there be a pair of matching jfets or ldrs to match?

Try looking at some designs first to get an idea, there are only so many ways to skin a cat.

[Knobby]

I've had vague thoughts around this before, and wondered about the effect on battery life of using a vactrol or LED + LDR and whether the slower response would limit the possibility of using fast tremolo speeds to get helicopter like effects. On the other side, I also wondered whether the direct connection of a FET based solution would be more likely to result in break through of the LFO on the output. Like many things that I think of I never get around to testing them to find out!

[antonis]

What about juxtaposition

Wha?

I think he means "placing side by side" but I'm not the most skilled person in English locutions, for sure..

[Mark Hammer]

With contemporary ultra-bright LEDs, current draw for the illumination source can be minimal.  Not micro-amps, but not more than a few ma at worst.  I can't speak to what Vactrols use.  Keep in mind they are often used as replacements for something older, so I'm not expecting them to be super-efficient.  Rolling your own can offer some flexibility and improved efficiency.

The "sluggishness" of LDRs can be an advantage.  Yes, they might constrain modulation speed (as well as depth at fastest speeds), but:
1) They don't distort...ever...under any circumstances, because ALL they are is a resistor.
2) The sluggishness in response to very fast transitions can make the "turnaround" (the extreme ends of the sweep) softer, rounder, and a little more musical.

[garcho]

^ right, and a non-related part of how LDRs don't distort is their sluggishness helping to smooth out ripple

[ElectricDruid]

Do I understand correctly that there will be no difference in the sound?

There may be subtle differences in the sound. All three methods have distinctive distortion profiles. If you keep all designs in the "audibly clean" region, you won't hear any difference. If you start to drive each a little harder, you might to hear the individual "character" of each method. FETs distort in a particular way. OTAs have a soft tanh distortion curve. LDRs are slow but probably cleanest of the lot, os you get distortion (smoothing mostly) of the LFO waveofrm rather than the audio.

Will all 3 ways handle any waveform?

LDRs are intrinsically slow. They will tend to "blur" any LFO waveform, so you may not hear the difference between a triangle and a sine at faster speeds, for example. OTAs do what you tell them, and if you feed them sharp edges, you will get clicks in the output - the fast response is a curse as well as a blessing.

How about noise? Clipping? or does it depend on the design? What about juxtaposition?

Noise and clipping probably mostly depends on the design. OTAs are pretty noisy in general, sinply because their signal levels are so low. You choose either more distortion or more noise with most OTAs. The recently produced AS3080 is the cleanest I've used, I have to say. In my view, audibly less hissy than earlier chips - certainly less than the original CA3080.

Will there be a pair of matching jfets or ldrs to match?

No, I don't see why you'd have to match any components. A typical tremolo is only a single stage, so it's not like a phaser where you have four or more stages that all have to work the same (unless it's a univibe!).

HTH

[POTL]

thanks for answers. In fact, I understand that OTA, LDR & Jfet (you can also add 4066/4016 chips) offer a fairly wide range of effects, so I decided to ask in advance about matching. In the Envelope filters circuits, we use a pair of variable resistors to change the peak in the filter, we can use any device that can work as a variable resistor, the main thing is to set the desired operating range. I know that OTA does not require matching, as it is done in the design of the chip, but what about the LDR? How demanding is the circuit on the identity of the resistors? I don't remember seeing messages somewhere about the need to match two LDR or Vactrols.

In phaser circuits we have to find 4 (or more) matching transistors, however I have never heard of finding 4 (or more) matching LDR or Vactrols. Indeed I am trying to find generic components that require less time searching for exact specifications and matching, which is available in stores and is still produced. By the way, now I thought about OTA and such circuits as Small Stone, Mutron micro V and remembered one interesting solution in the engineers thumb compressor, when the OTA worked as a variable resistor, why, with all the OTA problems (noise, small margin at the input amplifier) ​​and their advantages (matching, availability in stores, no need to adjust the offset, etc.) are they not used as variable resistors in the Envelope and Phasers circuits? this is the best of both worlds

Or did I not take into account something?

[POTL]

What about juxtaposition

Wha?

I think he means "placing side by side" but I'm not the most skilled person in English locutions, for sure..

Haha I use Google translate I am sure that the translation Russian English Greek sometimes turns out to be very strange

This is a strange word (Google never ceases to amaze me, it's time to change it to Yandex) MATCHING LOOOOOOOOOOOOOOOOL

[duck_arse]

have you juxtaposed googoo and yandex?

[ElectricDruid]

thanks for answers. In fact, I understand that OTA, LDR & Jfet (you can also add 4066/4016 chips) offer a fairly wide range of effects, so I decided to ask in advance about matching. In the Envelope filters circuits, we use a pair of variable resistors to change the peak in the filter, we can use any device that can work as a variable resistor, the main thing is to set the desired operating range. I know that OTA does not require matching, as it is done in the design of the chip, but what about the LDR? How demanding is the circuit on the identity of the resistors? I don't remember seeing messages somewhere about the need to match two LDR or Vactrols.

No matching is required because pretty much every tremolo circuit I've ever seen only has a single variable element. For envelope filters, a 12dB filter is pretty typical, and that needs *two* control elements, and it usually helps if they're reasonably well matched (although "similar" will often do!).

In phaser circuits we have to find 4 (or more) matching transistors, however I have never heard of finding 4 (or more) matching LDR or Vactrols. Indeed I am trying to find generic components that require less time searching for exact specifications and matching, which is available in stores and is still produced. By the way, now I thought about OTA and such circuits as Small Stone, Mutron micro V and remembered one interesting solution in the engineers thumb compressor, when the OTA worked as a variable resistor, why, with all the OTA problems (noise, small margin at the input amplifier) ​​and their advantages (matching, availability in stores, no need to adjust the offset, etc.) are they not used as variable resistors in the Envelope and Phasers circuits? this is the best of both worlds

OTAs are certainly common for envelope filters. There are a ton of those based on the LM13700 and a state variable filter design.

There are plenty of OTA phasers too, but the noise problems get worse as you add more and more stages, so there's a limit. FETs were probably preferred in the 70s because they were cheaper than OTAs, which were still pretty exotic parts back then.

LDRs can be used too, but I wouldn't even bother trying to match them. You'd have to match dynamic charateristics as well as static "this much resistance at this much light level", so it's complicated. Plus the driving LED adds another twist - how well matched are the LEDs? Do they all produce the same amount of light for the same current?
My own FilterFX pedal is an example of a vactrol-based filter, and the Univibe is a LDR-based phaser, so all these things are possible.

The final option which you haven't mentioned is actual VCA chips. These generally offer better performance than OTAs, but often need larger power rails (whereas OTAs are happy enough on 0-9V or much less). If I was building a filter or a phaser for studio use, I'd use a +/-15V supply and base it on some VCA chips.

[PRR]

What about juxtaposition

Wha?

Juxtaposition is a perfectly good word in English.
https://en.wikipedia.org/wiki/Juxtaposition

We "NEVER" use it in casual speech. It is just too hard to say; there are usually simpler phrases.

Casual writing like on forums, rare but not unheard of. In fact a Search on this forum finds 13 examples. Unshockingly, three are from our cunning linguists Mark Hammer and R.G.

[akc1973]

"Cunning Linguists" is also something we never use in casual speech. Can get *very* dangerous...

What about juxtaposition

Wha?

Juxtaposition is a perfectly good word in English.
https://en.wikipedia.org/wiki/Juxtaposition

We "NEVER" use it in casual speech. It is just too hard to say; there are usually simpler phrases.

Casual writing like on forums, rare but not unheard of. In fact a Search on this forum finds 13 examples. Unshockingly, three are from our cunning linguists Mark Hammer and R.G.

[POTL]

ElectricDruid I did not consider VCA due to their cost and rarity. There may be cheap chips, but I only know about THAT and the cost of an effect with multiple chips would be very expensive.

[vigilante397]

I don't have anything super meaningful to add, but I but I'm too lazy to use LED/LDR combos so I buy my vactrols at DigiKey. They're out there.

[ElectricDruid]

ElectricDruid I did not consider VCA due to their cost and rarity. There may be cheap chips, but I only know about THAT and the cost of an effect with multiple chips would be very expensive.

It's true that the THAT Corp parts are expensive. Cheaper options include:

AS3360 (Dual, linear or log control)
V2162 (Dual, log control)
AS2164 or V2164 (Quad, log control)

Or since we're building an allpass filter, you could use a dedicated filter chip like the AS3320 / V3320. That can be wired as an allpass.

Most of these chips are produced by both Alfa (AS versions) and Coolaudio (V versions).

[POTL]

I don't have anything super meaningful to add, but I but I'm too lazy to use LED/LDR combos so I buy my vactrols at DigiKey. They're out there.

Which models?

[POTL]

ElectricDruid I did not consider VCA due to their cost and rarity. There may be cheap chips, but I only know about THAT and the cost of an effect with multiple chips would be very expensive.

It's true that the THAT Corp parts are expensive. Cheaper options include:

AS3360 (Dual, linear or log control)
V2162 (Dual, log control)
AS2164 or V2164 (Quad, log control)

Or since we're building an allpass filter, you could use a dedicated filter chip like the AS3320 / V3320. That can be wired as an allpass.

Most of these chips are produced by both Alfa (AS versions) and Coolaudio (V versions).

unfortunately I can only buy That from local distributors. CoolAudio is almost out of stock.