Audio Delay

About

This is exactly what it says on the lid: an audio delay. Delay effects are super fun to play around with, and there even exist dedicated chips. There are cool expensive bucket-brigade-device delay chips - they sound nice, but require a complex schematic and usually provide tiny delay times, having to be cascaded for longer delays. However, there also is a notoriously crappy dedicated delay IC called PT2399. Many shame it for its lo-fi sound and bad overall quality - but personally i praise those characteristics instead! PT2399 is a very cheap, all-in-one delay IC, and you just need a dual op amp and some discrete parts to get it running in the basic configuration. I decided to take this device a couple notches up by adding various features and making some parts of it more reliable.

This delay's time is voltage controllable. TCV1 (attnuated) and TCV2 (direct) inputs are summed with the big TIME dial and determine the delay time. Two audio inputs, one of which also has an attenuator, and the feedback signal, are all summed up to the delay chip's input. The delayed output is broken out to the FEEDBACK O/I jacks - when nothing is patched into it, it is a plain feedback loop, but there are countless possibilities of inserting other FXs like filters, other delays, distortions and whatnot, to colourize the feedback tail. Finally, there is buffered wet and dry/wet-mix outputs, like on classic delays.

CV is vactrol-based, hence lacks a bit of linearity, but i decided to use it after countless experiments with transistor based CV: all of those made the delay chip shift pitch in a weird way while operating, so i decided to leave this for further studies. Two trimmers on the back set up the minimal delay time (useful to not let the delay block out or do those noisy suction sounds) and the time control span setting, so the active CV range can be set up quite precisely.

This delay is not entirely original and is based on literally most other PT2399 delays that are out there in the internet: while not being too special, this device works fairly well and has the custom functionality that i need. Feel free to let your imagination flow, mod it, update it -- or just build one for yourself, because audio delay effects are always great fun.

Schematic

The circuit is a frankenstein monster of some other circuits plus my own thought and experimenting.

IC3C and IC3D are a unity gain summator with the feedback loop sticked in the second stage. This is done so that the first stage's output could be used for the dry/wet mix knob - this eliminates the need for a second mixer. This might be the cause of the dry/wet knob not going 100% dry when fully CCW, and having just a tiny, but sometimes noticeable delayed sound behind. The output voltage is then divided by about 3 and is input to the delay circuit: it's a copy of René Schmitz' PT2399 Delay "Delta T" part, bar some capacitor change-ups. The output of the delay chip is DC decoupled, passed through the external effects (if patched) and re-amplified back to the original sound's level (about 3x). Finally, the outputs are simply wet output taken from the amplifier IC1A, and the dry/wet-mixture is a passive potentiometer crossfader that slides between the wet out and the first (inverted) stage of the input summator. The phase of the dry sound on this output thus will be inverted: not a big deal for me, but you can fix it by adding another inverter between C1 and the MIX potentiometer upper leg.

The time control circuit is also a two-opamp summator, but with a slight twist. First, the inputs and the initial TIME setting are invert-summed up with an inverting unity-gain summator, then only the negative part of the result is passed forward. This is done so that the output (this signal, inverted yet again by IC1B) ends up being only positive. Finally, there's a negative offset potentiometer added up to the rectified TCV sum: it sets the lowest possible voltage to drive the vactrol. It's done because at some point, the vactrol's resistance gets too big, and when it's at full dark resistance, the PT2399 just starts doing those fart noises, or just goes silent. I could put a 1M trimmer around the vactrol, but i also found out that this arrangement improves the response of the time knob/CV - to my perception, at least.

Calibrating this section is easy. Unpatch any CVs, turn the TCV1 and TIME knobs all the way down. Put some sharp sound into IN1, raise its input volume to noon. Set feedback to about 10 o'clock, mix to noon, patch the dry/wet-mix output to your output device. Trim TCV-MINIMUM until you get steady delays: then go towards the side where the delays get slower. Slowly tune it up to the point it starts being noisy (or to the point when it still is alive: at some voltage, it will either go silent or start doing weird fart noises.) Then turn the TIME knob all the way up, and dial the desired CV range with the TCV-AMOUNT trimmer. The TIME knob should evidently change the delay time all the way from fully CCW to fully CW for best effect.

Vactrols could be either off-shelf, or (my way!) a combination of GL5528 light dependant resistor and an orange/yellow 3mm LED in heatshrink tube package.

Media

Basic demo - some bandpass filter bongos through the delay. No CV on anything - just tuning time and feedback by hand.

A pattern set up with the 8-step phrase programmer - one row controls the FM pair, the other is patched to time CV. Slew outputs are used to give a bit of slidiness to the sequence.

A no-input demo of the broken out feedback loop: Doepfer VCF and VCA are patched in the path and are controlled by two independent LFOs, making for a dynamically evolving feedback path.

The unit can as well produce shorter delays, useful for chorus-like application. Slight triangular LFO is wobbling the delay time back and forth, generating an additional layer of sound around a simple sawtooth wave. Sometimes, it gives the sound a slightly reverby character, too.

Pictures

Module
left side
Module
back side
Module
right side

To top