Vibing a Phase 90?

[kaycee]

Hi,

having had a go at the vibe mods to the Phase 45 circuit and enjoyed the results I'd like to try the same idea with the Phase 90. I'm looking at Harald Sabro's stripboard layout:

http://www.sabrotone.com/wp-content/uploads/2010/08/MXR-Phase-90.gif

And the source schematic at Tonepad:

http://tonepad.com/getFile.asp?id=42

Unfortunately the schematic doesn't relate numbering wise to the layout. I know that I need to change the values of four caps from 50nf to 10nf and have tentatively identified C2,4,5 and 7 as the ones I need to switch, am I right?

Secondly, have I got the values right for vibing? Should they all be the same value, or would it be better to have different values at each stage?

Thanks for any help folks.

[DougH]

Find a vibe schematic and you can verify the cap values there. Yes, they are all different.

FWIW, I tried this when I built my Phase90 and had it on the breadboard and IMO it wasn't very impressive. There's much more to the vibe sound than the tuning of the filters. I think the optical aspect has a lot to do with it. A Phase100 is an optical phase shifter that can produce some very "vibey" qualities for example, and it's filters are all tuned the same, a la Phase90, small stone and other typical phase shifters.

IMO give it a try, but don't make a permanent commitment to it until you hear it first (breadboard or otherwise prototype it first).

[Mark Hammer]

The optical aspect plays a role in the "shape" of the modulation and the feel that imparts.  The filtering aspect involves the production of two wide and shallow dips in the spectrum, rather than tightly focussed "notches".  As with many effects, what is critical is where the user/listener's attention is directed by those differences.  In the case of a vibe, the use of broad shallow dips imparts a degree of subtle animation, rather than filtering.

I would imagine that where the dips occur is also critical to what one's attention is shifted to.  Certainly with phasers, when a notch or two moves through the region where the note fundamentals lives, the effect created is that the instrument being phased has somehow "moved away" from the listener, and then returned after the notch moves somewhere else.  Vibe units never do that.

[DougH]

As with many effects, what is critical is where the user/listener's attention is directed by those differences. 

Well said, Mark.

I will also add that although there is a setting on the Phase100 rotary switch that sounds pretty "vibey" to me, that may in fact have nothing to do with the optics. It may have more to do with the way the feedback and other circuit parameters are set for that switch position. In any case it sounds very "liquid" and "smooth" in that setting, similar to my vibe in some ways, but more like a "typical" phase shifter in others.

[kaycee]

Thanks for the comments so far.   :icon_lol:Yes, I've built enough disapointments to know not to start designing the label and painting the box just yet... I'll socket the 'vibe' set of caps and see how it goes. I know that it won't be a univibe as such, but the vibed P45 was really nice, not quite as good as the magnatone, but still nice. That uses two caps the same value for the vibe section, but I saw discussion elsewhere about cascading the values and it got me a wonderin'

Anyhow, can anyone confirm that I've got the right caps in my sights?

[Mark Hammer]

As with many effects, what is critical is where the user/listener's attention is directed by those differences. 

Well said, Mark.

I will also add that although there is a setting on the Phase100 rotary switch that sounds pretty "vibey" to me, that may in fact have nothing to do with the optics. It may have more to do with the way the feedback and other circuit parameters are set for that switch position. In any case it sounds very "liquid" and "smooth" in that setting, similar to my vibe in some ways, but more like a "typical" phase shifter in others.

A P90 clone I made, with a set of moderately-matched 5952s, includes a '"range" control (i.e., replacing the 1M fixed resistor off the bias trimpot with a 330k and a 1M pot).  Adjusting the current fed to the FET gates so that the sweep is more gurgly, and reducing the sweep width, starts to yield a more vibey sound.  Stick it ahead of a mild overdrive, and Trower-tone starts to become more easily realized.

[jdub]

Anyhow, can anyone confirm that I've got the right caps in my sights?

Unfortunately (and annoyingly) I can't open up the Tonepad site from work, but if you're referring to the .05uF caps connected to the non-inverting inputs of IC1b, IC2 a & b and IC3a, in parallel with the 10k resistors, then those are the ones.  If you're interested, I did a perf layout for a Phase 90 that shows the caps mounted on a 4PDT (http://www.aronnelson.com/gallery/main.php/v/jdub/Phase+90+perf.PNG.html.  Doug and Mark are right, though- after all that effort, it wasn't that great.  It does get across some of the throb of a 'Vibe, though.   

[kaycee]

Mmmm...definately getting the feeling that I'm pressing ahead with this against the advice of those who've been there, done that and didn't bother buying the T-shirt

I don't know how to do the multiple quotes in a reply, so Mark, yep, got that mod, the 'depth' control, easy enought to hack that in. I'm not going to bother with the 'resonance control' I'll stick with the Block/script switch already on there.

J, yes I've got your perf layout in my files already, thanks for sharing it .

I know which caps on the schematic, its finding them on the vero - I'm sure of three, just the last one has me a bit confused, but theres only one other one it could be so easy enough to stick another set of flying leads for testing. Not sure I understand how you've wired the 4PDT though? I was going to take a line from the board to the pole, fix one leg of each cap to the lugs, join the other legs and run back to the board.

I'd just like to add a bit of 'value' to a self build, if I wanted a straight P90 I can pick up a beringer or joyo for under £40 and be done with it, but wheres the fun in that

[Eb7+9]

start with the values used in the Small-Stone mods // the original "Univibed" project ...
go from there ... the key is to have two groups, each pair a decade in size from each other ...

http://www.lynx.bc.ca/~jc/moddedSmlStnSchm.gif

[DougH]

As with many effects, what is critical is where the user/listener's attention is directed by those differences. 

Well said, Mark.

I will also add that although there is a setting on the Phase100 rotary switch that sounds pretty "vibey" to me, that may in fact have nothing to do with the optics. It may have more to do with the way the feedback and other circuit parameters are set for that switch position. In any case it sounds very "liquid" and "smooth" in that setting, similar to my vibe in some ways, but more like a "typical" phase shifter in others.

A P90 clone I made, with a set of moderately-matched 5952s, includes a '"range" control (i.e., replacing the 1M fixed resistor off the bias trimpot with a 330k and a 1M pot).  Adjusting the current fed to the FET gates so that the sweep is more gurgly, and reducing the sweep width, starts to yield a more vibey sound.  Stick it ahead of a mild overdrive, and Trower-tone starts to become more easily realized.

I've actually gotten somewhat "Trower-y" effects with my standard Phase 90. There's a little wobble in the LFO already, the rest is getting the speed right. I discovered that speed control is very powerful. So much of my perception of "how something sounds" has more to do with the speed than I realized.

The ultimate is to use the leslie, then depending on how you set the amp, guitar, and speed, you can hear all the modulation effects - tremolo, phasing, chorus, flanging, vibe. Different settings of those variables accentuate different effects. All these are just different aspects of the leslie sound.

[kaycee]

start with the values used in the Small-Stone mods // the original "Univibed" project ...
go from there ... the key is to have two groups, each pair a decade in size from each other ...

Yep! Nice one. Now we're getting there. confirm my selection of caps to change, with these values: 150nf, 220nf, 4n7 and 470pf. Actually, there are a lot of possibilities here, having all the filter caps at 10nf was worthwhile too..

I've got the depth/intensity wired in, take out R17, series resistor of @470k to 1M pot, bridge an outer lug. Seems to work the same both ways round, not sure that I understand that?? I'm thinking maybe another switch to kill the dry signal for vibe mode. Once I'm done with it I'll stick up a drawing of the layout with the hacks for anyone interested.

[Mark Hammer]

Take into account that, since vibe units produces shallow broad dips rather than focussed notches, the effect is very hard to detect at the sorts of slower speeds desirable for phasers.  You may not find anything of any interest at LFO rates slower than 1hz.  As such, a Speed pot of 100kC  or even 50KC is recommended for dialability of useful sounds (fastest speeds being in the low-resistance range).

[Eb7+9]

Yep! Nice one. Now we're getting there. confirm my selection of caps to change, with these values: 150nf, 220nf, 4n7 and 470pf. Actually, there are a lot of possibilities here, having all the filter caps at 10nf was worthwhile too..

I think you meant 15nF, 220nF, 4n7 and 470p


I won't provide a complete solution for you but I'll tell you this is a problem for Spice ... in a Univibe each stage causes an amplitude and phase variation // if we base our math on the phase response 90degree cross-over point (and notice how far it spans in both directions with varying resistance) we can get an idea of how to "model" the Univibe response first of all ... let's assume that Univibe cells traverse a 5k to 500k cell resistance range // furthermore let's assume that 50k represents a fair middle-point in the response (considering that cell-to-light transfer evolves in a logarithmic manner) ... run your spice sims with that 50k cell resistance in there and see where the 90degree cross-over points lie for eqach cap value ...

now take the p90 and consider that the jFETs are paralleled by 24k resistors, and consider that jFET's don't go below about 100 ohms in resistance ... we can roughly assume that the equivalent resistance seen across the jFET's would be something like 240 ohms to 24k ohms ... again jFET's have an anti-reciprocal transfer profile that can be said to be roughly logarithmic ... and so, let's choose 2k4 as a rough mid-point ...

yes, it's cavalier thinking at its best ...

now, run a spice sim on a single op-amp phasor stage using that 2k4 equivalent resistance and see where the 90degree crossover point lies ... set each capacitor value so that 90degree crossover point matches that of each corresponding Univibe stage

we don't want to cloud the issue with the notch response of a full circuit with dry/wet mixed together and the like as we already know these are going to be different by virtue of the circuit architecture differences ... what we need is a basic argument of comparison // and the 90degree shift point of each (open loop) phasor stage is the one to use here ...

IMO this is how you should proceed to extract your vibed-P90 caps values // of course, you can always hunt around in pairs to get something pleasing that's not a full-on math emulation ...

[kaycee]

   Flippin' Heck! You've confused me with someone who has a clue as to what they are doing! The only Spice I have is some out of date nutmeg.....

Right, so your saying to get a real univibe-ish sweep I need to tinker with the 24k resistor on each stage. Bear in mind that at this point I'm pretty chuffed that a) it wobbles at all, b) I throw a switch and it wobbles differently   Is this something that I could set by ear via trimmers?

Don't think I'm not grateful for you taking the time to give a thoughtful and informative answer, I am and I'm sure others reading will pick things up from your ideas, and also note I'm not asking for you to provide the answers either, I'm all for 'fishing myself' as well, but its a bit beyond my knowledge to go that deep at the moment - plus it would take the switching into another realm for me as well, something other than mechanical...But like I said, many thanks

Mark, yes I've been messing with the rate pot values for this anyway, I don't like the way the C500K works on these as is and have subbed it for 20/100k B already at the breadboard stage.

[Nocaster Cat]

I may be way off base but I believe the Sweetsound MoFaux is a vibe'd Phase 90. I traced one out awhile back but have no idea where I put the drawing. I'll keep digging though.

[kaycee]

A short update on this one.

Got the phaser/vibe switch working, added a depth control, added a mix control - working on a couple of other little issues such as volume drop and the in/out cap values. All hacked into Sabro's layout, will post the mods when I have it boxed up and tested out.

[Eb7+9]

I haven't done this very sim myself, that's why I was giving you the method ...

you could always try plugging in the Rotovibe values ...

330pF, 3300pF, 0.0015uF, 0.015uF ... (the ordering doesn't matter in a cascade of op-amp based all-pass circuits, eg., P90)

[kaycee]

Still playing around with this on and off. I've noticed that on Haralds vero the input through C1 enters the opamp via pin 3, the inverting (+1 in) input. On another similar phaser that I'm copping some ideas off of the input is via pin 2 non-inverting (-1 in). I know that Pins 1 & 5 are interchangable as long as you pair with their respective outs, but does inverting and non-inverting make a difference?

[Mark Hammer]

The phase of the input stage should make little audible difference in the sound of the unit on its own.

Incidentally, I forget if I mentioned it earlier, but you will never see a feedback path or control on a vibe unit.  The reason is that feedback in a phaser more tightly focusses the nothes and peaks, but in a vibe you want them to be as diffuse and shallow as possible.  So, kill the feedback path entirely.

[kaycee]

Thanks Mark, so the feedback route is the 'resonance' control?

http://www.sabrotone.com/wp-content/uploads/2010/08/MXR-Phase-90.gif

On Harald's vero the output at pin 7 of IC2 routed through the switch via R18 to pin 6 of IC1 (the Block/script switch)? This kills the path or runs it through the 22K R18.