PWM phaser in the works

[Freppo]

Hi there folks

Here is a phaser project that i'm working on.  
It uses a CD4066 CMOS switch insted as variable resistors and a CD40106 for both the PWM and LFO duties (for now).

The distortion is from my camera and the hiss in the beginning is from the fans in my oscilloscope.
Around 1:08 I turn up the feedback (regen)

The schematic is just a first draft of what I have on the breadboard now. Things like Vref and power connections are not shown.

Full size: http://parasitstudio.weebly.com/uploads/2/4/4/9/2449159/parasit-phaser.pdf

I really like how I managed to utilize the CD40106 and keep the parts down, but I think I'll have to ditch the CD40106 LFO.  
The tringle wave from the CD40106 is kinda distorted and weak. A better LFO will allow the PWM sweep to go deeper and have a better sweep.
Now the pulse wave frequency is very high (about 95khz) to work with the low output LFO.. which also means high current consumption.

I will also add some noise reducing (emphasis/de-emphasis) and maybe a fixed phase shift stage in the feedback path, ala Ross Phaser.

Anyway, just though I'd share. Perhaps someone will find this interesting

cheers
/ Freppo

[mth5044]

Awesome idea! I look forward to your updates.

[armdnrdy]

Good job..it's nice to see someone working on something a bit outside of the norm.

I don't know if you've seen the ADA Final Phase schematic. The Final Phase uses CD4016s in the same fashion as the 4066....to switch in/out resistance in the phase stages.

[jatalahd]

Sounds good and the idea is definitely interesting. Thanks for sharing!

[Freppo]

Good job..it's nice to see someone working on something a bit outside of the norm.

I don't know if you've seen the ADA Final Phase schematic. The Final Phase uses CD4016s in the same fashion as the 4066....to switch in/out resistance in the phase stages.

Thanks. I'll check out that ADA Phaser, sounds interesting!
I started this because I wanted to build a phaser that doesn't need matched JFETS or LDRs.
The CD4066 solution is nothing new, but it works really well.

I'm on the fence if I should keep it as simple as possible..
...or go all out and add another bunch of stages or make it a bi-phaser and run it on +/- 12 volts and put it in a rack..
The extra headroom would sure be welcome.. hmm.. ideas ideas.

/ Freppo

[anotherjim]

Interesting.

Of that LFO, I'm sure you know the amplitude will be fixed by the threshold of the gate used. If you have an inverter spare, maybe you can overcome those thresholds by making a Schmitt trigger out of 2 of them in series, with a feedback resistor from the second output back to first input and a resistor in series with the input. The amplitude should depend on the ratio between feedback & input resistor. Input resistance has to be lower than the feedback. I know this works with ordinary CMOS gates, but I've never tried with ones that already contain Schmitt triggers.

The shape of the LFO may be worth a tweak. I've noticed the waveform on the charging capacitor in those circuits can vary according to the magnitude of the RC values. Have you tried a smaller capacitor and higher resistance? The wave always looks better to me with less charge/discharge current in the timing cap.

[stringsthings]

sounds good so far to me!     looking forward to putting this one together.  ( currently working on the Raygun Youth Chaos Fuzz )

[Freppo]

If you have an inverter spare, maybe you can overcome those thresholds by making a Schmitt trigger out of 2 of them in series

Great idea! But I can't quite wrap me head around how to do it. Do you have any example I can have a look at?
CMOS logic is still pretty new to me and i'm being ill and should be lying in bed today so my head is kinda mushy

Have you tried a smaller capacitor and higher resistance?

Yes, I've played around with all kinds of combinations and smaller caps makes the sweep more uneven at slow rates

I've been playing around with another LFO today, a wien-brige sinewave LFO. It has a much stronger output and nicer sweep.
But it add more complexity to the circuit and needs a dual-pot.. and I would rather keep it simple.

I think a charge pump could be a good solution to beef up the CD4106 LFO, and the headroom would not hurt for the s/n ratio.
But then again, added complexity.. Maybe I'll have to realize that there is no easy way to make a great phaser 

[Freppo]

sounds good so far to me!     looking forward to putting this one together.  ( currently working on the Raygun Youth Chaos Fuzz )

Awesome!    
This one might not be very DIY friendly though. It will certainly be too large for vero
But I will share a PCB layout when it's completed

[amptramp]

It looks like your design may converge on the Paul Nelson phaser of 2009.  It has a high-frequency astable made from one TL064 and another for the LFO.  I copied the schematic once but I don't have the link to where I got it.

[anotherjim]

a and b make a non-inverting gate. R1 is the feedback and R2 the input. This completes a Schmitt trigger. c is the integrator here making a conventional tri/sqr oscillator. This is with 4069/4049UB inverters. Schmitt should work using any CMOS logic so long as the feedback is positive (non-inverting).

So, I'm wondering if it will make an equivalent "single schmitt inverter" as you have with the 40106 by building the Schmitt with 2 inverters with resistors and feed the timing cap from the output of the first inverter (because to oscillate, it must be inverting).

[kingswayguitar]

the audio sounds pretty good to me freppo
cool seeing the scope at the same instant
cheers

[Freppo]

It looks like your design may converge on the Paul Nelson phaser of 2009.  It has a high-frequency astable made from one TL064 and another for the LFO.  I copied the schematic once but I don't have the link to where I got it.

I got the idea to use a CD4066 swith as a variable resistor from an article by Rod Elliott, but I've seen the Nelson phaser. At first I tried to use a astable oscillator followed by a comparator pwm. It worked ok, but had a limited pulse width because the triangle wave got a bit distorted at high frequencies. Probably because I did something wrong. The CD40106 is far simpler and the pulse wave is perfect so it seems like a easier solution.

The easiest way would be to use a PIC chip for the PWM/LFO (as R.G. Keen has done with a similar phaser), but I know nothing about programming so..

gate. R1 is the feedback and R2 the input. This completes a Schmitt trigger. c is the integrator here making a conventional tri/sqr oscillator.

Thanks! I'll give it a try

[deadastronaut]

sounds great freppo,

[Kipper4]

Sounds excellant. Thanks for sharing.

[Digital Larry]

Brilliant!   

[Freppo]

Thanks all for the positive comments

I'm aiming at doing a stompbox friendly version and a more extreme version (for the studio)
The only problem is that i'm running out of space on my breadboard 

/ Freppo

[Mark Hammer]

Nice.

1) Why are there two 100nf caps I the feedback loop?  Isn't one sufficient?

2) If you've going to use 3 chips, may as well make the third chip a quad, rather than dual, op-amp.  I know the 4066 only has 4 switches on it, but you can use the additional pair of op-amps for two fixed stages, as Rick Holt did for the Causality phaser.

3) The 40106 has 6 inverter stages, but you've only used 4.  That's not wrong, but there is the option of adding a second LFO to sum with the first one, and introducing some aperiodicity, or simply having two different rate presets to select between.

[Freppo]

Nice ideas Mark!

Why are there two 100nf caps I the feedback loop?  Isn't one sufficient?

Yes, I've actually been alittle confused about the purpose of the second cap myself. I took the feedback (more or less) out of another phaser that had two caps like that.
But as you say, one should be sufficient so I will remove the second one.

2) If you've going to use 3 chips, may as well make the third chip a quad, rather than dual, op-amp.  I know the 4066 only has 4 switches on it, but you can use the additional pair of op-amps for two fixed stages, as Rick Holt did for the Causality phaser.

I was going to make the third chip a quad anyway, but for other (less interesting) uses - one to buffer Vref (seem to reduce most of the LFO ticking bleeding into the signal)
and maybe that last op amp to buffer or beef up the LFO. Seems like it could be a good idea? But for the studio version I'll go all out and definitly add a couple of more passive stages aswell.

3. Great idea! I'll play around with a second LFO.
Thanks for the suggestions! 

[nelson]

I haven't visited the forum in years and the first thread I click on mentions me.

It's a sign, I must start building stompboxes again.

Good luck with the phaser. I haven't done any stompbox stuff in years, so sorry I don't have much to add.