Rebote 3 mod query

[Steve Blackdog]

I was wondering whether the Rebote 3 design could be successfully modified to work with two pt2399s instead of three. From the schematic, there does not seem to be any particular reason to triple the IC circuit, as opposed to doubling, or for that matter quadrupling it.

what d’ya reckon? tonepad_rebote3delay.pdf

[Enzo]

Just my quick scan...

Each of the three chips is a delay, so three in series gives triple the delay of just one. The three seem to be seriesed to add up to the desired amount of delay. Eliminate one and you would limit yourself to shorter delays.

[Steve Blackdog]

Just my quick scan...

Each of the three chips is a delay, so three in series gives triple the delay of just one. The three seem to be seriesed to add up to the desired amount of delay. Eliminate one and you would limit yourself to shorter delays.

The single pt2399 gives delay of around 0.5 seconds. I understand that the 3 chip version gives 2 seconds.

Therefore, am I right in assuming that the expansion has the following effect on the length of delay:-

ic = n
1 x n = 0.5 seconds
2 x n = 1.0 s
3 x n = 2.0 s
4 x n = 4.0 s etc?

[tboy]

...am I right in assuming that the expansion has the following effect on the length of delay:-

ic = n
1 x n = 0.5 seconds
2 x n = 1.0 s
3 x n = 2.0 s
4 x n = 4.0 s etc?

All else being equal, if you need two chips to make one second of delay, shouldn't you then need eight chips to make four seconds?

[Mick Bailey]

I can't see how three chips gives 2 seconds if one gives 0.5. My thought is that the maximum delay time is 0.5 x n and 3 chips = 1.5 seconds.

[Mark Hammer]

I built one, and yes, it's around 1.5 seconds delay, stock. One CAN probably squeeze out a little more time if you're willing to trim back on the high-frequencies. I will note that, at longest delay times, the sound quality tends to suffer. Keep in mind that these delay chips EACH do A/D and D/A conversion before feeding the next chip. It's not like simply adding more RAM to a digital delay. Take the output of the last chip, feed it back and recirculate it 3 or 4 times, and the signal will have been encoded and decoded about a dozen times.

The inflection point for what one DOES with a delay, appears to be around 800msec. That is, the things many use a delay for can all be done with less than 800msec delay time. People who want more tend to use it for different sorts of tasks. If you need more delay, go digital. These Princeton chips can be bought for <50 cents. Be realistic in what you expect from them.

[g1]

According to this article, the chips are about 340mSec each in 'stock' configuration, but can do up to a second each in a degraded mode.
So the 2 seconds is just a compromise between the 1020mSec and 3 second options.
https://www.electrosmash.com/pt2399-analysis

[J M Fahey]

Didnīt check that particular circuit but in general the idea of using 2 - 3 - 4 chips is not simply getting longer delays, in fact way too long becomes musically useless or at least boring, BUT to set them at *different* delays, so sound becomes richer, more complex, better simulating multiple wall reflections in a room, instead of just a longer passageway.

That was done by Belton in their incredibly successful Belton Brick, a very credible imitation of a spring reverb.

They use two to simulate both springs, and a third modulated one to give it "boingyness", very clever.

[Mick Bailey]

The dual delay I built it had independent controls for each section. Far more musical and useful than just a long delay. It had an asymmetric modulation. In retrospect I should have made the waveform adjustable as well. It gave that offset wobble of a tape delay or went to an extreme pitch shift.