I know this is a mind boggler for some, but does anyone know of anyone who has developed a reverb that doesnt use IC's, just transistors (and passives).
Let me know if you know of anything.
Andrew.

 

I built a tube guitar amp for a friend back in 1960 and built a spring reverb into it. I made the reverb unit out of a 10" long spring and a couple of old headphone magnetic ear pieces. Here is a transistorized version:
https://www.afiata.com/simple-spring-reverb-circuit/

 

Yeah i was going to mention the spring reverb too.
A pure transistor version would be possible but require a lot of transistors. If you look back a the "bucket brigade" type of integrated circuit using CMOS technology you can see how to make one. A lot of transistors though to do with just discrete transistors. Good luck with it.

 

What is the reverb effect on a waveform anyway? If a distortion pedal is clipping and doing that which I can not describe right now (I'm picturing the shape)

I can't remember the diff bettween echo's and reverb on the physic level (I have reverb for guitar tho), is it just adding really really small echo's, into a noise floor in essence ? ?

I need to scope my guitar processor

 

What is the reverb effect on a waveform anyway? If a distortion pedal is clipping and doing that which I can not describe right now (I'm picturing the shape)

I can't remember the diff bettween echo's and reverb on the physic level (I have reverb for guitar tho), is it just adding really really small echo's, into a noise floor in essence ? ?

I need to scope my guitar processor

Delay, echo, and reverb are all terms with a little flexibility and overlap. Each word is often misused (at least in my opinion) to describe one of the other effects.

Generally speaking delay means literally just playing back the same sound again with a certain amount of time in between. There's often a provision for feeding the delay output back to the input at a lower level so that you hear many evenly spaced delays at diminishing levels. The simplest physical delays would be either tape delay machines, or even just using two mics to capture the same source, but with one mic closer to the sound source so that there's a time difference between when the two mics capture it. Delay times can be set short enough that you don't perceive two distinct sounds, and instead here a comb filtering effect (phasing, flanging, etc.) They can also be long enough to register as discrete echos, maybe even timed to match the beat of a song.

Reverb is composed of tons overlapping delay paths (probably with widely varying delay times) that wash together until you can't pick out discrete delays. Simple mechanical reverbs are usually either plate or spring based.

I'm not really sure how I'd define echo. I'd expect to hear long, discrete delays, quite possibly also with some reverb on it, but I think echo can mean anything depending on who's selling it!

 

To put it more simply, reverb gives the effect of having numerous different delay paths of different lengths, all mixed together.

I can't even begin to imagine accomplishing that strictly with discrete electronics. I think you'd either need something mechanical (spring, plate, etc.) or a microprocessor with some nice code. Transistors alone might get you simple delay (with massive effort and a massive device,) but it's hard to imagine getting the pseudo-random effect of differing path lengths strictly with basic electronics (although I'd love to be proven wrong!)

 

Ok thanks, and I just lately watched a video on bucket-brigde delay's. It's another experiment I should try on breadboard. I should set aside some time everyday, just for little projects on breadboard in real life. I get lazy with it, I think because my BB's are cheap,

 

Here is a little something I was working on last summer using a common Bucket-Brigade IC for a Guitar pedal...
The IC is pretty flexible

 

Thanks guys

 

Let's look practically what would be needed. Let's also assume you're talking about analog, not digital.

As stated above, reverb is a bunch of different echoes (each with different delay times) mixed together. For each delay you would basically need to recreate a bucket brigade device. I don't know off the top of my head what the longest delay would be in, say, a 2 second reverb, but let's say it's around 200ms (feedback producing the longer overall reverb time).

CoolAudio makes some BBD IC's including a 4096-stage one capable of 204.8ms delay at a 10kHz clock frequency (which means a corresponding 5kHz audio bandwidth). The slower the clock frequency, the longer the delay. This topology appears to be the "complementary output topology" according to this page. The core of the delay circuit is 1 capacitor and 2 transistors per stage. That's over 8000 transistors for a 200ms delay. And that's only one delay. You would need at least several different delays, at least 10 probably, to make a useable reverb.

Then you also surely need some amplifiers, so if you're set on the discrete transistor circuitry, you've got a bunch more there.

Are you prepared to make a circuit with tens of thousands of discrete transistors?