Corroboration of Zombie mods

[Mark Hammer]

While doing some, er, "toilet reading" today, I was looking at a Japanese schematic for a MN3007-based flanger.  Like the Hollis Zombie, it too used a Vref coming from a 10k/15k divider.  This produces an offset voltage that goes to the op-amp just ahead of the BBD chip and provides adequate biasing, replacing the commonly found trimpot.

As many of you will recall, LFO-clicking in the Zombie has been a source of great annoyance to many.  In discussing the tick with Tim Larwill, he pointed out that he had been able to eliminate the ticking in a BBD-based unit he had made by making sure the Vref was very close to V+/2.  I did this on my own Zombie, and sure enough that fixed the problem.  I posted this mod at my site.

The Japanese Flanger I was looking at *also* uses a different Vref for the LFO, with equal-value resistors (3k3) instead of tapping off the 10k/15k junction.  So there you go, a published circuit which uses two *different* Vrefs, one for the BBD and input circuitry, and another for the LFO.  Great minds think alike.

[george]

Hi Mark

while we're on the subject of zombie mods, I've tried to implement what I think is your mod on my perfboard zombie build but the LFO stops sweeping.  I had tried to map your mod to RG's circuit diagram - the output on u3B just goes down to about 1.65V and stays there, whereas before it was sweeping up and down from about 4.65V to about 7.5V (as far as I could tell just using DMM)

Can you maybe do a quick circuit diagram or describe your mod in terms of RG's circuit diagram.

BTW I originally worked off John Hollis' handdrawn schemo, I notice that the vref resistor pair was 10K/12K, on RG's schemo this was changed to 10K/15K - would you know why.

As always you remain the modulation maestro,

Thanks and Regards
George

[Tim L]

Hi Mark,

How are you. I had also noticed this from the CE-2 schematic as well. You'll see that it also implemented sepearate bias for LFO and BBD. The vref actually used two 4.7K with a 10K trim sandwiched in between  for fine tuning.

Tim

[Mark Hammer]

I don't know why the one schem shows a 10k/12k pair and the other shows a 10k/15k pair, but what they share in common is that Vref (Vb) is not precisely half the supply voltage.

In the stock design, this network and resulting Vref goes to U1a via R2.  You will note that there are no caps between pin 3 of U1a and the input to the BBD, which means that any DC offset added to the input of the op-amp safely finds its way to the input of the BBD.  Were there any such series cap in the signal path, this trick would not work and one would need to resort to the standard trimpot on pin 3 of the MN3007, which would feed a portion of the supply voltage to that pin as a bias or DC offset voltage.

While the audio signal portion of the circuit is very tolerant of this asymmetrical arrangement (i.e., the signal can swing more on one side of the Vref than the other), the LFO is not.  It will work, of course, but when it swings to the "short" side of Vb/Vref it wants to swing wider and is prevented from doing so by lack of headroom.  This is the tick we hear.  The mod provides an alternate Vref to the LFO alone so that it behaves itself.  The audio portion remains untouched.

It occurs to me that it might be appropriate to reference all Vb contacts in the LFO circuit to the same 4.5v Vref.  The mod that I have shown here (http://hammer.ampage.org/files/DETICK~1.jpg) only provides this new Vref for pins 2 and 5 which are normally tied directly to Vb.  It provides considerable tick reduction, but there is still a wee bit of LFO noise in mine and maybe this might cure it.

The alternative is to replace R13 with a 10k resistor to provide the same 4.5v Vb for the whole circuit and attempt to kluge on a trimpot for setting the bias on pin 3.  The remainder of the circuit would behave nicely but the MN3007 would be off.  The kluge add-on would fix this.  You could mount it on a small piece of perfboard with 3 leads.  One goes to V+, one to ground, and the third provides the bias voltage to pin 3 of the MN3007.  As for what goes on the mini-board, stick on a 10k-25k trimpot, a 100k resistor, and a 1-10uf electrolytic cap.  One side of the trimpot goes to V+ and the other to gnd.  The 100k resistor goes from the trimpot wiper to pin 3 of the MN3007.  The cap goes from the wiper to ground (negative side to ground, pos to wiper).  The trimpot gets adjusted until you hear a chorus effect.  Variations in bias voltage won't screw up the MN3007 but you won't get any audio signal out of it unless the bias is set right.

I hope this is clear.

[george]

Hi Mark

what about putting the LFO on Vcc/2 and just keeping the input and output of U1a on the original Vref?

Regards
George

[Mark Hammer]

Yeah, that's fine, and is the mod I posted at my site.  There are essentially two objectives to be met - evening up voltage swing for the LFO and biasing the BBD - and any way you accomplish that is fine.  I suggested the retrofit mini-board because it involves the least amount of additional components or board changes.  I used "flying" components to fix the Vb issue on my own but it looks a little too "early Electro-Harmonix" in construction technique :wink: for my tastes. By simply using equal value resistors for R12/R13, you've essentially fixed the Vref issue without any physical change to the board itself.  The added bias trimpot also allows you to use other supply voltages for a little more headroom by simply adjusting the bias voltage, which I think is a better solution than relying entirely on the existing R12/R13 to do it for you and being constrained to +9v supply forever.  Remember, even though they will work fine with +9v, the MN3007 and 4046 are quite happy up to as much as +15vdc.

But again, anything that gets you to a fix faster by addressing the two issues noted is fine.

[swt]

so mark....can i just run my zombie with 15 volts?. I mean can i use the same circuit, no vlues changing at that voltage?. Thanks!
PS: remember i was asking for a dual supply chorus?. Maybe higher voltage on this circuit is the answer :wink:

[Mark Hammer]

Not so fast.

Certainly the audio portion and LFO can run fine at 15vdc, but the BBD needs to be biased a little more carefully than that.  I'm also unsure as to whether the clock frequencies produced by the 4046 hold true at 15vdc as well as at 9vdc.

So, "function"?  Yes.  Function identically or better with no other changes than supply voltage?  Not necessarily.

I'd appreciate some input from the CMOS gurus with respect to the relationship between supply voltage and clock rate here.  For instance if the timing cap needs to change as a function of supply voltage, that is an easy change to make but it still needs to be made in order to retain the same delay range and tone.

[Ed Rembold]

I'm no guru,
but all things being equal-  a 4046 does Not change output freq.
when changing supply B+ , say from 9v to 15v.   but all things are seldom equal.

Ed R.

[swt]

Another one...what if i leave lfo and 4046, with 9volts, and put 15 or more volts in the audio portion?. or maybe even using the 3007 with 9 volts. Is it possible?. I just want more voltage in the audio section. How can i do this?