Nyquist Versus The Zombie

[Joe Kramer]

Hey Friends,

Looking for an easy-to-build chorus and the Zombie looks dead simple.  My question is, how is this circuit getting away with absolutely no pre-BBD filtering whatsoever without rudely awakening grumpy Old Nyquist (alias Alias) from his comfortable nap?  Sure, there's not much above 4kHz on a clean guitar signal, but I've been known to patch a fuzz before the chorus now and then.  I suppose I'll find out when I breadboard, but I'd appreciate a few thoughts beforehand too.

Thanks,
Joe

[solderman]

Hey Friends,

Looking for an easy-to-build chorus and the Zombie looks dead simple.  My question is, how is this circuit getting away with absolutely no pre-BBD filtering whatsoever without rudely awakening grumpy Old Nyquist (alias Alias) from his comfortable nap?  Sure, there's not much above 4kHz on a clean guitar signal, but I've been known to patch a fuzz before the chorus now and then.  I suppose I'll find out when I breadboard, but I'd appreciate a few thoughts beforehand too.

Thanks,
Joe

Hi
This is of grate interest too me to since i am about to build this one from my 1590A layout. I chose the Zombie cause it was the only chorus i found that had so five components it wold fit in a 1590A without using DD pcb and surfice mount components. The only design thing i thought about was the lack of bias trimpot. but since iv seen and herd a five and the sounded good i was contempt with that. Now I have 4 3007 on there way and the cost a mill so I hope it works. 

//Solderman 

[brett]

Hi
the Zombie uses masses of post BBD filtering to take out transients ("stairs" and "blips", if you like).  General consensus is that it still has planty of noise and stuff.  The NA really appeals to me, but I haven't built one yet.  Must be getting lazy or something....
have a great day

[Mark Hammer]

The sound quality is actually pretty good, and having made two of them, my ears couldn't find a reason to increase lowpass filtering.  The bigger problem to address is the LFO ticking.

[Joe Kramer]

Thanks for the thoughts guys.

Solderman: You might want to look at this version of the Zombie, which is adapted for a 3207 BBD:

http://www.synthdiy.com/files/2007/zombie_chorus_mn3207_based.jpg

This is the one I'm thinking to build, since I'm a battery-user and low current-draw is a big factor.  This 3207 version adds a bias trimmer.  I suppose you could also add a clock cancel trimmer on the BBD output if you heard a need.

Brett: Sorry, what's "the NA?"  Yes, the post filtering quells that inevitable output clock noise.  I was just trying to understand how the possible input aliasing seems ignored here.  Say you clock the chip at 10K, which gives you about 50ms of delay and an upper frequency limit of about 5K before Nyquist starts tossing and turning.  I'm not sure what the Zombie is clocking at, but let's suppose it's something like 20K with about 25 ms of delay--a  good range for chorus.  Then your upper limit is about 10K, plenty comfortable for Nyquist's beauty sleep, until you send a fuzz into it and all hell breaks loose.  Or not.  I'll find out I guess.

Mark: I hold any sound-quality assessment of yours in high regard.  As for the ticking, I've had pretty good results with LFO-based stuff, as you may recall with my tremolo.  Attention to layout, PS decoupling, and input/output shielding usually does the trick.  It's also worth considering that those types of LFOs have a rather high output impedance, and so it helps to reduce ticking if you lighten the load to around 1Meg or so.

Thanks!
Joe

[snap]

NA

[Joe Kramer]

Ah, now I see how my topic could be misunderstood.  Serves me right for trying to be clever.  As they say, no good pun goes unpunished. 

By "Nyquist" I am referring to the man's theory, which says you must have at least two samples to meaningfully represent anything in digital terms.  Less than that and you get "aliasing."  That translates to digital audio (and to BBDs) by requiring twice the sample rate for the highest frequency of the input signal.  Without limiting the input frequency to about half the sampling rate, you run the risk of aliasing noise/distortion, which is where Mr. Nyquist has warned us not to go unawares.  So, just for clarity's sake, the Nyquist Aliaser, per se, doesn't interest me at the moment.  Sorry for the mix up.

Joe

[solderman]

Thanks for the thoughts guys.

Solderman: You might want to look at this version of the Zombie, which is adapted for a 3207 BBD:

http://www.synthdiy.com/files/2007/zombie_chorus_mn3207_based.jpg

This is the one I'm thinking to build, since I'm a battery-user and low current-draw is a big factor.  This 3207 version adds a bias trimmer.  I suppose you could also add a clock cancel trimmer on the BBD output if you heard a need.

Brett: Sorry, what's "the NA?"  Yes, the post filtering quells that inevitable output clock noise.  I was just trying to understand how the possible input aliasing seems ignored here.  Say you clock the chip at 10K, which gives you about 50ms of delay and an upper frequency limit of about 5K before Nyquist starts tossing and turning.  I'm not sure what the Zombie is clocking at, but let's suppose it's something like 20K with about 25 ms of delay--a  good range for chorus.  Then your upper limit is about 10K, plenty comfortable for Nyquist's beauty sleep, until you send a fuzz into it and all hell breaks loose.  Or not.  I'll find out I guess.

Mark: I hold any sound-quality assessment of yours in high regard.  As for the ticking, I've had pretty good results with LFO-based stuff, as you may recall with my tremolo.  Attention to layout, PS decoupling, and input/output shielding usually does the trick.  It's also worth considering that those types of LFOs have a rather high output impedance, and so it helps to reduce ticking if you lighten the load to around 1Meg or so.

Thanks!
Joe

Thx
That's grate to know. I havent seen that schematics. I can bye 3 3207 on one 3007 but i have allready 4 coming in with the mail Monday. The layout work 70%. Curious to here yore result with those.

//Solderman

[Mark Hammer]

Hi Joe,
Actually, I recently perfed the 3207 version and it sounds fine.  The only thing I couldn't figure out is why the cap values for setting the clock frequency seemed to be off. I had to drop them down to a max of 470pf to get anything that didn't sounded like bad echo.  For slow Leslie tones, I had to drop the cap value down to less then 220pf.  Not sure why.  Other than that, however, it was problem-free.

[Joe Kramer]

Thanks for the head's up on the 3207 version Mark, good to know.  Any idea what it's clocking at with your cap values instead of the stock .001s?  I guess anything beyond about 50ms delay will start to sound like slapback.  BTW, too bad that 4046 chip has to be controlled by discrete caps instead of a continuous pot, otherwise it's sort of nicer that the 3101/3102 as far as current draw and parts count.

[brett]

Hi
sorry for the confusion.  I thought you were running an LFO into the Nyquist Aliser (NA) to get some sort of chorus happening.
cheers

[Lurco]

Thanks for the head's up on the 3207 version Mark, good to know.  Any idea what it's clocking at with your cap values instead of the stock .001s?  I guess anything beyond about 50ms delay will start to sound like slapback.  BTW, too bad that 4046 chip has to be controlled by discrete caps instead of a continuous pot, otherwise it's sort of nicer that the 3101/3102 as far as current draw and parts count.

get yourself a 4046 datasheet with application notes and go voltage control after studying them!

[Joe Kramer]

Brett: All good, brother!

Lurco: Hm.  I Googled and thumbed through a few books, but didn't find anything that indicates a way to get continuous control over the 4046 VCO.  Every circuit I've seen uses caps and a switch. That'll should be fine for a quickie like the Zombie, but for a full-featured chorus or other, I guess I'll stick with the 3102/3101.

Joe

[puretube]

pin#9 in comb. w/ pins#11&12 resistors...

[Joe Kramer]

Ah, thanks for the pointers PT!

So, based on this info:

Back to the 74HC4046 VCO, . . .  Resistor R1 [pin 11 to GND] and capacitor C1 determine the frequency range of the VCO.

If I use, say, a 100K pot from pin 11 to ground, then I should get some continuous control of the VCO?  If this hypothesis is anywhere in the ballpark, just nod your head.   

Joe

[puretube]

 

[puretube]

err, that`s "ballpark", of course...

IIRC, gez had some better info, recently...

[Joe Kramer]

BAM!  That thread contains the info I was hoping to find--Thanks!  Though I didn't know the science, I could intuit the usefulness of the CD4046 just by seeing it applied in the Zombie.  I see Hollis's Ultra Flanger uses the 4046 as well (though the PDF schemo glaringly omits the value of the pin 11 resistor!).  Looks like a 1M pot on pin 11 might be give a better range of control.  The exactitudes of the math don't concern me too much, it's just great to find there's a cheap and versatile option for BBD clocking.  I'm excited about this.  Time to hit the breadboard!

Thanks again PT!

Joe

[solderman]

it's just great to find there's a cheap and versatile option for BBD clocking.  I'm excited about this.  Time to hit the breadboard!

Thanks again PT!

Joe

Yes, Really exiting.....BUT the BDD itself is still a safer investment than gold and an as limited as the oil reserv thees days so all price reducing efforts are welcome. But if any one has an idee' how to supst the BBD in analog chorus/flanger/delay etc: That wold be bingo for the DIY wallet.

//Solderman

[Joe Kramer]

Solderman, heard about these guys?

http://www.coolaudio.com/products.html

Their BBD subs are about two-thirds the way down the page.  Very affordable.