Part of circuit (filter in omnidrive) causes noise loss of signal

[mark2]

When the "filter" (circled in the schematic) is engaged, I get a loss of signal and noise on both pins 1 an 7 of U1.  The sound is similar to when the board wasn't grounded.



From this document.

V on pin 3 of U1 varies a bit (3 vs 3.3 IIRC, but can double check if needed) depending on whether switch is engaged, but I believe this is expected.

Things I've checked and seem ok:
* V on pins 8 and 4. (9ish and 0)
* Visually inspected for solder bridges
* Continuity with ground at many many points
* Replaced C3 and C4
* Reflowed U1 pins 5-7
* Resistance between pins 1 and 5 of U1, indicates R6 and R7 are ok
* Resistance across R8 and R9, on the board, indicates no cold joints.
* Switch is operating fine
* Replaced U1

Any tips on what to check next? At this point I'm ready to just pull everything out of the filter section, make sure it still works without it engaged, then redo all of it.

[Scruffie]

C3 should go to pin 7 of U1b, seems that's wrong on Hollis' schematic too.

[mark2]

C3 between 7 and the switch? And just jump where it currently sits?
Makes me wonder how others haven't stumbled on this before.

Thanks!

[mark2]

The more I've thought about this the more confused I am.  I found the pedalpcb infinidrive and it has the same schematic I'm using, albeit with one tl074 instead of two tl072: https://www.pedalpcb.com/product/infinidrive/

[j_flanders]

C3 between 7 and the switch? And just jump where it currently sits?

It's a Sallen Key filter.
This shows where its components go when using an op amp and the low pass filter version:

Your C3 is CA in the image above.
Also make sure R8 is connected to Vbias and not to ground (since you have no DC blocking cap to ground there)

[Scruffie]

The more I've thought about this the more confused I am.  I found the pedalpcb infinidrive and it has the same schematic I'm using, albeit with one tl074 instead of two tl072: https://www.pedalpcb.com/product/infinidrive/

They've dropped C4 (in your schematic) from theirs, obviously because it doesn't work as drawn and that was the modification they used instead.

But what I and j_flanders have said is the right way to do it.

[mark2]

Thank you both. I didn't see that C4 (on pedalpcb) was replaced with a resistor, and I appreciate the name for the filter circuit; this is a great learning opportunity to do some more reading to understand how this works.

I will try doing it as recommended.  On the offchance that doesn't work, I'll try copying pedalpcb's design.

Thanks again. I'll report back after I give it a try.

[mimmotronics]

Definitely seems like a Sallen Key topology. Here's a unity gain low-pass version:



And here is the Omnidrive version:


R8 and R9 are used for gain control, but the above answers are correct in saying that C3 should be connected to pin 7 instead of pin 6.

[mark2]

I tried connecting c3 to pin 7, and it reduced some of the noise on pins 1 and 7, but the signal is still really low at those points, and the output of the pedal is basically silent.

I haven't had time to try pedalpcb's version yet, though.

[j_flanders]

and the output of the pedal is basically silent.

Also make sure R8 is connected to Vbias and not to ground (since you have no DC blocking cap to ground there)

It's the only connection to Vbias that doesn't explicitly say "Vb" on your schematic, so it's an easy mistake to make and connect R8 to ground and end up with no output.

[mark2]

Thanks that's a good tip unfortunately not the problem.   R8 is connected to VB and reading a 4.something V.

[mark2]

As always, the problem is the most boneheaded and repeatedly overlooked one...

I had the switch wired wrong. 

I used the Sallen Key configuration you all recommended, and it sounds really good.  The filtering is really pronounced. Perhaps I'll go back later and try the original schematic and the pedalpcb again to see how they compare, but only out of curiosity.

Now onto the next (and final!) problem.  The octave switch has a barely perceptible impact on the sound, and I don't hear any octave up at all.  I substituted 4148 instead of the recommended 914 diodes.  Any chance that might be the cause? 

Or should I create a new thread for this?

Thanks again everyone.

[mark2]

I reduced the gain and probed to really isolate what's going on and the octave circuit is imparting a bit of distortion, so it's definitely doing something.

The octave up, if it exists, isn't really noticeable to me.  Unfortunately I don't have a scope so I can't watch what's going on.

[mark2]

I fed a pure tone into it from my phone and it's definitely adding octaves, albeit a bit quietly. I guess my ears just don't appreciate octave effects.

My perception of it is that at some frequencies it almost sounds like an octave effect is being removed when that switch is opened.

[mark2]

Here is an updated (unofficial) PDF of the GEO layout and schematic, in case it's helpful to anyone who stumbles on this thread via google: https://drive.google.com/open?id=13T7uq87FSwbJxG5cmtsx8HaVcKiiQiL6

This moves C3 to pin 7.

I also sent it to John Hollis and RG Keen in case they want to update the official pages.