Tape delay Simulators

[sjaltenb]

I have done some searching on the forum and have not found anything substantial. I am just wondering if there are any projects out there involving Tape echo simulators such as the old Binson Echorec, etc.

Im not looking to build a physical tape machine, of course, just interested in something that offers the same parameters with regards to the the taps the repeat styles that vintage tape delay offer.

I didn't know if anyone had somehow incorporated this into convential PT-80 (or the like) designs with success. It seems like it would be a really fun project if there were any ideas out there.

Any information would, as always, been great! Also any suggestions on existing products on the market that do a good job of this would be cool too.

**Edit: a block diagram of the signal path of a binson would be very helpful!

Stephen

[grapefruit]

A low pass filter on the output would help. You could take your output signal from the output of the LPF, and also feed this input back into the input for repeats, so that they are low pass filtered on each repeat.

A not so clean compressor would probably help too. You could put that on the input to the delay and have the feedback signal going into the compressor too.


I just had a look at the PT80 schematic and realised it already has a LPF on the output, but instead of just having a compressor on the input it also has an expander on the output. It seems that the PT80 was designed to emulate BBD delay units so perhaps by just having a compressor on the input it would sound more like a tape delay unit, as if you push signal hard onto tape it compresses. 

Cheers,
Stew.

[sjaltenb]

Sorry, I finally found this: and it looks like something I would like to attemp to build in a 1U rack:




So the question would be, what is the best DIY circuit to incorporate into this design. The circiut would not need a volume control, it could be set to MAX and the mixer volumes and switches would control that aspect.

The feedback would be neat to have, to control the amount of feedback on each "tape head", so that could be left alone. And the delay times would control the "spacing" of each head, but they coudl maybe be modded to have a range of say:

10-100
100-200
200-300
300-~ ms

So they dont stack up. I imagine this could be done fairly easily. The mixing and buffering would be a Jack Orman buffer/splitter and a bunch of SPST switches for both the feedback and output selection. Mixing for both the outputs and the feedback loops could be accomplished with a GGG Simple Mixer.

An internal power supply with 18 and 9V would simplify some of the circuits, especially the PT-80. one problem may be the actual TONE of the repeat in this design

Thoughts?

[jacobyjd]

That looks really interesting--how would you simulate the tape speed? If I'm thinking about this right, you'd basically have to lengthen or shorten the feedback for all 4 delays at an equal rate.

Are you thinking of doing this with 4 separate 2399's?

[slacker]

So the question would be, what is the best DIY circuit to incorporate into this design. The circiut would not need a volume control, it could be set to MAX and the mixer volumes and switches would control that aspect.

My initial idea was to use a rebote style circuit as the basis for it. That gives you the input buffer and the mixer. Then you just duplicate the PT2399 part so you've got 4 of them in parallel and sent the outputs to the mixer either through a master level control or individual ones. The feedback pot could just be like in the rebote but with a switch to decide which stage you want to connect it to.
You could possibly reduce the parts count by having single input and output filter stages for the PT2399s instead of just duplicating everything.

That looks really interesting--how would you simulate the tape speed? If I'm thinking about this right, you'd basically have to lengthen or shorten the feedback for all 4 delays at an equal rate.

The original idea of that diagram was to simulate an Echorec and I don't think that had variable speed so I didn't have to worry about that. It should be possible to vary the speed of all the delays at once using LDRs or transistors to control the speed of each one using a master pot. Might be tricky to keep the ratios the same though.

[orangetones]

This does look like an interesting prospect!

[jacobyjd]

So the question would be, what is the best DIY circuit to incorporate into this design. The circiut would not need a volume control, it could be set to MAX and the mixer volumes and switches would control that aspect.

My initial idea was to use a rebote style circuit as the basis for it. That gives you the input buffer and the mixer. Then you just duplicate the PT2399 part so you've got 4 of them in parallel and sent the outputs to the mixer either through a master level control or individual ones. The feedback pot could just be like in the rebote but with a switch to decide which stage you want to connect it to.
You could possibly reduce the parts count by having single input and output filter stages for the PT2399s instead of just duplicating everything.

That looks really interesting--how would you simulate the tape speed? If I'm thinking about this right, you'd basically have to lengthen or shorten the feedback for all 4 delays at an equal rate.

The original idea of that diagram was to simulate an Echorec and I don't think that had variable speed so I didn't have to worry about that. It should be possible to vary the speed of all the delays at once using LDRs or transistors to control the speed of each one using a master pot. Might be tricky to keep the ratios the same though.

So basically, 4 2399 chips with appropriate LPF setups with switchable feedback, and yeah--I was thinking LDRs for a master 'speed' control (i'd want this  if possible )--or perhaps external pots for each 'head'...that could get a little tedious to adjust though...sorry--just thinking out loud...

Adjustable LPFs would be interesting, too.

[cloudscapes]

I think Colin was thinking out-loud about something liek this a few months back on another forum. its definatelly interesting!

[aziltz]

i'm just curious, how sonically different would this be?  you're still using analog or digital delay, you're just changing how the feedback (multiple repeats) are configured, right?  Slightly different controls but the same beast, 4x?

Not trying to shoot it down, i think its interesting.  I'm just thinking outloud, if its not that different sonically, will it be worth the extra cash for 3 extra circuits?

please don't take offense i'm just curious.

[sjaltenb]

Here is a good explanation of the capabilities here...

http://www.youtube.com/watch?v=ApEdnkiEGOY

I am trying to think of the best way to use the circuits to replicate this. There would be a lot of signal routing that would need to be done. I think a more expanded block diagram of what is going on here is necessary to understand the best route to understake this project.

Slacker's might be exactly right, I'm just not thinking about it correctly. It seems there needs to almost be 4 delays for the first heads at pre-programmed times, then feeding into 4 new delays with a single feedback control, but again, pre-prammed delay times to match the first delay times of the "repeat"

Looks like there is a lot of interest!!!

[aziltz]

Here is a good explanation of the capabilities here...

http://www.youtube.com/watch?v=ApEdnkiEGOY

I am trying to think of the best way to use the circuits to replicate this. There would be a lot of signal routing that would need to be done. I think a more expanded block diagram of what is going on here is necessary to understand the best route to understake this project.

Slacker's might be exactly right, I'm just not thinking about it correctly. It seems there needs to almost be 4 delays for the first heads at pre-programmed times, then feeding into a whole new series of delays with a variable feedback....

Looks like there is a lot of interest!!!

OK, i see the difference features now.

you might be on to something big here.

[slacker]

Slacker's might be exactly right, I'm just not thinking about it correctly. It seems there needs to almost be 4 delays for the first heads at pre-programmed times, then feeding into a whole new series of delays with a variable feedback....

I think the block diagram is roughly correct, it was only based on that youtube video though so it could be wrong.
For everyone else here's the thread the diagram is from http://www.diystompboxes.com/smfforum/index.php?topic=73717.0.

In the original the time taken for the signal to go round the drum from the record head to which ever play head is selected sets the delay time. For the feedback the signal is sent from which ever play head you like back to the record head, where it's mixed with the incoming signal. If you just use one head for playing and feedback then that's exactly the same as what a normal delay pedal does, but with chips instead of a drum.
The cool thing is that you can get feedback from a different head than the one you're listening to, so you can listen to head 1 which might be say 50ms but then take the feedback signal from head 2 which might be 100ms. That's what the 4 parallel delays in the digram are supposed to simulate.

If I can find enough breadboards and time I might try and build something to see how it works.

[sjaltenb]

Ok so im trying to figure this out a little bit more:   Head=analog circuit

I understand that the switches on the right side of your block diagram are essentially the Top row in the Binson video...they control the On/off of each head

And basically the switches that turn each Head to feedback into the input are like the bottom row of switches on the Binson

And you would set the time for each Head like the actual Binson (or variable, but they would have to be  limited to 0-1, 1-2, 2-3, 3-4 so they don't "fire" out of order)

But if each circuit has to be set to a single repeat, where does the "length of swell" control come in without having a totally seperate group of delays? How would that be controlled?

The rest would be pretty straight forward, and you could have unlimited tweakability like Stereo panners to send each head to either L or R, volume for each feedback, etc etc etc

*Is this the part of the Rebote circuit that is the actual Delay section, ignoring the buffer and Power section? It would be very easy/cheap to duplicate as many times as necessary to make this project work

[grapefruit]

Ahh, so you want it like a multiple head tape delay...

The signal path could be something like this...

1. Op amp summer. Inputs are your main input and the feedback inputs.
2. Compressor. One compressor for all "heads". The output of this is the equivalent of the record head.
3. Four separate PT2399 delay circuits with LPF's would be your record head.
4. Mix the outputs of the delay circuits and dry signal for output.
5. Mix the outputs of delay circuits for feedback.

By having a compressor after the input/regen mixer it might make it more "tape like" and handle the feedback better. Possibly a compressor that distorts a bit when driven hard would be best.
You may need an inverter to keep the feedback signal in phase.

Cheers,
Stew.

[jacobyjd]

Ahh, so you want it like a multiple head tape delay...

The signal path could be something like this...

1. Op amp summer. Inputs are your main input and the feedback inputs.
2. Compressor. One compressor for all "heads". The output of this is the equivalent of the record head.
3. Four separate PT2399 delay circuits with LPF's would be your record head.
4. Mix the outputs of the delay circuits and dry signal for output.
5. Mix the outputs of delay circuits for feedback.

By having a compressor after the input/regen mixer it might make it more "tape like" and handle the feedback better. Possibly a compressor that distorts a bit when driven hard would be best.
You may need an inverter to keep the feedback signal in phase.

Cheers,
Stew.

Thanks for organizing my thoughts on this--that's exactly what I was thinking, sans compressor (which is a perfect idea).

I've had 2399s on the mind all day, so it's all a bit jumbled

[sjaltenb]

I apologize for being dense here (seriously)...but I am still not convinced this is accurate...

The Echo mode (top switches) is simple, each circuit is set for a different delay time (say, 50, 150, 200, and 310) with a single repeat, which are connected in parallel and are mixed back together to the output. Switches turn them on/off. Boom, done.

However, For the Repeat mode I just cannot seem to understand how it would work. If a delay is fed back into itself, it will simply keep repeating over and over and over until it oscillates, even if the first initial repeat is 1.

I just dont see how it would work properly without the other set of delays for repeat and swell modes...

Again, sorry if i am being really dense about this!!!

[aziltz]

I apologize for being dense here (seriously)...but I am still not convinced this is accurate...

The Echo mode (top switches) is simple, each circuit is set for a different delay time (say, 50, 150, 200, and 310) with a single repeat, which are connected in parallel and are mixed back together to the output. Switches turn them on/off. Boom, done.

However, For the Repeat mode I just cannot seem to understand how it would work. If a delay is fed back into itself, it will simply keep repeating over and over and over until it oscillates, even if the first initial repeat is 1.

I just dont see how it would work properly without the other set of delays for repeat and swell modes...

Again, sorry if i am being really dense about this!!!

The level of the repeat is attenuated everytime its run through the delay.

[sjaltenb]

The level of the repeat is attenuated everytime its run through the delay.

[grapefruit]

The level of the repeat is attenuated everytime its run through the delay.

It's not shown in the block diagram above, but all you need is a pot set up like a volume control to control the feedback level but make sure it never gets to greater than 1. You can have it set at unity gain if you want it to repeat forever...

Stew.

[jacobyjd]

So just for everyone's clarification, controls for what (I think) we're thinking would be as follows (some aren't in the block diagram)

-Compression (unless it's fixed in an orange-squeezer-like circuit)
-Some sort of pre-gain, possibly
-spst switches for each delay 'head's feedback
-spst switches for each delay head's output
-feedback attenuation for each delay head (or some sort of LDR setup for a universal control)
-delay time pots for each head--staggered for no overlap
-output mixer controls

This looks like it would give full control without anything getting out of hand.

There also might be a possibility for some modulation in here as well. Perhaps take a cue from the mid-fi clari(not) that I've been looking at so much and add a really subtle envelope-controlled vibrato to simulate tape wear.