Designing a "Cascaded Delay" PCB

[ianmgull]

Well I've got a fairly ambitious idea (for me at least). In my style of playing I often have two or even three delays on at the same time to give a "blurry wall of sound" type effect. I want to create a single PCB that has two PT2399 delays on it. They will be laid out in series for a cascaded effect. I've done a couple PCB layouts before (not quite this involved) and read RG's PCB Layout book, but have questions:

First: Assuming I use the Rebote 2.5 circuit; Are there any redundant parts I can omit as a result of having both effects on the same PCB (Filter caps, voltage divider, 5v regulator)? Or could I run into noise problems if I don't filter each power rail individually?

Second: Looking at the schematic, I was wondering if it would be possible to omit the output buffer of the "first in line" and the input buffer of the "second in line", effectively dropping one whole TL072?

btw... I'm using the Tonepad schem as a reference:

http://tonepad.com/getFileInfo.asp?id=98

[aziltz]

hey you might want to look into the BYOC Ping-Pong delay for some ideas, as well as the PTAP2 that was created to give it tap tempo.  It lets you cascade two different delay timings off the same tempo, very cool!!

[ianmgull]

Thanks for the tip, the BYOC circuit looks cool.  I like how they are using two PT2399s to effectively double the delay time. I'm thinking more along the lines of having a two independent delays in the same box, on the same PCB (for space considerations).

After staring blankly at the tonepad schematic for the past 10 minutes it looks to me like i could make due with with only one pair of buffers, sandwiching the actual delay portion of the circuit between the one set of input/output buffers.

I'm also thinking I could share the 5v regulator. I was looking at the datasheet and I'm not sure what the difference between "line regulation" and "load regulation" are. Does anybody know off hand about how much current a PT2399 draws?

[aziltz]

Thanks for the tip, the BYOC circuit looks cool.  I like how they are using two PT2399s to effectively double the delay time. I'm thinking more along the lines of having a two independent delays in the same box, on the same PCB (for space considerations).

It does both actually.  I was thinking of the parallel mode when I suggested it.  Yeah, they run them in series for longer delay time, but flip the switch and you've got two delays in parallel, which is what you want I think?  Anyway, it was just for your reference.  I'd love for someone to share a parallel delay PCB layout (I just got into etching).

Good Luck.

[JKowalski]

Does anybody know off hand about how much current a PT2399 draws?

My ICL8038/Pt2399 modulated delay pedal draws 30mA current, I very roughly estimate minus all the non-PT2399 parts the current draw would be around 15mA?

[ianmgull]

My ICL8038/Pt2399 modulated delay pedal draws 30mA current, I very roughly estimate minus all the non-PT2399 parts the current draw would be around 15mA?

Thanks, that's reassuring. I don't think I'll be anywhere close to the regulators limits.

It does both actually.  I was thinking of the parallel mode when I suggested it.  Yeah, they run them in series for longer delay time, but flip the switch and you've got two delays in parallel, which is what you want I think? 

As cool as that sounds I'm actually shooting for something much simpler. Essentially I just want to put to Rebote 2.5s on the same PCB, in the same box, independently switchable of one another, and eliminate any redundant components.

[aziltz]

ok, I follow now.  you want to toggle between them and/or have the option to use one or the other.  All on the same PCB, etc...

[jacobyjd]

You did mention that you want to cascade them. Parallel won't do that for you. You'll want to put 2 delays in series in the same box.

[Mark Hammer]

My vote is for two complete independent delays in the same box, with a mixer stage at the front end of each, and a splitter at their individual outputs.  This will give you maximum flexibility, permitting things like reprocessing prior to mixing back in.

Ideally, the advantages of such an approach really don't start to emerge unless one is looking at delays of an appreciable length, where the changes over time can be heard.  While long enough to produce audible repeats, 330msec is not really of sufficient length to show off what the reprocessing can do.  On the other hand, with both in series, you'll have well over a half second, which puts you into the range, for reprocessed outputs of delay B to be heard when fed back to the input of delay A.  That's why you'll need a mixer for each.

[ianmgull]

Mark,

If I'm not mistaken, you are referring to two delays in parallel with the ability to feed the output of one back into another? I probably should have explained what I was after better from the start:

My style of music is heavily dependent on running multiple delays in series. It is also dependent on the delay times in these different delays being essentially random. In this way the final output is a blurry messy wall of sound. Obviously parallel processing isn't the best way to achieve that. I'm not going for the "edge" type tempo based stuff here. If any of you are familiar with any of the Toronto based "post-rock" groups (GY!BE, etc) you know that this means rapid tremolo picking with two or three delays in series to melt your guitar into something that sounds very orchestral in nature. So in a way, yes, I want two completely independent delays in one box but because pedalboard real-estate is at a premium I wanted to condense it down to one PCB. I should probably just draw up a schematic later to eliminate any ambiguity.

Thanks for the help everyone,

Ian

[jacobyjd]

Mark,

If I'm not mistaken, you are referring to two delays in parallel with the ability to feed the output of one back into another? I probably should have explained what I was after better from the start:

My style of music is heavily dependent on running multiple delays in series. It is also dependent on the delay times in these different delays being essentially random. In this way the final output is a blurry messy wall of sound. Obviously parallel processing isn't the best way to achieve that. I'm not going for the "edge" type tempo based stuff here. If any of you are familiar with any of the Toronto based "post-rock" groups (GY!BE, etc) you know that this means rapid tremolo picking with two or three delays in series to melt your guitar into something that sounds very orchestral in nature. So in a way, yes, I want two completely independent delays in one box but because pedalboard real-estate is at a premium I wanted to condense it down to one PCB. I should probably just draw up a schematic later to eliminate any ambiguity.

Thanks for the help everyone,

Ian

I understand what you mean--I've used the same kind of sound before (tremolo picking and two series delays to create a wall of sound). I like to run my analog delay set to a slapback w/ feedback just short of oscillation into a quarter note digital delay with like 6 or so repeats. I like that dynamics become less of a 'how hard you pick' to more of a 'how often/fast you pick'.

That said, it sounds like you want two pt2399 delays in series, in the same box. I know Solderman put together a TINY layout to fit a pcb in a 1590a enclosure--that might be what you're looking for--you could use 9mm pots and get 6 of 'em into a B-size box w/ 2 etches of Solderman's board on 1 piece of stock.

[ianmgull]

Hey Josh, I just checked out Soldersound's page and that is a great place for me to start. Thanks for the help.

[jacobyjd]

Hey Josh, I just checked out Soldersound's page and that is a great place for me to start. Thanks for the help.

No problem! Good luck getting your project together

Now you've got me thinking...that's never safe...

[Mark Hammer]

Use a 1590C chassis and layer the boards on top of each other with spacers.  Plenty of vertical space for doing what you need to do, with a footprint that isn't much bigger than a 1590BB.

Actually I was thinking in terms of running multiple delays in series.  But if you have a pair of mixable inputs and a pair of reasonably isolated outputs from each delay, the options it opens up are huge.  You don't HAVE to use them, though.  You could treat it as if it wre just two complete self-contained delays.  But if you didn't feel like it for a moment, you could go really crazy...in a good way.

[jacobyjd]

Use a 1590C chassis and layer the boards on top of each other with spacers.  Plenty of vertical space for doing what you need to do, with a footprint that isn't much bigger than a 1590BB.

Actually I was thinking in terms of running multiple delays in series.  But if you have a pair of mixable inputs and a pair of reasonably isolated outputs from each delay, the options it opens up are huge.  You don't HAVE to use them, though.  You could treat it as if it wre just two complete self-contained delays.  But if you didn't feel like it for a moment, you could go really crazy...in a good way.

Throw a micro-POG in there somewhere and you've got some chimey nirvana waiting...

[ianmgull]

Ok, so I've started on my own PCB anyway. Here's a question for you:

On the tonepad schematic for the rebote 2.5 there are three caps going from pin 1 of the PT2399 to ground. Pin 1 is also connected to the +5v rail from the voltage regulator. If I put two Rebotes on the same PCB sharing the same +5v power supply it wouldn't be necessary to ad these caps for each of the PT2399s would it? It seems like one set could be removed but I don't really know what I'm talking about so someone please let me know what you think.


Thanks!!!


schem:

http://tonepad.com/getFileInfo.asp?id=98

[sean k]

I think the basic jist of two caps in parallel is that the internal resistance of two is less than the internal resistance of 1 at double the capacity so they, the two, become better at combatting any ripple or feedback on the earth lines. You'd look into ESR of caps to get the inside line.

That said you could most probably get away with 1 x 100uf for both pin 1's and even 1 x 47uf for both pin 2's as I think this just sits on the internal bias but you'd be pushing it because a simple rule for quietness is lots of capacitance to help isolate IC's from each other so the internal crosstalk on the earths is kept to a minimum. I always put a 100uf and a 100nf on every IC and on a mixer I did years ago I also added a 1nf cap as well. It really depends how clean you want the output and how much decay is involved. If the feedback is way up then I wouldn't skip on isolation and smoothing.

[sean k]

I like this idea
I'm going to build this... including about a square meter of room for the pots! well nine actually.

[ianmgull]

OK. So this is my first time doing a PCB of this magnitude so a few more sets of eyes would be GREATLY appreciated. I know some of the labels are hard to read so please ask if you need help identifying any of the parts. I'm going to try to shrink things a bit more (don't know how I ended up with all that space in the middle). In case it isn't clear, the two ICs on the right are oriented with pin 1 facing DOWN. Also, the +5v supply will have to be jumpered to each pt2399 (unless anybody see's a path i don't).

[ianmgull]

I tried to squeeze this together a little more. I'll also post a couple different views to make it easier to read. Again, If anybody else traces this and looks for mistakes it would be very much appreciated. Just to clarify this layout as of yet is UNVERIFIED. Thanks.