Schematic critique, please - multi distortion/boost pedal

[PRR]

> will R22 and C26 turn into a low pass filter

Is a hi-pass (low-cut) so bass does not mudd-up too much. Is hi-pass at any setting of the pot.

[patricks]

Apologies, I was looking at an earlier schematic saved on my computer.
It's R39 and C40 on this schematic, and I think I got my high- and low-pass filters mixed up, I think this will be a high pass/low cut with a corner frequency of 159Hz, which will be totally fine 

[antonis]

It's a bit more complicated: 

[patricks]

Hmm, that is a bit tricky.
I'm breadboarding atm before I build up the PCB, I'll test it out in different configurations and report back 

[patricks]

UPDATE

I've tested out various aspects of the circuit, and I'm really happy with the outcomes. Can't say how grateful I am to everyone who's contributed corrections and suggestions, I've learned a lot in the process.
Version 2.0 is ready now, changes include:

• Power supply streamlined; components to double to 18V have been removed, the higher voltage made precisely no difference to the sound, to my ears.
• Gain pot on the final "dirty boost" stage changed to A10K, 50K just turned it into a distortion.
• The "warp" control (Thanks Jack Orman!) has changed from the previous schematic, now the 1k resistor and capacitor (R22 and C26 in this schematic) are in parallel with each other across lugs 2 and 3 of the pot. This gives a nice taper to the sweep of the pot, and tightens up the distortion in the bass frequencies.
  

The boost section is very much dirty, there's a high frequency hash there probably related to the poor audio handling of the JRC2902. I tried various filtering methods but anything that controlled the background distortion sound overlapped with the guitar's frequency range. I've just run with it as-is, and that sound is part of its "character".
It fixes itself a bit with the boost turned up a little way, so the partial solution is more loud = more good (I could add a 1-2k resistor in series with the pot to give it a certain minimum amount of boost but elected to leave it off for simplicity).

I'm going to build it up using the PCBs I've already ordered and make the mods on that board before I redesign the PCBs and order more.
Now the breadboard's free to set up and test the next project

[/list]

[patricks]

Another question (and happy holidays, all!)
I've built it up using one of the version 1 PCBs, and it worked first time, which is fantastic. At higher gain settings it oscillates a fair bit, I'll try lower value pots for the gain controls and see how that goes. It's probably also partly due to crosstalk between tracks in some areas, and hopefully version 2 will help with that, with more careful attention to layout.

One thing that puzzled me, though, was that I swapped in a TL074 to see what a higher slew rate op amp would do and... nothing.
Well not quite nothing, if I hit a chord really hard it would belch out a spitty bit of noise and then go silent again.
It should be pin compatible with the JRC2902 so I'm not sure what could be the cause of the difference in output between the two chips.
If you're browsing the forum over the holidays and can chip in any thoughts I'd be most grateful 🙂

[patricks]

Reviving the thread, I've rebuilt the whole circuit on the breadboard to test, and confirmed the issue with the prototype on the PCB is due to the PCB design.

On the breadboard the circuit works as intended, with some oscillation at higher gain and volume settings when pushing one stage into another as expected (also because of the flying leads for connections to pots and jacks working as antennas), but it all does what it should. I tried swapping in a TL074 for the JRC2902 and that also worked (contrary to what happened when I plugged it into the PCB), and sounded "smoother" as expected.

When I redesign the PCB it'll probably be for a larger enclosure (e.g. 1590XX) to give some more space for component layout and distance between traces

[ElectricDruid]

While more distance between traces is great, the measurements we tend to deal in for pedal PCBs are all "big" anyway, so you don't need to worry too much. If oscillations are the problem, I'd look carefully at the inputs and outputs of the various stages and if there's anywhere where those *particular* tracks/traces get close to each other. Rerouting those so they're further away might help, but I'd doubt that generally just spacing things out would help much.

Also think about the off-board wiring. It's no good having the board all tidy and then twisting the input and output wires together to "make them look neat" and thereby encouraging oscillations.

RG's the master of pedal PCB layout though, so maybe he'll step in.

BTW, can we see a pic of the lovely purple Rev.1 boards, please?!? 

[R.G.]

While more distance between traces is great, the measurements we tend to deal in for pedal PCBs are all "big" anyway, so you don't need to worry too much. If oscillations are the problem, I'd look carefully at the inputs and outputs of the various stages and if there's anywhere where those *particular* tracks/traces get close to each other. Rerouting those so they're further away might help, but I'd doubt that generally just spacing things out would help much.

Good evaluation. I regularly use design rules of 0.006" traces with 0.006" spacing between them. The particular signals on the traces have a much bigger effect than the spacing. Also there's a big effect if a sensitive, high impedance and high gain input trace runs parallel to a high voltage carrying output trace for a long distance. Inter-trace capacitance goes up linearly with parallel distance, but drops off with the square of the spacing. Relatively little space increase dramatically drops the inter-trace capacitance.
So just going for more spacing might not produce much reduction in coupling. Reducing the parallel lengths is pretty reliable.

Also think about the off-board wiring. It's no good having the board all tidy and then twisting the input and output wires together to "make them look neat" and thereby encouraging oscillations.

Yep!

Another thing to think about is decoupling. Every opamp package on a PCB should have space allotted for a 0.01 to 0.1 ceramic capacitor between the power supply(...ies) and ground. Opamps are better at power supply rejection today than they once were, but the manufacturers and design engineers much better than I am insist on decoupling caps at least once per IC package to start with, until proven unnecessary. The decoupling cap should be as close to the power supply pins as you can get them. Ideally, for DIP packages the decoupling caps would be at either end of the package, or at the pin-1/pin-8 end for a single supply. If you have a ground plane defined, you can move them to just beside the power pins.

Standard DIP layout was once upon a time to stack DIPs end to end with decoupling caps between the chips and a line of power/ground traces running up the middle under the chip between the rows of pins.

oops, nattering on about power and ground again... :-)

[patricks]

Thanks very much, both!

Here are some photos, the populated board in the enclosure (minus the chip and the jets, they're on the breadboard), plus the top and bottom of the V1 PCB.
This project is very much a learning experience for me, so thanks for bearing with me, and any "facepalm" reactions to the board design are most welcome  This is my first PCB design, and the crowded layout plus using "auto placement" to fit everything in and "auto route" for the tracks is the jumping off point for a multitude of sins.

I loooove the purple PCB tho
You'll see the resistors standing on end all over the place (does this make them function as little antennas, picking up stray signals?), tracks bunched together under the chip, and running close to pot inputs and outputs.

While more distance between traces is great, the measurements we tend to deal in for pedal PCBs are all "big" anyway, so you don't need to worry too much. If oscillations are the problem, I'd look carefully at the inputs and outputs of the various stages and if there's anywhere where those *particular* tracks/traces get close to each other. Rerouting those so they're further away might help, but I'd doubt that generally just spacing things out would help much.

Good evaluation. I regularly use design rules of 0.006" traces with 0.006" spacing between them. The particular signals on the traces have a much bigger effect than the spacing. Also there's a big effect if a sensitive, high impedance and high gain input trace runs parallel to a high voltage carrying output trace for a long distance. Inter-trace capacitance goes up linearly with parallel distance, but drops off with the square of the spacing. Relatively little space increase dramatically drops the inter-trace capacitance.
So just going for more spacing might not produce much reduction in coupling. Reducing the parallel lengths is pretty reliable.

Also think about the off-board wiring. It's no good having the board all tidy and then twisting the input and output wires together to "make them look neat" and thereby encouraging oscillations.

Yep!

Another thing to think about is decoupling. Every opamp package on a PCB should have space allotted for a 0.01 to 0.1 ceramic capacitor between the power supply(...ies) and ground. Opamps are better at power supply rejection today than they once were, but the manufacturers and design engineers much better than I am insist on decoupling caps at least once per IC package to start with, until proven unnecessary. The decoupling cap should be as close to the power supply pins as you can get them. Ideally, for DIP packages the decoupling caps would be at either end of the package, or at the pin-1/pin-8 end for a single supply. If you have a ground plane defined, you can move them to just beside the power pins.

Standard DIP layout was once upon a time to stack DIPs end to end with decoupling caps between the chips and a line of power/ground traces running up the middle under the chip between the rows of pins.

oops, nattering on about power and ground again... :-)

Those are great points, thank you!
I was originally trying to keep the power supply components off to the side of the board ABs away from signal traces but I'll revise that with the decoupling cap.
The proximity of the jacks and the pots probably doesn't help, either.