4th order delay as an alternative to equal-staged phasers!

[Peter Snowberg]

Switch selectable. :mrgreen:

[puretube]

information lost...

[StephenGiles]

When I am able to post the original Eventide Instant Phaser circuit, I'd be greatful if a fresh pair of eyes would check my redraw. I'm pretty certain it's OK but you never know!
Stephen

[Vsat]

Here's a couple of phaser response plots...

http://img.photobucket.com/albums/v484/icmax/equi-stag.gif

The upper plot is the response of a 16-stage first order phaser with identical sections.

The lower plot is a 16-stage phaser using four identical cascaded  "univibe" style allpass sections (each section consists of four staggered allpass stages, as in univibe).
Regards, Mike

[puretube]

information lost...

[puretube]

information lost...

[Vsat]

Puretube (and list),
Here's the same data as before but presented with linear (Hz) scale (OK, not quite the same data - to make the details more visible chopped off everything above 10 KHz):

http://img.photobucket.com/albums/v484/icmax/equi-staglin.gif

The Hartmann paper shows mathematically that using a bunch of identical stages gives a more flanger-like response - these plots using just 16 stages give a glimpse of this phenomenon. Use of 2nd order stages is not covered in the paper, although briefly mentioned.

Enough graphing for today!
Regards, Mike

[stm]

I had curiosity and plotted the amplitude response of my previous 0.6 msec delay composed of two 7th order stages (14 stages in total), and an ideal 0.6 msec delay.

Both delayed signals are combined with the undelayed signal.  As it can be seen, the allpass delay follows perfectly the notches up to 8 kHz (in blue), whereas the ideal delay departs exhibiting infinite notches after (in red).




Many years ago I tried the phaser approach using PWM modulated resistors as in the eventide instant phaser. The problem I encountered at that time was noise feedthrough (a kind of white noise appeared), and also I wasn't able to change radically the PWM duty cycle to achieve a wide range (I moved between 10% and 90% or so).

Maybe this is an accurate way to have resistors track in the 4th order delay circuit.  It could even be extended to the 0.6 msec allpass, provided the resistor in parallel with the analog switches is proportional to the resistor that's being modulated, such as 100 times its values.

Take care.

STM

[Vsat]

stm - VERY neat plot!
Regards, Mike

[Peter Snowberg]

This gets my vote for coolest thread in a long time!

Nice plot stm!

[Boofhead]

For the unwary:  some high order circuits, like the one shown,  can be very noisy (even though they use less opamps).  Look at the output stage.  It has a gain of 8 or so. Roughly speaking that means the noise of the three opamps is multiplied by about 8 from the start.  The noise of the first order circuits isn't too bad.

[stm]

As puretube requested, I am including PHASE plots for the three 4th order phasers.

Blue: 4th order delay (stm)
Red: Equal capactors (small stone)
Green: Staggered capacitors (univibe)

This is the LOG frequency plot:



And this is the LIN frequency plot:



I don't see on the green (univibe) curves an attemp to have flat phase.  Rather than this, the univibe curve on the LOG frequency scale (first plot) is rather a sloped straight line, whereas the two other circuits are curved.

STM.

[stm]

And for the grand finale, I am fulfilling my promise to puretube of simulating the notches produced by connecting four 4th-order stages, i.e. 16 stages.

Again, we have:

Blue: 4th order delay (stm)
Red: Equal capacitors (small stone)
Green: Staggered capacitors (univibe)



Have fun,

STM  :wink:

[puretube]

well: wow, gentlemen, this is stuff to chew on for a while!
Thank you a whole bunch, STM & Vsat     !!

Stephen: how do we interpret the cap-designations: 0.01µ-0.1µ in the Ev.?

[puretube]

information lost...

[StephenGiles]

Ton, you are up early, can't you sleep?


Stephen: how do we interpret the cap-designations: 0.01µ-0.1µ in the Ev.?

Ton, it states on the circuit "values will vary: typically.01 - .1uf, and I also notice that it states CD 4066 cannot be replaced with CD4016.

I am picking up my camera memory cards in half an hour so I may get the original circuits up later.
Stephen[/quote]

[puretube]

early to bed, and early to rise
makes a man healthy, wealthy, and nice!


well - I`m having a little nap, now and then...

:wink:

[StephenGiles]

Ton see my previous post
Stephen

[R.G.]

... why wouldn't you use four PWM resistors implemented with analog switches?

These can be made to match beautifully; LDRs are touchy at best to get matched parts.

Does anyone else notice the similarity of the feedback networks to a bridged-T filter network? Approximately the same components makes for two bridged-T notch filters, which is tuned by the same components changing. I suspect these things are cousins in the math.

[StephenGiles]

Here you go - Eventide Instant Phaser!
http://img.photobucket.com/albums/v317/StephenGiles/Etide_contrl.jpg
http://img.photobucket.com/albums/v317/StephenGiles/inst_phaser_shift.gif
http://img.photobucket.com/albums/v317/StephenGiles/inst_phaser_modulator.gif

Please compare these with my drawing so that I can correct any errors.
Stephen