Parallel Delays.... buffers?

[finetuned]

Hi,

I'm building an effects unit with some effects in it. The chain starts with two delay units. I want them to run in parallel to each other. That's why I need to split my guitar signal into these two delay units. I want to be able to play just one of the two, or both on at the same time.

Now this problem arises:
If I turn off delay 2 (bypass with spillover), will the signal still pass trough delay 1?

See this image:


I'm guessing I need some buffers in this, but where do I put them?

Thanks!
Stan.

[Seljer]

Not really sure if its the best way do it but a while ago I drew up this for someone else, its a slightly modifiedsplitter-blend from Runoffgroove

http://www2.arnes.si/~spribe/slike/aorbaplusbloopbox.gif
with 2 DPDT footswitches you can choose either channel A, B, or A+B
add a third footswitch to bypass the entire thing

[Seljer]

I realised things could probably be a bit simpler so I spent 25 minutes drawing this

simple buffer -> 2 fx loops -> mixer
not 100% sure that it'd always retain unity gain (i think with that 150k resistor over the opamp it gives x1.5 voltage gain, change it to 100k for unity...I basically ripped of the mixer thing on AMZ Mixer)

but then it occured to me that with 2 3DPT switches you could get the scheme from your diagram but slightly better in regards to true-bypassish stuff

No indicator LEDs though, I think you'd want to place a buffer on the input side, I'm not sure what you'd do with the output side

edit: for the LEDs in that last one- with the buffer in front, I think you could get rid of the sections of the switches on the left side of the diagram (just connect evething together), then use them for the LEDs without too many problems

[finetuned]

Thanks a lot Seljer, I'll take a good look at those shemes. I really appreciate your help!

[finetuned]

Oh, but you can confirm that the way I originally drew it up will not work?
(Oh, an to your last image: I didn't use the 3pdt's because I wanted delay spillover into my output signal. I am actually using 2PDT's so the input of the delay connects to the ground when deactivated.)

[d95err]

Oh, but you can confirm that the way I originally drew it up will not work?

The original design has two potential flaws.

First, the lack of a proper mixer. Just connecting the return signals from the delays together could cause unpredictable results. You need mixing resistors to ensure that they are mixed equally. A 10k resistor in series with the output from each delay circuit should work. The active mixer in Seljer's design would probably be better though.

Second, the non-true bypass design without a buffer infront. When one delay is bypassed, the input of the whole circuit will be connected directly to the output of the other delay (just follow the lines...). At this output, a low impedance signal is probably expected, so there is probably a low resistance (e.g. 10k or 100k) to ground somewhere (e.g. through the volume/level pot). This low resistance will load down the input of the circuit and cause loss of treble.

[Seljer]

if I get what you're saying:

when you turn on one of the delays you want it to start from scratch? so its no allready delaying whatever you were playing one bar earlier

and

when you turn off one of the delays you want wants to left of the delays to trail off naturally, not just abruptly stop?

[finetuned]

First, the lack of a proper mixer. Just connecting the return signals from the delays together could cause unpredictable results. You need mixing resistors to ensure that they are mixed equally. A 10k resistor in series with the output from each delay circuit should work. The active mixer in Seljer's design would probably be better though.

Okay, I understand. Adding the active mixer to the end of the chain seems easy. I'll just use the one from Seljers drawing. (What transistor do I need there, a FET of some kind?)

Second, the non-true bypass design without a buffer infront. When one delay is bypassed, the input of the whole circuit will be connected directly to the output of the other delay (just follow the lines...). At this output, a low impedance signal is probably expected, so there is probably a low resistance (e.g. 10k or 100k) to ground somewhere (e.g. through the volume/level pot). This low resistance will load down the input of the circuit and cause loss of treble.

I see, I understand.

So I go from the input into one buffer, and out of that buffer to the two delays? And I'll put an active mixer at the end of the chain.

if I get what you're saying:
when you turn on one of the delays you want it to start from scratch? so its no allready delaying whatever you were playing one bar earlier
and
when you turn off one of the delays you want wants to left of the delays to trail off naturally, not just abruptly stop?

Yes, thats it, exactly, both things you said.

[Seljer]

Okay, I understand. Adding the active mixer to the end of the chain seems easy. I'll just use the one from Seljers drawing. (What transistor do I need there, a FET of some kind?)

the triangular thing in the diagram is an opamp. You can use something like a TL072 which is common and has 2 seperate opamps in the single package, go look at the datasheet to find out which pin is what.

Though you can also make a simle buffer/mixer out of a JFET, shouldn't be hard to find a schematic.

I see, I understand.

So I go from the input into one buffer, and out of that buffer to the two delays? And I'll put an active mixer at the end of the chain.

well, to slightly refine the design:
Copy the left section of http://www.runoffgroove.com/splitter-blend.html (up to the 2 SENDS)
from the schematic I drew up http://www2.arnes.si/~spribe/slike/dualparallelloop.GIF , move the switches after the 2 delays (these basically mute the channel you're not using) so they're in front of the delays.
Keep the mixer section as it is.

though that doesn't keep the delays trailing if you turn both of them off, but that wouldn't be too hard to add (more buffers and stuff...)

[Seljer]

Ok, I'm not 100% sure on this but I think this should do what you want, I haven't tested it (could somebody smarter than me could check the workings of it if possible? only off the schematic if there are any mistakes....)



I'm pretty sure some things could be refined...

basically I set up 3 buffers at the input there, 2 of them go to the sends for each effect loop. The switches are setup so they mute the delay you're not using but whevenever one of loops is turned on then the clean line is also muted (so if both delays are turned off the clean line goes through normally). Then its the same deal as before with the mixer.

[Samugo]

For Seljer:

Can you post the eagle file? So i can make the pcb...Thanks

[finetuned]

Thanks Seljer, I love the way you figured out the switching! It looks like what I need.
Just a newbie-question: what do all the capacitors and resistors in the signal path do? Are they highpass filters of some kind or what?

Thanks again!
Stan.

PS: I read the NE5532 'sounds' better than the TL072....any comments on that?

Stan.