Phase 90 clone: Depth vs. Blend

[woulfer]

I built a phase 90 clone using a BYOC board. It sound great. I did the "Depth Knob" option, and I found it far more subtle than I expected. I thought it would give me enough control to put the effect barely audible to full "phasation".

So, in my search for possible solutions, I found mentioned here a "blend" pot. Would that better fit my needs in this case? Would it blend the straight signal with the effect, bringing me from completely unaffected to full on phase?  It would be the input from the jack to one end, the output from the board to the other, and the output jack to the wiper, corrcect? If so, what value pot should be used?

Any input would be aprreciated.

[Mark Hammer]

These are two ways of achieving subtlety.

A "Depth" control generally refers to an attenuation of the LFO output which forces the sweep of the control elements (in this case FETs).  Typically, in modulated effects less depth/sweep makes for a subtler effect and is often used to offset changes to the speed (faster speeds sound better with less sweep depth).  Normally, this attenuation simply trims back on divergences from some standard starting point.  Ideally, what one wants is variable offset or initial, such that the sweep can be set to start from such and such a point and go a little bit higher or a lot higher than that start-point to achieve different tones/feels.  The "bias" contro can be used to suib for such an offset/initial control to a limited degree.

In contrast, "blend" refers to the combination of two signals - wet and dry.  In the case of phasers, having much more dry than wet reduces the apparent intensity of the effect, regardless of the speed, regen, or sweep depth/width.  Having much more wet than dry, however, results in a vibrato or pitch-warbling effect.

Personally I find that variations in trimming back the wet signal from a 50/50 blend are helpful but if I want toshift the balance in favour of wet, I'd rather just use a toggle to cancel the dry signal.  60%-dry and 40%-wet is useful, as is 70% dry and 30% wet.  Anything between 0% and 50% dry does nothing AFAIC.  The dry and wet are balanced to a 50/50 blend using a pair of 150k resistors.  Being able to trim back the wet to achieve a broad range of subtler sounds might be achieved by replacing the 150k on the wet side with a 120k and 1meg pot in series.  This will let you dial in a *perfect* 50/50 blend (just because the two resistors *say* 150k doesn't make them identical or balance off levels flawlessly) as well as scaling back the wet signal considerably.

[woulfer]

Wow, thanks Mark. I always come away from reading your posts a little smarter than before.

I will play with your suggestions and report back with my progress. I'm sure I'll end up stumped on something. 

[woulfer]

So many options. I may try this, again posted by Mr. Hammer.

A second, that Tele_guitarist noted, is to vary the amount of regeneration.  Many phasers have a variable regen control, but the P90 did not.  The amount of regeneration, and subsequent emphasis of the peaks and notches, is governed by the value of the famed resistor R28.  In the Tonepad schematic this is the 22k resistor in red that straddles the 4th and 2nd phase-shift stages.  The lower the value, the more signal is fed back earlier in the signal path, and the greater the emphasis of the notches and peaks.  It is extremely unwise to have a resistor value LESS than the value of the resistor found in either the feedback loop or input of the phase-shift stages, since this will introduce oscillation.  However, a wide range of values greater than that can produce useful variations in tone.  T'wer I, swap the 22k fixed resistor for a 12k or 15k fixed resistor and a 50k pot wired in series as a variable resistor.  This will give a feedback resistance of 12k (or 15k) to roughly 60k (those 50k pots are really about 47k), which should provide a nice range of near-piercing resonance to imperceptible regeneration/resonance.  Probably a good idea to stick a cap in series with that path as well.  The cap would go between the output of the last phase-shift stage and the pot/resistor.  To restrict feedback at lower frequencies, a good value might be .047uf, though you can consider moving up to .1uf or even down to .033uf.

[Mark Hammer]

Note that some folks are simply content with two regen/resonance settings.  For instance the little magic button on the current EVH P90 phaser switches between two different regen settings (among other things).  Those folks who hav only a little space to work with may find that using a simple on-off-on SPDT minitoggle to get 3 different regen settings (none, a bit, lots) may be as much variation as they need.

[woulfer]

Now that sounds right up my alley.

Can you tell me how the resistor value affects the regens, meaning does the low value increase or decrease the regenerations? And if you'd be so kind, could you suggest a min and max value that I probably wouldn't want to exceed?

Thanks again.

[Mark Hammer]

Stock is something like 22k or thereabouts (I lose track of which issue has what).  Open circuit (or anything higher than 150k or so, really) gets you no audible regen.  A value of 10k will get you oscillation so you don't want to go that low.  Maybe 15k would be a decent place to aim for to get a high regen sound. A value of 27k would be less regen than stock.

So, looking at the Tonepad layout ( http://www.tonepad.com/getFile.asp?id=42 ), you would want to wire up the common/centre lug of the 3-way toggle to where the regen signal returns on IC2b (pin 6 or its equivalent on your pedal).  The existing feedback resistor is now disconneted from that point and tied to one of the outside lugs of the toggle.  A lower-value resistor is then installed connecting the other outside toggle lug and the far side of the stock resistor.  In the middle position the feedback connection is broken.  To one side you have medium feedback (which could be stock setting but doesn't need to be), and to the other side you have intense/high feedback with the lower value resistor being connected.

Again, you can use whatever parts values you want, simply keeping in mind that the closer to unity (in this case, 10k) the feedback path is, the greater the likelihood of uncontrolled oscillation, and the farther from unity (values bigger than 47k or so) the subtler and inaudible any feedback will be.

As I've described it, the "middle" switch position is no feedback.  You may well decide that you like at least a little, in which case the existing 22k resistor is swapper for something like a 47k, and the toggle now adds other resistor values in parallel.  A second 47k unit switched in parallel gets you the equivalent of 23.5k, and a 33k unit switched in parallel with 47k gets you the equivalent of 19.4k, which should provide audible differences.

Finally, sometimes it is helpul to restrict the bandwidth of the feedback signal such that there is more regen for higher frequency content than for lower.  This can help reduce the annoying metallic or boxey quality of higher regen settings.  As an example, using an 18k resistor in series with a .01uf cap would get you higher regen, but attenuated for content below around 880hz.