stereo echobase delay

[cpm]

This is my attempt at a stero ping-pong delay



I started out with the echobase schematic, duplicating stages. the stereo ping pong is achieved feeding the output from delay stage A into delay stage B. The feedback goes from B output to A input.
this way, if both delays are set to "T" delay time, repeats go at time T:outA, 2T:outB, 3T:outA... and so on.
The spatial effect when using two amps is great, and the vibrato thing adds a nice chorus to the sound.

Some problems i had was that the signal got easily distorted, notably with humbuckers, so I changed some resistors. The red marked areas are those changes from the original echobase made to clean up the sound. i think i got a nice setup with this.

Green areas are some additions:
The LFO on/off: disconecting the main LFO output is not a satisfactory method, because there will be a fixed 39k at the delay time, so the switching must eliminate that resistor.
The tails switch: I made a variation to the original switching. Input must be muted whenever the pedal is OFF, regardless of tails mode, so that no delay tails are generated while OFF.
Switchable caps at pins 9-10: this asymetric configuration will filter the delays on a more dramatic high-pass style. Long feedbacks will fade into a more treble echo. This will also increase output level. It can be done on pins 11-12 to go into a deep low-pass feedback.

All four cmos switches are in use, and there is one left opamp if using dual chips, which could be used as another buffer for a second input.

There's a lot of possibilities on providing wet, dry or mixed outputs, working out some kind of jack switching.
i am waiting your comments and corrections about this, since i'm sure there may be a lot of improvements to be done

thanks!

[Auke Haarsma]

Thanks for posting your idea!

You can bypass the LFO by taking the 1uF out of the circuit at the LFO. So, no need to disconnect that 39k resistor.

Maybe it is cool to add another feature:

There is a mod out here (coop from Slacker and Valoosj iirc) that doubles the delay-time by adding another PT2399 to the Echobase. It could/should be possible to add a switch to your schematic so you can choose: ping-pong delay or double delay time. This makes maximum use of the additional PT2399.

[slacker]

Nice work, I love the hand drawn schemo.

[cpm]

You can bypass the LFO by taking the 1uF out of the circuit at the LFO. So, no need to disconnect that 39k resistor.

There is a mod out here (coop from Slacker and Valoosj iirc) that doubles the delay-time by adding another PT2399 to the Echobase. It could/should be possible to add a switch to your schematic so you can choose: ping-pong delay or double delay time. This makes maximum use of the additional PT2399.

Nice tip on the 1uF, it resets the LFO and keeps the delay time setting static. Thanks!

there can be mono or stereo ping pong with a couple of switches ruled by the output jacks (using the stereo pin to ground as trigger). That way, for example: only connecting out A, you have  amix of "ping and pong", both outputs is for the full stereo panning, and B only is for a long delay time using both chips. Can be made with more 4066 cells, or some kind of FET switching.

A dual pot on delay time will give symetric stereo delays, or two independent pot may give a more complex pattern, like "taa-ta-taa-ta-taaa" 

Nice work, I love the hand drawn schemo.

there are still some typos there, and a couple of resistors missing
i still have a problem when bypassing the delay lines. I will try to breadboard a stock echobase to see how it behaves, for i may have misplaced something.
When closing the cmos switches for the output, i get a sudden increase in feedback level, so it goes over unity and feedback gets out of hand, all this while being "bypassed". So turning it on again turns up in a big and loud noise havoc.
It also happens when turning down the output pot at the delay out (prior to the mixer opamp). As the wiper turns lower, feedback level increases too. 

[cpm]

I tested the original echobase with the following results:

- with the 47k resistors it produces some distortion when using humbuckers (anybody else? or is it just me?)
- The feedback level is stable when coling the output switch or reducing output level

When changing both 47k to 22k, the sound is much cleaner, which is a big improvement. The rebote schematic has 10k values there, but it suffers from a lot of background noise though.
The original echobase with the two 22k resistors has the same problem with the feedback level, so its easily reproducible. Feedback growth is proportional to the lowering of output level, and closing the output switch will make it grow even faster.

I guess it has to do to some relation between resistor values, including internal values on the chip.
Coincidentially, the PT80 delay also uses 22k there, so we could find some clues there.

i would really appreciate some help here, since the echobase is a great desing.

[slacker]

That's bit strange, I'll have a think about it and see if I can work out what is happening.

[slacker]

I don't know if it's just a mistake on your schematic but you've drawn it with the 1M5 biasing resistors for the 4066 connected to ground, they should go to 5 volts. I don't know if that would cause the problem you're having though.

[cpm]

I don't know if it's just a mistake on your schematic but you've drawn it with the 1M5 biasing resistors for the 4066 connected to ground, they should go to 5 volts. I don't know if that would cause the problem you're having though.

    i misread that on the first time and took for granted from then on.
that is a very probable cause of problems. there are coupling caps though  i will try tomorrow and see what happens
strange is that those 4066 worked fine with the signal biased to ground, there is supposed to be a heavy clipping in that configuration.

thanks indeed Slacker

[cpm]

biasing hasnt solved anything. Even taking the 4066 out of the circuit doesnt make any difference.

Looking at the PT80 schematic, it has a buffer right at the output, probably to make the compander happy.
Borrowing the idea and placing a buffer right before the output bifurcations seems to have solved the issue, maybe in not a such elegant way.
i really dont know if those resistor values are appropiate (any help there?), at least it sounds ok.



I am still concerned about the distortion i noticed in the original echobase. Is it a common isue or is it just me doing something wrong?
What about all those 100n caps? For what i've seen, the usual is 1uF and bigger

thanks!

[slacker]

I can't understand how making the 47k resistors smaller would affect the feedback level.
The only thing I can think is that the lower value is allowing some signal to take a different path through the circuit and that is causing the problem.
Try putting the 47ks back in and replace the 10ks on pins 16 and 13 with 22ks, that will give the stages about the same amount of amplification. See if that solves or reduces the problem.

No one else has mentioned anything about distortion so perhaps you have a problem somewhere. Is the bypassed sound clean?

About the 100n caps, I used them originally because I had a lot of them. However, I've done the maths and combined with the other components they are large enough to let all useful frequencies through, so there's no need to make them any bigger.

[sshrugg]

there are still some typos there, and a couple of resistors missing

It's so twentieth century to call your hand-written errors "typos" 

This is incredible work.  I'm really just posting so I can encourage you guys and have it pop up under "show new replies" whenever something comes up.  I plan to build my echobase as soon as I get together enough $$$.  I'm afraid I can't offer much until then.  Keep up the awesome work, though!

there can be mono or stereo ping pong with a couple of switches ruled by the output jacks (using the stereo pin to ground as trigger). That way, for example: only connecting out A, you have  amix of "ping and pong", both outputs is for the full stereo panning, and B only is for a long delay time using both chips. Can be made with more 4066 cells, or some kind of FET switching.

A dual pot on delay time will give symetric stereo delays, or two independent pot may give a more complex pattern, like "taa-ta-taa-ta-taaa" 

  WOW!  I am TOTALLY building this.

[cpm]

a provisional schem: http://bayimg.com/LalglAabk

this is how i build it up. Its basically a doubled echobase and some additional workarounds
the schem is not verified though. There may be also some mistakes or ugly things regarding components and design.

[Radamus]

Has anyone verified the schematic yet?

[cpm]

this is the updated schematic which i think is the one that will stay for me.

[Radamus]

Does that schematic include the extra delay time?

[cpm]

it does two PT2399 in series, so the delay path goes U1->U2->U1->U2->U1...
playng with the control signals you can have the output at out_B which is U1+U2 (long delay), or out_A with the mixdown from each U1, U2, ... (shorter times), or both outs A & B with separate U1, U2 times respectively, also with clean mixed in or just the delayed (wet) sound

[Radamus]

Very cool. Now I just have to figure out how to mix this schematic with all the mods in the other echobase forum.
Thanks much,
Conrad

[Radamus]

I've printed out the schematic and I'm trying to figure out how all of this connects. If anyone could give me a brief walk-through of the signal path, I would greatly appreciate it. Also, I can't seem to figure out where the ping-pong vs. extra delay switch is.

Thanks

[Radamus]

I'm doing my best to figure out this schematic, and I've compared it to the original. I'm getting quite a bit confused. I've used the 4066 chips on most of my recent effects, and I don't understand some of their purposes here. I can tell that one of them is for the LED. The next two (moving up and to the right) are the input bypass. The two at the top and right are the output bypass, controlled by the boss/tails swicth (I'm assuming).  The one in the middle of the upper right is whether or not the two ends of the stereo are connected, which amounts to the extended delay vs. ping pong? If there are six of these, then two 4066's are needed, and more than one switch is required to activate them. I'm assuming that, if I'm using a DPDT, the LED and the in/out controls can be powered with it. The mix can be a spdt, and not necessarily a foot switch. The LED, the mix, and the boss/tails don't have to be 4066 switches

The Out_CLEAN section is also confusing. I'm guessing that gets connected to both ends (or whatever you want) to mix in the clean sound. Seems self explanatory, but I want to be certain before I make a mistake.

Both pt 2399 chips are connected back into each other, by the looks of it. So, every time the chip processes the sound, it is output to it's designated out, and also input into the other ship, which then processes the sound and puts it out to its own output and back into the other chip, etc. So that's how the ping pong part works. What changes to allow it to be just extra long delay? Or does the ping pong delay even have that as well?

What do switch 1 (a and b) and switch 3 do? It looks like switch 3 grounds the current on the control of the bottom left switch, which would turn off the LED? Does it also change the transistors in the bottom left? How do those transistors affect the circuit? Also, what is the significance of the jack rings (B and CL) in the schematic?

I have several switching jacks, and I am planning on having the two outputs go into one output jack unless the second jack is plugged in. Shouldn't be too hard to wire. I should check and see if I can find a stereo switching jack so that I can have the option of using a stereo quarter inch jack.

In the switch that separates or incorporates both sides of the stereo, what is the purpose of the capacitor connected to the right side? Is it merely to remove the voltage offset? Also, most of the input/output caps are 100nF, which is .1uF. Can I change these to larger values? I play bass, and people typically recommend that. I'm assuming 4.7 or 10uf (i've got plenty of these) will do fine. That won't change too much, right?

With the input and the output switches, why is there already a 9v supply going through a 220K resistor? How much does this drop the current? What purpose does that serve?

What is the difference between boss and tails mode? The tails allow the delay tails to taper off after the delay is bypassed. Does boss cut them off abruptly? I've never understood a lot of the switching options. I understand true bypass and I understand the 4066 chips, but I don't know what boss-style bypassing is.

Does having the feedback only on one of the two channels affect anything? Does is do the same thing as it would in the original echobase? I will admit I have no idea what the feedback does, or, even, what it really means in this case. If anyone can help me out here, that would be awesome.

Also, has anyone had success modifying this version to use the waveshape and envelope mods? I'm thinking that I want to build this thing with all of the whistles and bells, so I want to work both of those mods.

Switch 2 seems to be in charge of the boss/tails section, as well as disabling the LFO. Is this supposed to be two different switches? I'm trying to figure out how many SPDT's I need to order.

The wiring of the LFO into the two PT2399 chips is slightly different than on the original schematic. Is there any real difference in performance?

I think that's about all the questions I have. If any of the things I said are wrong, let me know.

I'm excited to get working on this. Thanks again.
Conrad

[cpm]

a lot of questions there... 

As you've guessed, chips are in series, one after the other, so delay time can be doubled as the sum of both delay times, and taking the "long" output at the out of the second chip (out B). Feedback here is called to the signal going from the output at the second PT to the input of the first one. At the output of the first chip (out A) it would be in the middle of the "long" delay time.
Using separately both outputs there is a stero ping-pong effect. Using only out B there is long delay time. Using both out A & B (mixed down into one physical output) there is short delay time.

The switching in the original echobase is ctl_A for the input to the delay line, ctl_B for the output of the delay line, and the led switching.
There is 9v through 220k into the 4066 to make a default logic "1", the control wires pull it down to ground in order to disable the 4066.
For the stereo version you can see switch B is duplicated. Also added more switches for controlling A&B mixdown (for short delays into out_A), and clean mixdown (for a 100% wet)

the original echobase uses a simple NPN inverter/buffer for the switching. Physical switches are still "bypass" and "LFO", both SPDT. Additional switching C & D is switched directly from the output jacks. When a mono jack is inserted in a stero jack its ring terminal connects to ground, so it can be used to control the 4066 switches. I used mosfet (as used in a millenium2 bypass) as a quick a dirty solution, but that can be improved with a NPN like the original bypass switching, a real CMOS inverter, mosfet pair... i guess its important to not flow significant current into the encloseure (

for the A & B control wires i changed the way the boss/tails switch behaves. i  care that no signal enters into the delay line when the effect is bypassed, in order to always get a "clean" start when the effect is engaged.

The out_clean output is just buffering the clean input (no effect), and switching automatically for a 100% wet output (no clean mix). my usage is for getting three outputs: A & B (delayed only), and C (clean output without effect).

Some 4066 are biased to GND instead of 5v. Thats weird but i settled on that to avoid noises when switching. Those made a loud "pop" when biased to 5v. It may be circumstancial or not, but it worked for me anyhow.

For the LFO part the only addition is a switch for the cap to enable/disable the oscillation