The simplest audio delay in the world

[scrawny git]

I'd like to have a go at building some very straightforward audio delay circuits. No variables, just a fixed time delay (upwards of 1 second) which I assume is achieved with resistor values, and continual one time playback. Does anyone know a good source of simple schematics for this?

So far I've found this one:**broken link removed**, which looks nice and easy once you have the right chips.

Any ideas/experience?

 

[blueroomelectronics]

Are you going to use 1 Bit ADC?

Audio can be a very complex waveform and it is no trivial feat to time delay it.

 

[scrawny git]

I take your point, but I am not necessarily interested in reproducing the sound with any attention to fidelity. Could I not make a test circuit with a 1-bit adc and replace it with a higher bit one if I can get it to do what I want? (I am assuming bits and £s are closely connected)

 

[blueroomelectronics]

A 1 bit ADC means its a form of PWM input to a digital output.
**broken link removed**

I would suggest something more straight forward if you're studying electronics.

 

[crutschow]

Yes, you need some sort of serial digital bit stream. That's what the note "Using a 1-bit oversample ADC" on your schematic means. The "Audio Input" is from the A/D converter. If you just feed audio directly into the circuit, you will get something that has little resemblence to the original audio.

1-bit does not mean low fidelity, it just means the information is in a single bit-stream. This can be a 1-bit ADC that gives a PWM signal as described by blueroomelectronics. You could also use the output of a standard A/D converter that has a serial output. If you wanted to listen to it after the delay, you would have to convert it back to analog by passing the output to a serial input D/A converter.

A PWM signal can be converted to analog by passing the bit stream through a low-pass filter and amplifying the filtered signal, since the signal has an average value equal to the orignial digitized signal. For best response a high order filter is needed, but a simple 1-pole filter will give a useable signal if you don't need maximum fidelity.

 

[kchriste]

Here is a schematic that is similar to the one above and uses ancient components that may be hard to find:
http://www.armory.com/~rstevew/Public/SoundSynth/Reverb/ddl.htm
There used to be analog bucket brigade chips such as the MN3011 or SAD1024, designed specifically for the generation of audio delays, but they are long obsolete. A microcontroller with an ADC, PWM/DAC and a decent amount of RAM (32K) would be one way to do it.

 

[scrawny git]

I have found the attached datasheet/schematic for a simple enough looking circuit based around the MX609. This looks easier than the schematic i found which uses the MN3011 replacement: NJU9702, and it gives longer delays (up to two seconds). So now you're going to tell me MX609 is obsolete? or does this look like a possibility? If so then what's the thing marked 1,024 MHZ

 

[crutschow]

what's the thing marked 1,024 MHZ

That thing is a 1.024MHz quartz crystal. XTAL is an abbreviation for crystal. It causes the oscillator in the MX609P to oscillate at a precise frequency. The MX609P data sheet should have more info on that.

 

[Roff]

Yes, you need some sort of serial digital bit stream. That's what the note "Using a 1-bit oversample ADC" on your schematic means. The "Audio Input" is from the A/D converter. If you just feed audio directly into the circuit, you will get something that has little resemblence to the original audio.

The circuit is a simple delta-sigma ADC, but there is an error in the schematic. IIRC, this was originally published in EDN Design Ideas, and was subsequently corrected, but errors live forever on the Internet.
With the correction (see attached), it will work as a crude ADC, and audio (not another ADC output) is the input signal.

 

[crutschow]

Ron is right of course. Should have taken a closer look at the schematic to realize something was wrong with the original.

Do you know what clock frequency was recommended?

 

[Roff]

Ron is right of course. Should have taken a closer look at the schematic to realize something was wrong with the original.

Do you know what clock frequency was recommended?

I don't remember the clock rate. I ran a sim on it this morning with Fclock=100kHz, and RxCx=100 usec. It looked OK to around Faudio=5kHz. I didn't try running the sim on a WAV file and then playing back the result (which you can do with LTspice by jumping through a few hoops).

 

[crutschow]

Found the article. It was in Electronic Design **broken link removed**. The clock and time constant suggested was the same as Ron used.

 

[scrawny git]

I have now built this circuit:
**broken link removed**

and it works, and delay times are adjustable by the value of R

BUT

I would like to only hear the delayed signal, and not the original sound, does anyone know how I might be able to add a dry/wet control? or do i need to build a separate timer circuit which switched the audio on after x milliseconds?

 

[crutschow]

I have now built this circuit:
**broken link removed**

and it works, and delay times are adjustable by the value of R

BUT

I would like to only hear the delayed signal, and not the original sound, does anyone know how I might be able to add a dry/wet control? or do i need to build a separate timer circuit which switched the audio on after x milliseconds?

The second schematic in the PT2399 data sheet is for a surround/delay application and that should do what you want.

 

[valvusmusicus]

Multi tap delay question

Hi Scrawny git and others
In relation to your MX609 circuit-how would you make this a multi-tap delay, like the MN3011 with 6 taps or the 3012 with 3 please

Dense of BH22