zombie... flanger?

[jimmy]

hi all

i was looking through my schematic library today and i discovered the chorus-flanger from the big book of japanese schems. there are notes down the bottom to change the pedal from a chorus to a flanger to a delay. it uses the mn3007, so i thought is there any way to make the zombie chorus a zombie flanger? i had a quick look at the schems but the zombie uses a 4046 as the clock driver, while the chorus-flanger uses a mn3101.

any ideas? the possibilites could be almost endless... small clone flanger, etc...

cheers
Jim

[george]

I've toyed with the idea of doing something with my tonepad small clone, to give it flanging capability, something along the lines of:

- increase the clock rate from on the 4047 by reducing the size of the capacitor (this idea inspired by Mark Hammer's mods to the zombie chorus) thus getting the shorter delays that give flanging sounds.
- put some sort of feedback loop in to create the regeneration effect most flangers have, by connecting the output of the BBD back to the input via a (say) 100K pot to vary the intensity?  I've just had a look at the ultraflanger schemo, it has just such a loop with a 4700pf cap and 270k resistor in series with a 100K pot.

mind you I haven't tried it yet ... no doubt Mark Hammer will be able to say if it'll fly or not ...

Bear in mind that while different LFOs use different ICs (MN3101 and 4046 like you mentioned), I think you'll find the clock rate in all cases is governed by an RC network.  If you get the datasheet for your particular clocking chip, it'll show you how to change the values of the resistor and capacitor to get different clock rates.

[ExpAnonColin]

- put some sort of feedback loop in to create the regeneration effect most flangers have, by connecting the output of the BBD back to the input via a (say) 100K pot to vary the intensity?  I've just had a look at the ultraflanger schemo, it has just such a loop with a 4700pf cap and 270k resistor in series with a 100K pot.

You'd need to be much more careful than just a 100k pot.  Using just a resistor to create feedback loops with flangers and choruses is very touchy.  I'd personally recommend an op amp between the input and output to give it a bit more control.

-Colin

[ExpAnonColin]

Or, on the chorus flanger schematic that this thread was started about, there is an input and output buffer that for the MN3207 that the feedback loop goes around.  That would probably help.

-Colin

[Mark Hammer]

The advice doled out so far is pretty good.

In many respects, it would be smarter to chor-ify an Ultraflanger than it would be to flang-ify a Zombie

There are a few things that need to be in place to make a modulated BBD suitable for flanging:

1) An LFO capable of producing suitable sweep frequencies (slow enough and fast enough)

2) A clock range suitable for producing the appropriate delay times

3) A means for feeding back the delay signal

The LFO speed range is easily addressed.  Since most choruses will not yieldan audible chorus effect if you sweep them too slow, they usually don't vary over a very wide range.  Indeed, a sweep range of maybe 0.8hz - 7hz is usually sufficient to produce all the different chorus sounds people want.  You can extend the sweep range at the slow end by paralleling the cap in the integrator stage.  This would be C8 in the Zombie, and the 2.2uf tantalum cap in the Small Clone.  Double the existing value and you cut the sweep speed in half.

As noted, the clock range depends on the small-value cap in the clock generator.  The Zombie uses a pair of 1nf caps in series and switch-shunts one of them to have effective capacitances of either 1nf or 500pf.  Drop that capacitnace value down to 120pf and you get into flanging territory.  For the Heladito/Small Clone, drop the cap to, say, 68pf and that should do the same.

Typically, flangers will have two op-amp stages in series at the input and a transistor-based lowpass filter (although the third stage doesn't HAVE to be transistor).  The dry signal is fed to the output  from the first stage and the wet signal is fed back to the second stage, which serves as a mixer of not-yet-delayed and previously-delayed signal.  You will note that the Zombie lacks such a stage, and so does the Small Clone.  Were you to try and feed back the output of the lowpass filter after the BBD, you'd quickly realize that this is exactly the same signal path that the dry signal follows to get to the mixing at the end.

On the other hand, you will note that the Ultraflanger only has a single op-ampbased input stage, with the dry signal going to the output/mixer from the output of the first stage, and the return signal for resonance/regen coming back to the *input* of that first stage.  So, in theory you could try tapping the signal from pin 7 of U1b on the Zombie and returning it to pin 3 on U1a, via a cap, pot and fixed resistor.

One of the things that many BBD-based devices with regeneration include is a quick and dirty limiter at the point where the regen signal returns.  That limiter is intended to keep the sum of the regen and input signal from ever getting too big for the BBD.  Usually it will be a back-to-back diode pair and a small fixed resistor in series, to produce soft clipping.  Although the first op-amp stage in the Zombie does introduce a little gain, it isn't much (about 1.5), so the clipper may or may not be necessary.  Let your ears decide.  Certainly if the feedback path has the low end chopped with a small value cap in the signal path, the likelihood of the regen signal getting too big is drastically reduced.

In the Heladito, the regen signal would likely be tapped off the junction of the 1uf cap and 20k resistor after Q2 and fed back to pin 3 of IC1a, again via a cap, pot and fixed resistor.  In this case, the input stage of the Heladito has a gain of just under 6 so the soft-clipper/limiter is probably a smart idea here.  A pair of 1N914's and a 2.2k resistor in series seems about right.

[puretube]

In many respects, it would be smarter to chor-ify an Ultraflanger than it would be to flang-ify a Zombie

I love that phrase, Mark!

It should be noted however, that in the first case, the output lowpass filter`s cutoff-frequency might need to be lowered due to the lower clock frequency.

[george]

Mark as always you have deep knowledge coupled expressed in really practical terms we can use right away ... thanks!

[Mark Hammer]

Thanks to both Ton and George.  

And, as ever, Ton is right (hey, we make a good team!).  Remember that BBD-based circuits need lowpass filtering to make sure that the clock signal and aliasing are not part of the audio output of the pedal (or at least are a very very small part of it).  But of course, since the clock rate needed to produce different delay ranges for different effects is also different, that means the filter is normally set to a different cutoff point.

The rule of thumb that companies/designers use is to permit the largest bandwidth possible, given what you need to filter out.  What you need to filter out is largely a function of the longest delay time attained/produced  That means that if a flanger were to sweep from 0.5 to 8msec and a chorus swept from 4 to 12msec, it is a safe bet that the chorus's lowpass filtering is set to a cutoff about a half octave lower than the flanger's. (the interval change between 8msec and 12msec).

Why is this important?  Because with flanging the critical aspect to its character is the production of an increasing then decreasing number of audible notches.  When the flange sweep starts from its highest point (i.e., shortest delay), you begin with a largely untainted signal which gradually becomes "infected" with more and more notches as it sweeps lower.  For those notches to be *audible*, the flanger needs to have as close to full (20khz) bandwidth as can practically be achieved.  If the regen signal were tapped after, say, 8 poles worth of lowpass filtering at 6khz, whatever notches happen to be potentially created above 6khz when the device has swept to its highest point  simply will not be heard.  It will certainly be a quiet pedal as far as clock noise goes, but not necessarily one that will give you thrilling sweeps.

This is all the long way of saying that Ton is correct in suggesting one may well need to alter the cutoff point of the filters when reverting from one sort of effect to the other; lower for chorus, higher for flanging.

For a device like a BF-2-to-CE-2 conversion, that's kind of a big deal since there are roughly 6 poles of filtering, hence 6 components that would need to be switched or adjusted to re-optimize.  In the case of the Zombie, there are only 3 poles of lowpass filtering, but they are set relatively low to compensate for how few stages of filtering there are.  For instance, the R7/C2 network forms one lowpass section with a 3db-down point of about 3.4khz.  My filter-calculation chops are not that sharp but it stands to reason the 2-pole filter built around U1b is not that far off in terms of cutoff frequency.  In the case of the Small Clone,  there are also 6-poles of lowpass filtering (3 before the BBD and 3 after).

So, Ton is likely quite accurate in his assessment that the Zombie cannot be *easily* switched from chorus to flanger via a few simply part swaps and connections like the one in the Japanese book and still deliver a great version of each effect type.  That also suggests the Japanese design is not going to blow you away with its sound, because it will either a) not have enough bandwidth for satisfying flanging, or b) will have too much clock noise for comfortable chorus or c) have acceptable noise and modest bandwidth in all settings but the delay ranges for the flanging and chorus function will be too similar in delay range and sound.

In retrospect, this all sounds a little too discouraging, but it only sounds discouraging if your goal is to have a completely optimized version of 2 effects in one design.  If, on the other hand, your goal is simply to take a given pedal and expand the range of sounds you can wring from it via a few simple mods, then why not?  No, a Zombie  or Small Clone will not instantly turn into an Electric Mistress or A/DA Flanger, but so what?  It will be a Zombie or Small Clone that does something no one else's Zombie/Clone does in addition to what it normally does.  

One shouldn't let the need to mimic or copy effects stand in the way.  The real goal is to expand the tonal palette of all players, those with money for toys, and those without.

[puretube]

I`m seriously thinking `bout building 2 different filters with diff.  cutoff-frequencies for a future project (where size doesn`t matter)
for easy switching between them for longer/shorter delaytimes;
or at least a staggered multipole filter with a tap after the first half...

[Mark Hammer]

This is why the switched-resistor/PWM tracking filter on the old MXR Analog Delay (and its Ross clone) were such pure brilliance.  The filters, all 6 or 8 poles of them, went wherever the clock did.

Of course, it bears noting that the Analog Delay had a fixed clock, and did not sweep up and down like a chorus or flanger.  I'm not sure what sorts of artifacts might be expected were one to use a PWM tracking filter that moved with the LFO.

What would be sweet would be if the SSM2040, 2044, or IR3109 were readily available.  Something that would let you stick what amounts to a brickwall filter in there that is easily retuned with a single pot.  You'd actually have both noise optimization and timbral control with a single knob.

[puretube]

yes, it depends whether "size" or "price" doesn`t matter....
There are some new SwitchedCapacitorFilters on the
market to take over from the MF10s or the Reticons;
will place a link as soon as they cross my way again...

[jimmy]

wow, thats alot of info to absorb, but i kinda get it. i have taken the liberty of drawing up a design based on the front end and LFO of the zombie, with the clock driver of the small clone, complete with guesses of values that might achive the desired chorus/flange/delay combo. i decided to have a crack at the delay part as well. the schem is pretty simple, so the regen loop comes from befre the output cap to right before the input opamp stage, because its somewhat of a buffer. take a look at stick in some filters and make any corrections you might like.  problem is ive got the schem on a notepad and no access to a scanner... ill zoom with my camera and see how we go... sorry in advance for the handwriting haha

cheers
Jim

[puretube]

yes, it depends whether "size" or "price" doesn`t matter....
There are some new SwitchedCapacitorFilters on the
market to take over from the MF10s or the Reticons;
will place a link as soon as they cross my way again...

so they crossed my path today:
http://pdfserv.maxim-ic.com/en/ds/MAX7400-MAX7407.pdf  ...

[jimmy]

thats a pretty schmick piece of equipment...

cheers
Jim