Modular-izing the Phase 90

[Taylor]

I am building something like a modular synth, but intended to process signals from guitar/bass. The idea is to have the complex and open-ended capabilities of a modular, but using circuits that run on 9v and accept instrument level signals.

I am building a couple of Phase 90-based circuits - using some of the ideas from R.G.'s Phase 180+. I plan to have selectable 4/8 stages (possibly 12 if I can find enough matching FETs), and the controls from the Phase 180+. What I'm curious about is what I could do to make the circuit more modular.

Would it make for interesting results if I added an input and output for each stage, so that I can patch together different feedback paths? What other patch points might give me interesting sounds? I plan to allow at least some of the parameters to be controlled from outside control sources.

[Mark Hammer]

One of the nice thngs about having a couple of phase shifters is that you can experiment with sticking them both before and after certain other effects.  And you are correct in assuming that dual positioning can yield some value over and above just having one phaser with more stages.

This can happen in several ways.  First, you can get some neat sounds by sweeping one phaser slowly (low to medium feedback) and a second phaser faster (with lots of feedback).  Second, phasers sound different when before a distortion than when after, and different again when placed in both locations and not synchronized.  Third, having two unsynchronized phasers in stereo creates interesting patterns of spatial movement.

If onewishes to have multiple modules, the matching does not have to be across all stages.  Matching could be within these 4 stages and those 4 stages.  In some respects, use of photocells or OTAs makes this easier since they do not need to be matched the same way as JFETs.  You might want to consider the standalone 4-stage module that Francisco Pena has at Tonepad.

[Taylor]

In addition to having 2 separate phasers with their own LFOs, I'm also talking about having the phase shift stages be patchable to each other, by which I mean:

Say we have a 4 stage phaser. Usually the feedback links the ouput of the 4th stage to the input of the 1st stage. What if I could patch feedback from the OP of the 3rd stage to the IP of the second for example? Would this create a different sound than having the feedback go through all 4 stages? Or would this not be anything too special? What about having an inverter in the feedback loop?

[notchboy]

Doepfer has a patchable phaser that's similar to what you're describing, plus support for both positive and negative regen:

http://www.doepfer.de/a100_man/a1013_man.pdf

IIRC there are mp3s but I don't have a link handy.

I think the main thing you need to include is the ability to independently select the number of phasing stages vs the number of regen stages, like the Moog rack phaser.  The next thing would be to have two separate outputs, where each mixes the dry signal with the output from a selectable number of stages, or just outputs the wet signal.  You can get some cool panning and imaging effects by having four stages on the left output and six on the right, etc.

Something that I intend to try if my f*%$#@! Ropez boards ever arrive [cough] is to allow cross-regen between the channels of a stereo phaser, something that I've only heard on digital phasers.

[Taylor]

Using the Doepfer as a model, could I just have an in put and output for each shift stage, then, if I want 12 stages, I just use the last output, or if I want 5 stages, I use the 5 stage output? And if I want to feed stage 10 into stage 6, can I just patch that directly? Will there be any issues with loading or anything? I don't want to have to have a buffer for each stage; that would really complicate this design.

As an aside, it is quite easy to make a 12-stage Phase 90. I'm sure this doesn't compare to the 12-stage Moog, since the topology is completely different, but for someone wanting to make a many-stage phaser and feeling daunted by looking at the Moog, I can suggest this. I was able to find a pretty good set of 12 matching FETs in my pack of 100.

Has anyone thought of using a transistor array for this? Seems like that would make building a FET matcher unnecessary. Is there some reason this wouldn't work? The arrays seem to be around $3-5, which, though pricy for an IC, is cheaper than 100 FETs, and easier.

[Mark Hammer]

Unless you are planning to feed it with complex wide-bandwidth signals like synths, drumkits or mixdowns, a 12-stage unit will likely disappoint.  The problem is that a guitar, even when fuzzed, does not possess quite the same bandwidth or spectrum coverage as either a multi-source mixdown or white noise.  So, yo tend to only get to hear a couple of notches, since the remainder are either out of the spectrum covered by the instrument, or else are so much less evident that you ca't really hear them.

[Taylor]

Yeah, I've heard 12 stages don't really add much over 8, but I went ahead and did 12 just to see. It's shaping up to be more than I was planning: right now I'm fitting jacks to have I/O for each stage pair, and I'll have a patchable attenuation knob with an inverter for pos/neg feedback.

The idea with modular stuff for me is that I can't always anticipate what I will want to be able to do in the future. I honestly don't know how useful this setup will be, but why not?

[Taylor]

To go further with this modular idea, I'd like to be able to control the phase shift with an external controller rather than the LFO. Would I just need to inject the control source into the point where the 3.3m resistor connects to the gates of all the JFETs?