Another strange idea (phaser/flanger/wah)

[Rob Strand]

i saw a thread about LiveSpice that runs in DAW's, is that a recomendable program/plugin?
i do a lot of electronic music production for a friend who's been working on a game for the last 7 years (i doubt he'll ever finish it tho)
i'd love to do some weird ass sounds and effects i have in real life, but not in my DAW.

The goal of LiveSpice is so you can do that *if you want to*.

Circuit simulators are for electrical engineers, some engineers have very little to do with audio!

Honestly,  LTspice is more than good enough for the average pedal developer.    It's entirely stand alone.   As far as LTSpice vs other products, some other products let you do sims a without having to know so much about spice.   However there's a lot of help on the web for LTspice so if you get stuck it's likely you will find a solution on the web.

There is a learning curve associated with using simulators.

[ElectricDruid]

Or is it a religious conflict? (House of Mac?)

No, that's not it. I use LTspice on a Mac. It's just as awkward and "quirky" as on any other platform, I'm sure. Still, does the job. For nothing, I can't argue.
The actual underlying software is good. It's just the interface that's dreadful. Which is a shame, since that's the part you use!!

[iainpunk]

There is a learning curve associated with using simulators.

i know, i used to use MultiSim a lot for school, but my new school doesn't include licence for any simulation software, so i used falstad a lot for pedal building/designing.

cheers, Iain

[Rob Strand]

The actual underlying software is good. It's just the interface that's dreadful. Which is a shame, since that's the part you use!!

I can see why they simplified the interface.   You don't want to spend you whole time programming the user interface but you do want the the simulation engine to work.   However it does have quite a few quirks.   If they fixed some of the quirks life would be a lot easier.

I've reprogrammed a lot of the hot keys to be similar to what I've used on Pspice.  It's a big help, but probably only helpful to me.

I guess my biggest beef are some "hidden" behaviour/features of the user interface, like pressing control to edit fields.    Copying from one sheet to another is a tricky one!   It's virtually impossible to get access to those features through normal use - we shouldn't have that type of stuff in modern interfaces.

[Rob Strand]

Here's a crude summary of the all-passes in this thread.




It's not a complete catalog of all all-pass circuits.    There's other circuits and ways of making small changes to the circuits I've mentioned as well.

The only thing I added was the "other type" of bridge-T based circuit which has equal resistor values instead of equal cap values.
It's possible to do the RC exchange on the other circuits as well.

[PRR]

....i used to use MultiSim a lot for school, but ....

Any relation to Multisim Live?
(I have not tried the off-line Multisim.)

Live seems to be free, but wants a decent web connection, and like all online tools there is some question of intellectual property leakage. (Who the hell would trawl through all the drek attempted in online simulators?)

AH! The Live version is limited to 25 components. Or maybe 78. Their "comparison" is confusing.
https://www.multisim.com/pricing/

[iainpunk]

....i used to use MultiSim a lot for school, but ....

Any relation to Multisim Live?
(I have not tried the off-line Multisim.)

Live seems to be free, but wants a decent web connection, and like all online tools there is some question of intellectual property leakage. (Who the hell would trawl through all the drek attempted in online simulators?)

AH! The Live version is limited to 25 components. Or maybe 78. Their "comparison" is confusing.
https://www.multisim.com/pricing/

that ''multisim Live'' has little to do with the real, full multisim and is a complete dumpsterfire.
yes, 25 components limit, it crashes for like no reason, and the UI is totally different compared to real MultiSim.

i remember the day i got the multi sim licence from school, that really kick started my pedal building in the early days. especially since there are a huge heap of common components in the library already and you don't have to manually add all of them.

i remember designing my first fuzz. it used 4.5V, a single 741 a diode and some resistors and capacitors. i believe i still have that perfboard (cannibalized) stuck to the wall behind my desk.

cheers, Iain

[ElectricDruid]

Here's a crude summary of the all-passes in this thread.



It's not a complete catalog of all all-pass circuits.    There's other circuits and ways of making small changes to the circuits I've mentioned as well.

The only thing I added was the "other type" of bridge-T based circuit which has equal resistor values instead of equal cap values.
It's possible to do the RC exchange on the other circuits as well.

Thanks for this, Rob.

Here's another example, a lowpass+dry signal to give all pass. You can do the same thing with highpass filters too.

https://www.njohnson.co.uk/index.php?menu=2&submenu=0&subsubmenu=3
https://www.njohnson.co.uk/pdf/cesyg/vcapf-sch.pdf

This is a pretty typical design for a synth filter, so deriving an allpass from it is a logical next step. The 2164 quad VCA will run from +/-4V, so you can build this on a 9V supply, just. It's a lot of parts for a single-pole stage though - one VCA and two op-amps per stage.

[Rob Strand]

Here's another example, a lowpass+dry signal to give all pass. You can do the same thing with highpass filters too.

I haven't yet gone over the circuit but from what I can see the filter Q varies with frequency -   maybe you know off hand?    It's second order because of the two caps but only one variable element (the VCA).

It does't bring up the idea of having the Q vary with frequency for a 2nd order all-pass.    The non swept 1st-order all-passes on the MXR Phase 100 unit does that but it's not the same as a Q > 0.5 second order doing it.

[garcho]

^ Noodling with multiple-feedback op amp filters made me wonder whether there might be a "musical" use of q and cutoff frequency changing in sync somehow. For instance, maybe in a wah pedal, one might prefer a lower q sweeping the higher and lower end and a higher q around 1kHz. Almost like taper for potentiometers. I can't imagine "dialing" that in with analog electronics; I wonder if you'd have to just test (topologies, components, etc.) and make do, you know, get creative. Might be a good experiment for digital, though.

[Rob Strand]

I haven't yet gone over the circuit but from what I can see the filter Q varies with frequency -   maybe you know off hand?    It's second order because of the two caps but only one variable element (the VCA).

OK after looking up the datasheet for the SSM2164 it's obvious the all-pass sections are only first order and the RC network is something recommended in the datasheet.  The SSM2165 current input nullifies the effect of the RC network, since it is a virtual ground.

So in short it's more of a conventional first-order all-pass design using an OTA/VCA.

^ Noodling with multiple-feedback op amp filters made me wonder whether there might be a "musical" use of q and cutoff frequency changing in sync somehow. For instance, maybe in a wah pedal, one might prefer a lower q sweeping the higher and lower end and a higher q around 1kHz. Almost like taper for potentiometers. I can't imagine "dialing" that in with analog electronics; I wonder if you'd have to just test (topologies, components, etc.) and make do, you know, get creative. Might be a good experiment for digital, though.

Actually the standard wah circuit has variable Q.    The Q is higher (narrow bandwidth) for low frequencies.   Since there's more energy in the guitar signal down low the variable Q still sounds even-ish in volume.   The circuits which do the opposite change in Q,  like some of the auto-wha, get very exaggerated at the top of the sweep.

[11-90-an]

Thanks a lot, Rob... Your research is more in-depth and more valuable than anything I could do in my lifetime...

I do, though, have more questions...
How would one dial the feedback in "correctly"? Trimpot, of course, but... just enough to stop oscillation? or.. Is there a way of knowing it without relying on ear?

And inverting the feedback... something like a simple inverting amplifier like this would work, right?



And, as if, I have not asked enough questions already, from your sim in reply #3, the resistors to ground... should they be to ground ground? or a virtual ground, perhaps a 4.5v VREF?

[Rob Strand]

How would one dial the feedback in "correctly"? Trimpot, of course, but... just enough to stop oscillation? or.. Is there a way of knowing it without relying on ear?

What would work when you want the highest peaks.

The other way, to to adjust the feedback for a given peak height, for example +20dB.   In practice you would probably need to freeze the LFO to do that.

And inverting the feedback... something like a simple inverting amplifier like this would work, right?

Yes that type of thing.   

In my reply #3 example I had non-inverting mixers at the output and for the feedback (identical circuits).   However in order to get the four peaks  there was also an inverting gain stage "K" which inverted the feedback.   If you follow the dry signal path you will see it is non-inverting.

It's possible to save an opamp by removing the gain-stage "K" and use inverting mixer for the feedback.   However, this also inverts the signal and to compensate for that the phase-shift part must connect to an inverting input of the output mixer.   Like this,



There's no "right" way to do it, just different ways of achieving the same result.

And, as if, I have not asked enough questions already, from your sim in reply #3, the resistors to ground... should they be to ground ground? or a virtual ground, perhaps a 4.5v VREF?

Yes, for an inverting stage you need to connect the non-inverting opamp input to Vref.

For the output mixer Reply #3  at least one of the resistors going to the opamp's non-inverting input need to connect to Vref, or another opamp which is biased to Vref.

So for the circuit I just posted, R26 would connect to Vref.   R33 could be AC coupled to the previous stage by a coupling cap, or, it can directly wired to the previous stage without a couple cap.   It's easier just to wire it without the coupling cap.

[Rob Strand]

FWIW, for completeness here's the two peak version of the 4 stage phaser.    The phase of the feedback is negative.

[ElectricDruid]

I haven't yet gone over the circuit but from what I can see the filter Q varies with frequency -   maybe you know off hand?    It's second order because of the two caps but only one variable element (the VCA).

OK after looking up the datasheet for the SSM2164 it's obvious the all-pass sections are only first order and the RC network is something recommended in the datasheet.  The SSM2165 current input nullifies the effect of the RC network, since it is a virtual ground.

So in short it's more of a conventional first-order all-pass design using an OTA/VCA.

Yes, that's it exactly. That 510R/560p are just for the 2164's stability, not a part of the allpass, so it's first-order sections and the Q should be fixed.

[11-90-an]

Thanks again, Rob for sharing your understanding... 

Another question...


The resistors circled in red... they should be connected to Vref, right? not directly to ground? my understanding is that when the resistance goes low (from an extra variable resistor tacked in parallel with each 10k), the op-amp would misbehave (especially TL07X opamps, i've heard) when put too close to the negative power rail? or am I missing something?

[Rob Strand]

The resistors circled in red... they should be connected to Vref, right? not directly to ground? my understanding is that when the resistance goes low (from an extra variable resistor tacked in parallel with each 10k), the op-amp would misbehave (especially TL07X opamps, i've heard) when put too close to the negative power rail? or am I missing something?

Yes correct.

You might want to look-up the Boss PH-1r schematic (4 -stages).  I'm not sure what sign it uses for the feedback.

[11-90-an]

Positive feedback... i think.. original signal gets inverted then feedback is not-inverted...



might there be a "limit" to how many notches there can be in the usable guitar range?

[Rob Strand]

Positive feedback... i think.. original signal gets inverted then feedback is not-inverted...

might there be a "limit" to how many notches there can be in the usable guitar range?

Yes, it looks like positive.

The feedback affects the peaks (which is what your first post was about).    I guess that's the experiment to see if having peaks sounds good.   Feedback phase doesn't affect the notches when the feedback is set to 0.

Interesting Boss use positive feedback instead of negative.   I do remember phase of the feedback sounded better than the other.    I'd have to find and read my old notes.

As for the number of stages.   Without feedback I distinctly remember 8-stages or more starts to sound like a flanger.    I just can't remember how the feedback sounded with with many stages.  Probably like a flanger but I'm just guessing.

Some of the fine details are getting fuzzy these days.