MultiPhase Project

Scott Stites

I've started a page concerning my ongoing multi-stage configurable phase shifter here:

http://mypeoplepc.com/members/scottnoanh/birthofasynth/id18.html

The idea has expanded from a dual 12 stage phaser to a dual 16 stage phaser. Vactrol based (surprised? Very Happy

).

The idea is to have two 16 stage sections that can be operated in parallel or series (for up to 32 stages), with selectable phase and regen points on each section (a la the Moog Stage Phaser, which wasn't a vactrol phaser).

I haven't put up schematics yet, because it's a work in progress. Lately (not up to date on the page) I've arrived at better cap values and parabolic waveshaping for the sweep.

Before I schplit for Illinois, I added a few samples I recorded a couple of weeks ago to the bottom of the page. The samples start out with the dry signal, stereo-phased, then stereo-phased run through the stereo inputs of the Dim C.

http://mypeoplepc.com/members/scottnoanh/birthofasynth/id18.html

Cheerios,
Scott

paul e. · Posted: Tue Mar 14, 2006 9:06 pm Post subject:

those demo's sound great !..nice work

Scott Stites · Posted: Wed Mar 15, 2006 7:26 pm Post subject:

Thanks, Paul!

I hope (after I get home) to post some more varied samples.

Take care,
Scott

elektro80 · Posted: Wed Mar 15, 2006 7:36 pm Post subject:

THis is an interesting project. I just might want to build this one myself.

Scott Stites · Posted: Wed Mar 15, 2006 8:12 pm Post subject:

It's really been a fun project. There are a few simple controls that add quite a bit.

One of these controls is a switch that selects whether the regeneration is inserted into the first stage of the chain or the second stage - in other words, regeneration occurs from an even or odd number of stages. If you look at a lot of phase shifter designs, you'll probably notice quite a few of them use an odd number of stages for regeneration, while others use an even number of stages. Even/Odd number of stages provide distinctly different effects. Juergen Haible explained the mechanics of it to me quite well a little while ago. I'll have to dig up my notes.

Also, simple inversion of the regenerated signal is another way of varying the effect.

Now, with a high number of stages, sometimes inverting the regen into an even number of stages and not-inverting it into an odd number of stages (and vice versa) can sound similar. As an experiement, I recorded a file that had a comparison to the two effects - they sounded similar, but there was a marked difference in certain points of the sweep (lower or higher end). I haven't touched on trying even/odd stages of the unregenerated wet signal.

Add to that the ability to tap regen from before or after the phase tap - this provides a rich source of timbre variance. I really am quite taken with the Compact A arrangement (inverted filtered regen around the first two stages) - I don't think this phase shifter will ever emulate the Compact A exactly, but the arrangement provides a unique timbre - the regen around those first two stages provides a single peak which sweeps up and down with however many stages of phase shift are selected.

The stereo phase taps sound really nice, and the phase shift is not lost in a stereo->mono mix.

My next goal is to see how compression of regen sounds - I've got some THAT Corp. RMS converters and VCA's I want to try out, just to see what gives.

God help me, but I've been flirting with the Dimension D schematics while on this trip.....the Dim C really adds a lot of twist to the sound (I think the cross mix in the Dim C plays as large a part as the Dim C anti-phase modulated delay lines themselves), I'm wondering how a Dim D would sound with it. I've noticed, other than the obvious increase in headroom, the Dim D arrives at mode 4 quite differently than the Dim C.

Cheers,
Scott

Uncle Krunkus · Posted: Fri Mar 17, 2006 4:41 pm Post subject:

It's a credit to you Scott not only that I could be bothered to download a 3Mb file to check out how your phaser is sounding, but that I was quite happy to leave it on repeat, turned up loud, for about an hour, while I woke up and had breakfast this morning. Smile

Scott Stites · Posted: Fri Mar 17, 2006 7:54 pm Post subject:

Thanks for that, Uncle K, I needed it today Very Happy

I'm calibrating these $25,000 com boxes while out on the road. Today, a slight slip of the probe, shorted two pins and let a brief bit of light and a lot of smoke out of one of 'em. Well, that one's going back to the factory.... Embarassed

Cheers,
Scott

toppobrillo · Posted: Sat Mar 18, 2006 1:20 am Post subject: nuthin but a DSP thang

mighty good phaserin ya done. your soundsamples are really good..

dude and i never told you this, but that mutantfilter is DABOMB i might build one, with home-done vactrolss. i been doing sometests on a bunch of assorted CDSs i have. it sure would be handy to have an extra ammeter or ohmmeter around...

32 stages! why, that's enough to do a shepard phasing with 4 stagers! YOWZA!

does buchla now have a barberpolable phaser? is it a DSP/MCU thang?
i think one of them new BOSS pedal phasers has an infinite up/down type of mode [DSP no doubt].

and i might have read someplace that the EH 'flanger hoax' was a phaser modulated with quadrature functions? you ever heard these things?

i actually have done a working design for generating the 8 shepard function pairs with CMOS and a little bit of analog. it is the article of john simonton's that got me thinking about it. my design scheme is similar, just implemented with different parts.. i am kind of stubborn that way.[someone into MCUs could do this so easily]

if i were to actually build one, i would make it voltage tunable, up/downable, and possibly use 2 ssm2164s as on board VCAs.

someday... but i want danged 32 stages of shifting ready to rock in parallel when she's done......

shepard, ho!

Uncle Krunkus · Posted: Sat Mar 18, 2006 4:58 am Post subject:

Yeah,
shit happens! Cool

I remember once I thought I could get away with installing a sound card with the computer powered down but the power still plugged in. Edge of the card bumped up against the bios chip and a whole load of that expensive smoke came out.
"Now remember little gaurdian angel of technical mishaps, if you see me about to do that again you'll give me a kick in the ribs okay?"
I could say that I'd love to build one of those phasers too, but the list is longer than ever, and Bub No.2 (Hanna) is due almost any day now, so to pretend that I'm going to get anywhere near being able to start something else soon would just be wind in the sails.
I should soon be able to post a new movie of the advanced Roboride.
And I have done a lot of work lately on Frankenstein's Oscillator. A running version which includes the eight modulators is not far off.

Scott Stites · Posted: Sun Mar 19, 2006 9:22 am Post subject:

Hey Topp,

Thanks! For the phase shifter, I want to measure my VTL5C3/2's as well just to see if I can sort them as to tolerance. The tolerance can vary quite a bit from device to device, so I'd like to see if I can formulate a test and classify them by how they react to diode current. I figure I'll take three readings (max CV, mid CV and max negative CV) and sort them by the resistance value measured at those points. Then I'm going to measure the caps used and try matching their tolerance inversely to the tolerance of the Vactrols. Not sure if this will make an audible difference, but it will be interesting.

What type of optocouplers are you using right now? I imagine just about any could be adapted to the Mutant Filter.

32 stages is certainly enough to do some serious barberpole phasing! I've never heard a barberpole phaser - that would be interesting to hear. I don't know how Buchla is doing it. I would bet the Boss stuff is using DSP. If I want to do it with this design, I'll have to come up with eight control sections (one for each phase of the Shepard function generator). Right now I only have two control sections in mind - one for each 16 channel section. I'm using an ARP style NPN/PNP expo transistor control circuit. I figure I'll have to match the transistors fairly well between the sections in order to get a good response out of serial chaining. I plan to match Vbe and beta for the transistors in each section.

I haven't posted any samples of it yet, but the day before I left on this trip, I put in a FET distortion circuit to create a hyper-triangular LFO out of a triangle waveform. It certainly added some really nice sweeps to some setups. I like both the normal triangle and hyper-triangular response, so I'll probably make this waveshaping switchable. I've also increased the cap value, which shifted the range down a bit to the lower register. That's the hardest part - trying to figure out what value one wants there. The problem is that there are 16 different caps to change to hear the full impact throughout the range (I have to re-do the breadboard in order to get all 32 stages in - haven't done that yet).

As I understand it, the Flanger Hoax has both a BBD delay and a phase shift section. I really don't know what the signal flow of it is, though.

Uncle K - Congrats on the impending arrival of the new bambino! *That* is the ultimate DIY project. Very Happy

Absolutely that takes top priority. We expect pictures.

Take care,
Scott

toppobrillo · Posted: Sun Mar 19, 2006 9:56 pm Post subject:

im using homemade combos of ldr/leds. testing them in a rubber piece of tube taped up with black tape.. for the testing i am just kind of getting the most useable range plotted, current sucked thru led vs. resistance.

these photocells im using have a really steep log curve, so once i actually epoxy everything together, i will find the best compromise 'tween the 2 [for the 2/3 arrangement]

do you have a capacitor test on your multimeter? i unfortunately do not. i should look around for a method.. theres circuits [like a dome filter] where its important to match 'em up.

*my apologies for this next section of othertopic, but related material.. skip it if you're like "what?" or i guess do a search for "synbal" here if you are curious about it...*

~also, i have decided to make two 3-LDR/1-LED, [a la 'hairball']
controls to at least test my idea of the 'spread' control dynamic of the synbal circuit. which i think [i hope] will be interesting enough to do. my plan is to inversely shift the frequency of the 3 highest and lowest oscillators... a la 'hairball crossfade circuit' with control voltage.

when the top 3 are getting higher, the bottom 3 are going lower, and vice-versa. also, the topmost osc will be affected more than the one below it, and so on... same for the bottom 3, where the lowest in freq. will be affected more than the others. this would be alot easier to diagram than write about!

i think it will definitely be good to have at least 4 different tap-points of this configuration brought out to a mixing section, maybe a matrix style mixer? or a joystick... hmm.. maybe thru simple bandpass filters. wow this just got alot bigger of a thing... sorry to ramble on as much about an entirely different project, ill keep y'all posted~

as for the control section of the phaser adapted for 8 cvs..there are several transistor arrays with common elements... that would be reasonable matching without all the trouble.
[unless you like trouble, heheh]

anyhow, yeah im excited about the filter... the bipolar attenuator is great..
the highpass is like the most exciting highpass i have ever heard. i think it would be awesome to do a single pcb layout that can be made to all 3 filters, what do you think?

josh

Uncle Krunkus · Posted: Thu Mar 23, 2006 6:39 am Post subject:

Hey Scott,
here's a barrage of questions for ya! Surprised

Is the multi-phase project an extension of the mutant vactrol filter? Cos I was wondering, if I were to do a stripboard layout of the mutant, whether it would be possible to make them chainable to build a multi-phase etc.
Also, I noticed that your schems which include the VTL5C2/3 show the LED as two LEDs in series. Is this what's actually inside? What's the drop acros the LED side?
And I also noticed that the LDR (even on the datasheet) is drawn as two LDRs wired as a pot. They both have the same co-efficient don't they? (ie both are higher resistance in the dark) Are the 3K-10M kind of values you spoke of for each half, or both together?

Scott Stites · Posted: Thu Mar 23, 2006 9:08 am Post subject:

Hey Uncle K,

Scott Stites · Posted: Fri Mar 24, 2006 7:40 am Post subject:

Hey Uncle K:

LED voltage drop of the VTL5C3/2 is 1.51V.

I've attached a quick sample of the MultiPhase configured for four stages of phase shift - this can be accomplished with one TL074 and two VTL5C3/2's if one ever wanted to build a four stage phase shifter. The values used are those in the schematic of my previous post above.

Anyway, the sample consists of four segments:

00:00 Dry signal (source is the DW6000)
00:12 Four stages of phase shift, no resonance
00:25 Four stages of phase shift, four stages of resonance
00:37 Four stages of phase shift, three stages of resonance (resonance is injected into stage 2)

This was recorded straight with no other effects - just the MultiPhase. Modulation on this is a triangle wave converted to a hypertriangular waveform through the FET distortion circuit.

Cheers,
Scott

Uncle Krunkus · Posted: Sat Mar 25, 2006 2:21 am Post subject:

What do the schematics for the resonance and modulation look like?
I've almost finished my first two hand made VTLs. I'm going to put two on a 14DIL package (isolated from each other) so I can do layouts based on a standard design.
BTW have you ever tried modulating the LEDS with different mod sources, I'm thinking 3 simple LFOs would be enough to create a complex mod signal. Then you could patch which LFO was modding which stage etc. This is gonna be fun! Cool

blue hell · Posted: Sat Mar 25, 2006 5:46 am Post subject:

Uncle Krunkus · Posted: Sat Mar 25, 2006 6:08 am Post subject:

I've just tested my first home made Vactrol.
LED Current-------0mA-----.03mA-----50mA
LDR Resistance----3M--------60K-------500R
This was just with a black pen cap over it.
I think this seems really good from what you were saying before Scott. What do you think?

bugbrand · Posted: Sat Mar 25, 2006 7:15 am Post subject:

Scott Stites · Posted: Sat Mar 25, 2006 8:24 am Post subject:

Modulation is a very important aspect of this project, and I haven't really discussed it any at all. The goal is to have at least two multi-waveform LFO's, with bipolar attenuators selecting the phase and amplitude of modulation for each 16 stage section of the phase shifter. In other words, each section will have a control pot and switch for each LFO source. Centering the control pot will attenuate fully, full CCW will be full inverted modulation, and full CW will be full non-inverted modulation. The switch will select the modulating waveform. In this manner, the two sections can be modulated together or anti-phase to each other.

I'd also like to add an envelope follower, and I'm thinking about putting in a quadrature generator, but that remains to be seen. A 90 degree phase shift in the modulation of both sections I think would provide a unique stereo configuration. Most definitely external modulation is something that is definitely desired. Whether the LFO's themselves will be voltage controlled or not hasn't been decided - the problem is not doing it, but at some point the number of controls is going to hit critical mass.

I'd like to create at least one set of controls that controls both sections simultaneously (it would make setting up sweeps in chain mode easier).

For the phase mixer (dry and wet) I'm using the mixer module on my modular. This is a simple unity gain TL072 based op amp mixer circuit - I just haven't put one on the breadboard. I think the same arrangement, only putting dry on one side of the pot and wet on the other will provide a one-control mixer (CCW full dry, center 50% mix, CW full wet). Right now, with two controls, I just set dry for 50% and wet for 50%. The resonance is buffered by the same kind of unity gain op amp circuit. The inverted output of the resonance is available from the first stage and the non-inverted resonance is available from the second stage - a switch selects between the two. On the inverted output, the signal is running through that low pass section from the Compact A schematic that V-un-V posted a while back. This signal, after the switch, is mixed into either the first stage of the all pass chain or the second stage of the all pass chain, through a 10K resistor. As an aside, I have all the THAT parts here now to try putting a compressor in, and I was thinking about compressing/limiting the res to see what effect that would have. High resonance can provide 'surprises' in amplitude that make getting a good recording level difficult.

LFO is the same deal - straight out of my modular. I have Ray Wilson cool new LFO's that I use. Note that right now I have an amplitude control and a center control controlling the sweep - very important to have both of these, especially when you get a lot of stages going and it begins to sound more flangerish. The cool new LFO's have been modified to provide a very slow sweep (again important for multi-stacked stages). The cap value was upped and made switchable to provide a very slow range.

The LDR values certainly look good to me, though I don't know as the control circuit will suck as much as 50 mA through the LED. What does it give at around 20 mA?

To be honest, I better look at how much current I'm sucking. Fuse is out on my ammeter (a constant condition due to always managing to put plug the leads into the wrong ##$% socket and popping the fuse). But, you always want to make sure your LED can handle the amount of current you're shooting for. My first tech job was repairing communications service monitors - some really cool gear I'm sure Mosc would love. My very first repair was fixing the SINAD/Distortion meter - replacing an inop CLM6000 optocoupler. Two sets later, I performed the same repair on another set. A day or so later, another set. And on and on. It was a common failure. Why? Because whomever had designed the distortion null circuit had a successfully operating circuit that just plain pulled too much current through the optocoupler LED. The guy obviously interpreted 'absolute maximum' as 'normal everyday'. They'd work a couple of years, but eventually that LED would just >pop<. As a result, I like to keep the current in LED's well away from the max. In this case, the current is going through a transistor in a design that has worked for 30 years in ARP synths, so I haven't been too worried about it.

I agree with Bug Brand - 14 DIL is an AWESOME idea. Seems to me the Countryman phase shifter had something like 8 pin DIL way back in the day.

When selecting your LDR's, you might take a look at the response times of the LDR's as well. Motohiko Takeda has a good setup for doing that.

Cheers,
Scott

Scott Stites · Posted: Sat Mar 25, 2006 8:26 am Post subject:

I thought I'd put this in a separate post:

I've just uploaded a composition to my site, the only link to it is this:

http://mypeoplepc.com/members/scottnoanh/sitebuildercontent/sitebuilderfiles/suffocation.mp3

Filesize is 2.65 MB.

I've never publicly posted this because it is a bit political in nature, and in these times I'm a bit paranoid. Words are juxtaposed. Mosc/V-un-V, if this is not appropriate in this forum, please let me know. I'm not here to make 'statements'.

Anyway, I post it now because I feel it's a good example of using anti-phase multi-LFO modulation in a stereo configuration. Boy, that's a mouth-full.

This was recorded back when I was using 2n7 cap values with 1.5M parallel resistors on the optocoupler elements. The musical portion consists of only one stereo track - the DW6000 run through the nascent MultiPhase circuit. Sorry, don't have a dry sample of the DW6000, but it was a fairly plain tone - no modulation. All of the growly/percussive modulation heard is provided by the MultiPhase.

The MultiPhase itself is divided into two 6 stage phasers - one for right and one for left. Each phaser has it's own regen and control circuit. I used two LFO's - a triangle wave and a pulse wave. I mixed the two LFO's together. The non-inverted LFO mix output modulates one channel and the inverted output of the LFO mix modulates the other channel. As the sample starts, only the triangle is modulating the phase shift stages, and then I mix in the pulse wave - this can obviously be heard as a percussive sound envelope. There are no musical overdubs or any other effect other than the MultiPhase used (often times, to me in these circumstances, it sometimes sound like a delay is being used - this is not the case). The vocals were selected, edited and overdubbed. I was in a mood.

Cheers,
Scott

bugbrand · Posted: Sat Mar 25, 2006 9:00 am Post subject:

All your phase sounds are supra-lovely!

Square-wave modulation for rhythmics is something I'm especially into - I may be somewhat wrong, but it doesn't seem to be used ?enough? -- I mean, modulation settings on devices are so often going for the smooth changing tri/sin shape, but try squarewave modulating a VCdelay and whoooah, repitching sounds!

Tieing this in with my control rack -- take a squarewave and divide it into many divisions (..i need replicants of some of the modules - more 2/4/8/etc dividers and more 3/4/5/etc ones too). You can get really wonderful rhythmic variations out of such a setup - keep a regular 4/4 rhythm but with pretty much ever changing emphasis / change. This sounds really great over-driving the CV inputs of filters, delays, surely other modules too that I just haven't built yet (... YES, I do want to make a phaser somepoint!)

So, yes, in essence, squarewave modulation mmmmm!

Scott Stites · Posted: Sat Mar 25, 2006 9:35 am Post subject:

I absolutely agree! Modulation style effects are capable of so much more if you give them the opportunity to be modulated by something other than smooth,cyclical waveforms. Sample and Hold modulation (unslewed) of a delay is really a cool effect, too, BTW.

A few years back, I created this circuit, since dubbed the Klee Sequencer by Romeo Fahl, who built one. 'Klee' is a reference to 'clay' - I'd made the comment that setting up a sequence on the thing was like molding something out of clay.

The circuit really is just a mad combination of Buchla's Quantized Random Voltage function from the 266 Source of Uncertainty and Ken Stone's Gated Comparator. It consists of a voltage input into a comparator and a clock feeding a CD4006 shift register. Four outputs of the shift register feed four pots, and these pots are all summed together to create a stepped voltage. The CD4006 can hold 16 bits, so with those four pots, one can create a 16 step sequence in loop mode.

You feed the comparator with a voltage source - LFO, noise, whatever. Whenever the voltage source is above the reference, the comparator output is 1, if the voltage source is below the reference, the comparator output is 0. Whenever a clock pulse arrives the value at the output of the comparator is shifted into the shift register. So, in this mode, the shift register is always clocking along a random pattern. The comparator reference could be set with a pot, or be an external CV. This function could be described in Buchlanese as a 'Probability of Change' control.

I had a 'loop' switch setup so that when an interesting pattern was within the shift register, I could switch the output of the shift register to feed the input. In this configuration, the same pattern would constantly shift through, creating the 16 step sequence mentioned above.

The four pots could be set up to process their individual bits in whatever increment one wanted. The combination of bits and these pot levels determined the output volatage and pattern. It was an amazingly effective device, though, as mentioned before, setting up the sequence was like molding wet clay into a sculpture.

The shift register outputs, along with creating stepped voltages, could also be used as gates or triggers. This was extremely useful - depending on when a gate fired off (controlling an EG for example), a certain segment of the sequence would be 'illuminated' for want of a better description. This circuit was great for controlling filters, too.

Here's my 'Shift Register Magnum Opus' I recorded back then. The only control source is the Klee sequencer. It contains random and looped sequences, some at low audio rate controlling the VCO's and filters. The pads are the only part that aren't controlled by the Klee - I just layered VCO's to create them.

http://mypeoplepc.com/members/scottnoanh/sitebuildercontent/sitebuilderfiles/final3n.mp3

It's 5.1 MB, but if sequencer/square wave stuff interests you, check it out (it's one of my favorite things I've ever recorded, other than the poor recording, I'm kinda proud of it).

I'm definitely going to make a Klee Sequencer module......

Cheerio,
Scott

bugbrand · Posted: Sat Mar 25, 2006 9:49 am Post subject:

The clay analogy is just spot on! What a wonderful way to work.

And a really great track there too - - its strange how analogue signals have been kind-of 'quantised by western musicality and scales' - I loved the way your sequences would bring in sounds and intervals that were so unexpected but still so right.

Scott, you're a freaking mine of great information!

Hehe - you make me remember that I've got to rehouse the S/H from one of my now defunct TestBoxes to get going with stepped waveforms again.

Supercheers!

blue hell · Posted: Sat Mar 25, 2006 1:05 pm Post subject:

mosc · Posted: Sat Mar 25, 2006 3:08 pm Post subject: