Lunetta Reverb Circuit

[Cynosure]

I had this idea for a while, but just got around to testing it out tonight. I used two 4006 shift registers to create a digital delay line. It works well with lunetta devices because it can only output square waves, but you can try it with other sound sources if you want. It might sound a little odd though - like the delay is being overdriven.

The attached sound demo starts with just the source waveform, and then I turn up the level of the reverb. Then I play with the pitch of the source. Lastly, I adjust the rate of the delay.

Here is the basic run-down of how it works.

A clock drives two 4006 shift registers. The rate of the clock is essentially the delay rate, but at the same time it is also the sample rate so things can sound a little odd at times.

The shift registers are all set up in series so the output of one goes to the input of the next. This creates a long delay line that spans 36 steps across the two chips.

The input of the first shift register is the sound source input. In my schematic, it is a 40106 oscillator with a momentary switch. That can be replaced with any sound source that goes above the chip's input threshold. Note the need for a pull-down resistor at the input.

I then tap the shift register outputs at four different points. The earliest point has no divider, and the rest have an increased voltage division as they go down the register chain. This makes the first tap point the loudest, and the last tap point the quietest.

All four tap points are mixed with a TL074.

The mixed delayed sounds are sent through a voltage divider volume pot and mixed with the source sound. Adjust the pot controls the level of the reverb.

I also added a 0.1uF cap to ground after the level pot to remove some of the harshness in the reverb sound. This is entirely optional. I like my reverb a little bit warm. I found it makes it sound more like it passed through an actual space, and masks the fact that it is just a repeating square wave of decreasing volume.


THINGS TO TRY

1. Use a complex sound source.

2. More or less tap points.

3. Add more 4006 shift registers to the chain for really long delay time.

4. Increase the clock rate so the sample rate also improves (requires more shift register steps).

5. Try another shift register with more steps.

6. Try filtering the reverb different ways.

---

4006delay.mp3
Description:	4006 Delay Sound Demo
Filesize:	4.87 MB
Viewed:	2396 Time(s)
▶ Play    ⇩ Download

---

---

4006delay.png
Description:
Filesize:	24.69 KB
Viewed:	1357 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[tony void]

Awesome! I've wondered about this idea before but didn't think it would work. This opens up so many possibilities.

[RingMad]

Hmnmm, interesting... Alas, the 4006 is out of production I believe, and getting harder to find for some. But what about the 4031, which is a 64-bit shift register?

Anyway, I'll have to check this out one day... too much stuff on the breadboards and waiting for enclosures.

James.

[elmegil]

4015 is a dual 4 bit register. I bet 2 or 3 of them could be subbed with the same use. Max clock is only 8MHz, but I don't see that as a drawback....

[Cynosure]

[analog_backlash]

This is brilliant! Another circuit to add to my impossible list of circuits to try before I die (circuit bucket list?).

[phobos]

redrawn the schematic a bit (and removed the 10K resistors between in- and outputs of the 4006)

---

4006reverb.gif
Description:	CD4006 reverb by Cynosure, redrawn by PHOBoS
Filesize:	42.69 KB
Viewed:	1235 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[Cynosure]

Thanks phobos! <3

James - replace the 4006 ic's with one 4031 and crank up the clock rate

[phobos]

Looking at the schematic again now and I think you can remove the two 1uF caps between the opamps, because you're using
a dual supply with a real GND. And you can remove the 47k resistor to GND aswell.

[LeftyLogic]

After seeing this thread and wondering, "What's a shift register?" I did some reading and it sounds like you could do some really fun things to this circuit with a 4031 chip; it has two inputs that it switches between depending on whether the voltage at the switch pin is high or low, so I bet you could put a toggle switch on that pin, hook up your oscillator to input 1, and then feed the output back into input two for some endless looping fun!
Or, for extra weirdness you could get rid of the toggle switch entirely, divide the input signal down into the LFO range and hook THAT up to the switch pin!

Although I have to admit, I'm seriously new to this whole SDIY/Lunetta stuff so I wouldn't be surprised if none of this works, but hey, it was fun to research

[kkissinger]

Brilliant idea! You could feed two such delays and have a stereophonic lunetta.

[corex]

Using a shift register like a BBD is clever. Good idea!

[bingmachine]

I just saw the datasheet of the 40105. It's essentially 4 16-bit shift-registers, with common clock - and then some more. It should be ideal for this application, shouldn't it?